diff --git a/cmd/zpool/zpool_main.c b/cmd/zpool/zpool_main.c index 2311d4f046f6..474f3bbb1096 100644 --- a/cmd/zpool/zpool_main.c +++ b/cmd/zpool/zpool_main.c @@ -1,10837 +1,10848 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2011, 2020 by Delphix. All rights reserved. * Copyright (c) 2012 by Frederik Wessels. All rights reserved. * Copyright (c) 2012 by Cyril Plisko. All rights reserved. * Copyright (c) 2013 by Prasad Joshi (sTec). All rights reserved. * Copyright 2016 Igor Kozhukhov . * Copyright (c) 2017 Datto Inc. * Copyright (c) 2017 Open-E, Inc. All Rights Reserved. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2019, loli10K * Copyright (c) 2021, Colm Buckley * Copyright [2021] Hewlett Packard Enterprise Development LP */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zpool_util.h" #include "zfs_comutil.h" #include "zfeature_common.h" #include "statcommon.h" libzfs_handle_t *g_zfs; static int zpool_do_create(int, char **); static int zpool_do_destroy(int, char **); static int zpool_do_add(int, char **); static int zpool_do_remove(int, char **); static int zpool_do_labelclear(int, char **); static int zpool_do_checkpoint(int, char **); static int zpool_do_list(int, char **); static int zpool_do_iostat(int, char **); static int zpool_do_status(int, char **); static int zpool_do_online(int, char **); static int zpool_do_offline(int, char **); static int zpool_do_clear(int, char **); static int zpool_do_reopen(int, char **); static int zpool_do_reguid(int, char **); static int zpool_do_attach(int, char **); static int zpool_do_detach(int, char **); static int zpool_do_replace(int, char **); static int zpool_do_split(int, char **); static int zpool_do_initialize(int, char **); static int zpool_do_scrub(int, char **); static int zpool_do_resilver(int, char **); static int zpool_do_trim(int, char **); static int zpool_do_import(int, char **); static int zpool_do_export(int, char **); static int zpool_do_upgrade(int, char **); static int zpool_do_history(int, char **); static int zpool_do_events(int, char **); static int zpool_do_get(int, char **); static int zpool_do_set(int, char **); static int zpool_do_sync(int, char **); static int zpool_do_version(int, char **); static int zpool_do_wait(int, char **); static zpool_compat_status_t zpool_do_load_compat( const char *, boolean_t *); /* * These libumem hooks provide a reasonable set of defaults for the allocator's * debugging facilities. */ #ifdef DEBUG const char * _umem_debug_init(void) { return ("default,verbose"); /* $UMEM_DEBUG setting */ } const char * _umem_logging_init(void) { return ("fail,contents"); /* $UMEM_LOGGING setting */ } #endif typedef enum { HELP_ADD, HELP_ATTACH, HELP_CLEAR, HELP_CREATE, HELP_CHECKPOINT, HELP_DESTROY, HELP_DETACH, HELP_EXPORT, HELP_HISTORY, HELP_IMPORT, HELP_IOSTAT, HELP_LABELCLEAR, HELP_LIST, HELP_OFFLINE, HELP_ONLINE, HELP_REPLACE, HELP_REMOVE, HELP_INITIALIZE, HELP_SCRUB, HELP_RESILVER, HELP_TRIM, HELP_STATUS, HELP_UPGRADE, HELP_EVENTS, HELP_GET, HELP_SET, HELP_SPLIT, HELP_SYNC, HELP_REGUID, HELP_REOPEN, HELP_VERSION, HELP_WAIT } zpool_help_t; /* * Flags for stats to display with "zpool iostats" */ enum iostat_type { IOS_DEFAULT = 0, IOS_LATENCY = 1, IOS_QUEUES = 2, IOS_L_HISTO = 3, IOS_RQ_HISTO = 4, IOS_COUNT, /* always last element */ }; /* iostat_type entries as bitmasks */ #define IOS_DEFAULT_M (1ULL << IOS_DEFAULT) #define IOS_LATENCY_M (1ULL << IOS_LATENCY) #define IOS_QUEUES_M (1ULL << IOS_QUEUES) #define IOS_L_HISTO_M (1ULL << IOS_L_HISTO) #define IOS_RQ_HISTO_M (1ULL << IOS_RQ_HISTO) /* Mask of all the histo bits */ #define IOS_ANYHISTO_M (IOS_L_HISTO_M | IOS_RQ_HISTO_M) /* * Lookup table for iostat flags to nvlist names. Basically a list * of all the nvlists a flag requires. Also specifies the order in * which data gets printed in zpool iostat. */ static const char *vsx_type_to_nvlist[IOS_COUNT][13] = { [IOS_L_HISTO] = { ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO, ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO, NULL}, [IOS_LATENCY] = { ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO, NULL}, [IOS_QUEUES] = { ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE, NULL}, [IOS_RQ_HISTO] = { ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO, ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO, ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO, ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO, ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO, ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO, ZPOOL_CONFIG_VDEV_IND_TRIM_HISTO, ZPOOL_CONFIG_VDEV_AGG_TRIM_HISTO, NULL}, }; /* * Given a cb->cb_flags with a histogram bit set, return the iostat_type. * Right now, only one histo bit is ever set at one time, so we can * just do a highbit64(a) */ #define IOS_HISTO_IDX(a) (highbit64(a & IOS_ANYHISTO_M) - 1) typedef struct zpool_command { const char *name; int (*func)(int, char **); zpool_help_t usage; } zpool_command_t; /* * Master command table. Each ZFS command has a name, associated function, and * usage message. The usage messages need to be internationalized, so we have * to have a function to return the usage message based on a command index. * * These commands are organized according to how they are displayed in the usage * message. An empty command (one with a NULL name) indicates an empty line in * the generic usage message. */ static zpool_command_t command_table[] = { { "version", zpool_do_version, HELP_VERSION }, { NULL }, { "create", zpool_do_create, HELP_CREATE }, { "destroy", zpool_do_destroy, HELP_DESTROY }, { NULL }, { "add", zpool_do_add, HELP_ADD }, { "remove", zpool_do_remove, HELP_REMOVE }, { NULL }, { "labelclear", zpool_do_labelclear, HELP_LABELCLEAR }, { NULL }, { "checkpoint", zpool_do_checkpoint, HELP_CHECKPOINT }, { NULL }, { "list", zpool_do_list, HELP_LIST }, { "iostat", zpool_do_iostat, HELP_IOSTAT }, { "status", zpool_do_status, HELP_STATUS }, { NULL }, { "online", zpool_do_online, HELP_ONLINE }, { "offline", zpool_do_offline, HELP_OFFLINE }, { "clear", zpool_do_clear, HELP_CLEAR }, { "reopen", zpool_do_reopen, HELP_REOPEN }, { NULL }, { "attach", zpool_do_attach, HELP_ATTACH }, { "detach", zpool_do_detach, HELP_DETACH }, { "replace", zpool_do_replace, HELP_REPLACE }, { "split", zpool_do_split, HELP_SPLIT }, { NULL }, { "initialize", zpool_do_initialize, HELP_INITIALIZE }, { "resilver", zpool_do_resilver, HELP_RESILVER }, { "scrub", zpool_do_scrub, HELP_SCRUB }, { "trim", zpool_do_trim, HELP_TRIM }, { NULL }, { "import", zpool_do_import, HELP_IMPORT }, { "export", zpool_do_export, HELP_EXPORT }, { "upgrade", zpool_do_upgrade, HELP_UPGRADE }, { "reguid", zpool_do_reguid, HELP_REGUID }, { NULL }, { "history", zpool_do_history, HELP_HISTORY }, { "events", zpool_do_events, HELP_EVENTS }, { NULL }, { "get", zpool_do_get, HELP_GET }, { "set", zpool_do_set, HELP_SET }, { "sync", zpool_do_sync, HELP_SYNC }, { NULL }, { "wait", zpool_do_wait, HELP_WAIT }, }; #define NCOMMAND (ARRAY_SIZE(command_table)) #define VDEV_ALLOC_CLASS_LOGS "logs" static zpool_command_t *current_command; static char history_str[HIS_MAX_RECORD_LEN]; static boolean_t log_history = B_TRUE; static uint_t timestamp_fmt = NODATE; static const char * get_usage(zpool_help_t idx) { switch (idx) { case HELP_ADD: return (gettext("\tadd [-fgLnP] [-o property=value] " " ...\n")); case HELP_ATTACH: return (gettext("\tattach [-fsw] [-o property=value] " " \n")); case HELP_CLEAR: return (gettext("\tclear [-nF] [device]\n")); case HELP_CREATE: return (gettext("\tcreate [-fnd] [-o property=value] ... \n" "\t [-O file-system-property=value] ... \n" "\t [-m mountpoint] [-R root] ...\n")); case HELP_CHECKPOINT: return (gettext("\tcheckpoint [-d [-w]] ...\n")); case HELP_DESTROY: return (gettext("\tdestroy [-f] \n")); case HELP_DETACH: return (gettext("\tdetach \n")); case HELP_EXPORT: return (gettext("\texport [-af] ...\n")); case HELP_HISTORY: return (gettext("\thistory [-il] [] ...\n")); case HELP_IMPORT: return (gettext("\timport [-d dir] [-D]\n" "\timport [-o mntopts] [-o property=value] ... \n" "\t [-d dir | -c cachefile] [-D] [-l] [-f] [-m] [-N] " "[-R root] [-F [-n]] -a\n" "\timport [-o mntopts] [-o property=value] ... \n" "\t [-d dir | -c cachefile] [-D] [-l] [-f] [-m] [-N] " "[-R root] [-F [-n]]\n" "\t [--rewind-to-checkpoint] [newpool]\n")); case HELP_IOSTAT: return (gettext("\tiostat [[[-c [script1,script2,...]" "[-lq]]|[-rw]] [-T d | u] [-ghHLpPvy]\n" "\t [[pool ...]|[pool vdev ...]|[vdev ...]]" " [[-n] interval [count]]\n")); case HELP_LABELCLEAR: return (gettext("\tlabelclear [-f] \n")); case HELP_LIST: return (gettext("\tlist [-gHLpPv] [-o property[,...]] " "[-T d|u] [pool] ... \n" "\t [interval [count]]\n")); case HELP_OFFLINE: return (gettext("\toffline [-f] [-t] ...\n")); case HELP_ONLINE: return (gettext("\tonline [-e] ...\n")); case HELP_REPLACE: return (gettext("\treplace [-fsw] [-o property=value] " " [new-device]\n")); case HELP_REMOVE: return (gettext("\tremove [-npsw] ...\n")); case HELP_REOPEN: return (gettext("\treopen [-n] \n")); case HELP_INITIALIZE: return (gettext("\tinitialize [-c | -s] [-w] " "[ ...]\n")); case HELP_SCRUB: return (gettext("\tscrub [-s | -p] [-w] ...\n")); case HELP_RESILVER: return (gettext("\tresilver ...\n")); case HELP_TRIM: return (gettext("\ttrim [-dw] [-r ] [-c | -s] " "[ ...]\n")); case HELP_STATUS: return (gettext("\tstatus [-c [script1,script2,...]] " "[-igLpPstvxD] [-T d|u] [pool] ... \n" "\t [interval [count]]\n")); case HELP_UPGRADE: return (gettext("\tupgrade\n" "\tupgrade -v\n" "\tupgrade [-V version] <-a | pool ...>\n")); case HELP_EVENTS: return (gettext("\tevents [-vHf [pool] | -c]\n")); case HELP_GET: return (gettext("\tget [-Hp] [-o \"all\" | field[,...]] " "<\"all\" | property[,...]> ...\n")); case HELP_SET: return (gettext("\tset \n")); case HELP_SPLIT: return (gettext("\tsplit [-gLnPl] [-R altroot] [-o mntopts]\n" "\t [-o property=value] " "[ ...]\n")); case HELP_REGUID: return (gettext("\treguid \n")); case HELP_SYNC: return (gettext("\tsync [pool] ...\n")); case HELP_VERSION: return (gettext("\tversion\n")); case HELP_WAIT: return (gettext("\twait [-Hp] [-T d|u] [-t [,...]] " " [interval]\n")); } abort(); /* NOTREACHED */ } static void zpool_collect_leaves(zpool_handle_t *zhp, nvlist_t *nvroot, nvlist_t *res) { uint_t children = 0; nvlist_t **child; uint_t i; (void) nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children); if (children == 0) { char *path = zpool_vdev_name(g_zfs, zhp, nvroot, VDEV_NAME_PATH); if (strcmp(path, VDEV_TYPE_INDIRECT) != 0 && strcmp(path, VDEV_TYPE_HOLE) != 0) fnvlist_add_boolean(res, path); free(path); return; } for (i = 0; i < children; i++) { zpool_collect_leaves(zhp, child[i], res); } } /* * Callback routine that will print out a pool property value. */ static int print_prop_cb(int prop, void *cb) { FILE *fp = cb; (void) fprintf(fp, "\t%-19s ", zpool_prop_to_name(prop)); if (zpool_prop_readonly(prop)) (void) fprintf(fp, " NO "); else (void) fprintf(fp, " YES "); if (zpool_prop_values(prop) == NULL) (void) fprintf(fp, "-\n"); else (void) fprintf(fp, "%s\n", zpool_prop_values(prop)); return (ZPROP_CONT); } /* * Display usage message. If we're inside a command, display only the usage for * that command. Otherwise, iterate over the entire command table and display * a complete usage message. */ static void usage(boolean_t requested) { FILE *fp = requested ? stdout : stderr; if (current_command == NULL) { int i; (void) fprintf(fp, gettext("usage: zpool command args ...\n")); (void) fprintf(fp, gettext("where 'command' is one of the following:\n\n")); for (i = 0; i < NCOMMAND; i++) { if (command_table[i].name == NULL) (void) fprintf(fp, "\n"); else (void) fprintf(fp, "%s", get_usage(command_table[i].usage)); } } else { (void) fprintf(fp, gettext("usage:\n")); (void) fprintf(fp, "%s", get_usage(current_command->usage)); } if (current_command != NULL && ((strcmp(current_command->name, "set") == 0) || (strcmp(current_command->name, "get") == 0) || (strcmp(current_command->name, "list") == 0))) { (void) fprintf(fp, gettext("\nthe following properties are supported:\n")); (void) fprintf(fp, "\n\t%-19s %s %s\n\n", "PROPERTY", "EDIT", "VALUES"); /* Iterate over all properties */ (void) zprop_iter(print_prop_cb, fp, B_FALSE, B_TRUE, ZFS_TYPE_POOL); (void) fprintf(fp, "\t%-19s ", "feature@..."); (void) fprintf(fp, "YES disabled | enabled | active\n"); (void) fprintf(fp, gettext("\nThe feature@ properties must be " "appended with a feature name.\nSee zpool-features(7).\n")); } /* * See comments at end of main(). */ if (getenv("ZFS_ABORT") != NULL) { (void) printf("dumping core by request\n"); abort(); } exit(requested ? 0 : 2); } /* * zpool initialize [-c | -s] [-w] [ ...] * Initialize all unused blocks in the specified vdevs, or all vdevs in the pool * if none specified. * * -c Cancel. Ends active initializing. * -s Suspend. Initializing can then be restarted with no flags. * -w Wait. Blocks until initializing has completed. */ int zpool_do_initialize(int argc, char **argv) { int c; char *poolname; zpool_handle_t *zhp; nvlist_t *vdevs; int err = 0; boolean_t wait = B_FALSE; struct option long_options[] = { {"cancel", no_argument, NULL, 'c'}, {"suspend", no_argument, NULL, 's'}, {"wait", no_argument, NULL, 'w'}, {0, 0, 0, 0} }; pool_initialize_func_t cmd_type = POOL_INITIALIZE_START; while ((c = getopt_long(argc, argv, "csw", long_options, NULL)) != -1) { switch (c) { case 'c': if (cmd_type != POOL_INITIALIZE_START && cmd_type != POOL_INITIALIZE_CANCEL) { (void) fprintf(stderr, gettext("-c cannot be " "combined with other options\n")); usage(B_FALSE); } cmd_type = POOL_INITIALIZE_CANCEL; break; case 's': if (cmd_type != POOL_INITIALIZE_START && cmd_type != POOL_INITIALIZE_SUSPEND) { (void) fprintf(stderr, gettext("-s cannot be " "combined with other options\n")); usage(B_FALSE); } cmd_type = POOL_INITIALIZE_SUSPEND; break; case 'w': wait = B_TRUE; break; case '?': if (optopt != 0) { (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); } else { (void) fprintf(stderr, gettext("invalid option '%s'\n"), argv[optind - 1]); } usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); return (-1); } if (wait && (cmd_type != POOL_INITIALIZE_START)) { (void) fprintf(stderr, gettext("-w cannot be used with -c or " "-s\n")); usage(B_FALSE); } poolname = argv[0]; zhp = zpool_open(g_zfs, poolname); if (zhp == NULL) return (-1); vdevs = fnvlist_alloc(); if (argc == 1) { /* no individual leaf vdevs specified, so add them all */ nvlist_t *config = zpool_get_config(zhp, NULL); nvlist_t *nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); zpool_collect_leaves(zhp, nvroot, vdevs); } else { for (int i = 1; i < argc; i++) { fnvlist_add_boolean(vdevs, argv[i]); } } if (wait) err = zpool_initialize_wait(zhp, cmd_type, vdevs); else err = zpool_initialize(zhp, cmd_type, vdevs); fnvlist_free(vdevs); zpool_close(zhp); return (err); } /* * print a pool vdev config for dry runs */ static void print_vdev_tree(zpool_handle_t *zhp, const char *name, nvlist_t *nv, int indent, const char *match, int name_flags) { nvlist_t **child; uint_t c, children; char *vname; boolean_t printed = B_FALSE; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) { if (name != NULL) (void) printf("\t%*s%s\n", indent, "", name); return; } for (c = 0; c < children; c++) { uint64_t is_log = B_FALSE, is_hole = B_FALSE; char *class = ""; (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE, &is_hole); if (is_hole == B_TRUE) { continue; } (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log); if (is_log) class = VDEV_ALLOC_BIAS_LOG; (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &class); if (strcmp(match, class) != 0) continue; if (!printed && name != NULL) { (void) printf("\t%*s%s\n", indent, "", name); printed = B_TRUE; } vname = zpool_vdev_name(g_zfs, zhp, child[c], name_flags); print_vdev_tree(zhp, vname, child[c], indent + 2, "", name_flags); free(vname); } } /* * Print the list of l2cache devices for dry runs. */ static void print_cache_list(nvlist_t *nv, int indent) { nvlist_t **child; uint_t c, children; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) == 0 && children > 0) { (void) printf("\t%*s%s\n", indent, "", "cache"); } else { return; } for (c = 0; c < children; c++) { char *vname; vname = zpool_vdev_name(g_zfs, NULL, child[c], 0); (void) printf("\t%*s%s\n", indent + 2, "", vname); free(vname); } } /* * Print the list of spares for dry runs. */ static void print_spare_list(nvlist_t *nv, int indent) { nvlist_t **child; uint_t c, children; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0 && children > 0) { (void) printf("\t%*s%s\n", indent, "", "spares"); } else { return; } for (c = 0; c < children; c++) { char *vname; vname = zpool_vdev_name(g_zfs, NULL, child[c], 0); (void) printf("\t%*s%s\n", indent + 2, "", vname); free(vname); } } static boolean_t prop_list_contains_feature(nvlist_t *proplist) { nvpair_t *nvp; for (nvp = nvlist_next_nvpair(proplist, NULL); NULL != nvp; nvp = nvlist_next_nvpair(proplist, nvp)) { if (zpool_prop_feature(nvpair_name(nvp))) return (B_TRUE); } return (B_FALSE); } /* * Add a property pair (name, string-value) into a property nvlist. */ static int add_prop_list(const char *propname, char *propval, nvlist_t **props, boolean_t poolprop) { zpool_prop_t prop = ZPOOL_PROP_INVAL; nvlist_t *proplist; const char *normnm; char *strval; if (*props == NULL && nvlist_alloc(props, NV_UNIQUE_NAME, 0) != 0) { (void) fprintf(stderr, gettext("internal error: out of memory\n")); return (1); } proplist = *props; if (poolprop) { const char *vname = zpool_prop_to_name(ZPOOL_PROP_VERSION); const char *cname = zpool_prop_to_name(ZPOOL_PROP_COMPATIBILITY); if ((prop = zpool_name_to_prop(propname)) == ZPOOL_PROP_INVAL && !zpool_prop_feature(propname)) { (void) fprintf(stderr, gettext("property '%s' is " "not a valid pool property\n"), propname); return (2); } /* * feature@ properties and version should not be specified * at the same time. */ if ((prop == ZPOOL_PROP_INVAL && zpool_prop_feature(propname) && nvlist_exists(proplist, vname)) || (prop == ZPOOL_PROP_VERSION && prop_list_contains_feature(proplist))) { (void) fprintf(stderr, gettext("'feature@' and " "'version' properties cannot be specified " "together\n")); return (2); } /* * if version is specified, only "legacy" compatibility * may be requested */ if ((prop == ZPOOL_PROP_COMPATIBILITY && strcmp(propval, ZPOOL_COMPAT_LEGACY) != 0 && nvlist_exists(proplist, vname)) || (prop == ZPOOL_PROP_VERSION && nvlist_exists(proplist, cname) && strcmp(fnvlist_lookup_string(proplist, cname), ZPOOL_COMPAT_LEGACY) != 0)) { (void) fprintf(stderr, gettext("when 'version' is " "specified, the 'compatibility' feature may only " "be set to '" ZPOOL_COMPAT_LEGACY "'\n")); return (2); } if (zpool_prop_feature(propname)) normnm = propname; else normnm = zpool_prop_to_name(prop); } else { zfs_prop_t fsprop = zfs_name_to_prop(propname); if (zfs_prop_valid_for_type(fsprop, ZFS_TYPE_FILESYSTEM, B_FALSE)) { normnm = zfs_prop_to_name(fsprop); } else if (zfs_prop_user(propname) || zfs_prop_userquota(propname)) { normnm = propname; } else { (void) fprintf(stderr, gettext("property '%s' is " "not a valid filesystem property\n"), propname); return (2); } } if (nvlist_lookup_string(proplist, normnm, &strval) == 0 && prop != ZPOOL_PROP_CACHEFILE) { (void) fprintf(stderr, gettext("property '%s' " "specified multiple times\n"), propname); return (2); } if (nvlist_add_string(proplist, normnm, propval) != 0) { (void) fprintf(stderr, gettext("internal " "error: out of memory\n")); return (1); } return (0); } /* * Set a default property pair (name, string-value) in a property nvlist */ static int add_prop_list_default(const char *propname, char *propval, nvlist_t **props, boolean_t poolprop) { char *pval; if (nvlist_lookup_string(*props, propname, &pval) == 0) return (0); return (add_prop_list(propname, propval, props, B_TRUE)); } /* * zpool add [-fgLnP] [-o property=value] ... * * -f Force addition of devices, even if they appear in use * -g Display guid for individual vdev name. * -L Follow links when resolving vdev path name. * -n Do not add the devices, but display the resulting layout if * they were to be added. * -o Set property=value. * -P Display full path for vdev name. * * Adds the given vdevs to 'pool'. As with create, the bulk of this work is * handled by make_root_vdev(), which constructs the nvlist needed to pass to * libzfs. */ int zpool_do_add(int argc, char **argv) { boolean_t force = B_FALSE; boolean_t dryrun = B_FALSE; int name_flags = 0; int c; nvlist_t *nvroot; char *poolname; int ret; zpool_handle_t *zhp; nvlist_t *config; nvlist_t *props = NULL; char *propval; /* check options */ while ((c = getopt(argc, argv, "fgLno:P")) != -1) { switch (c) { case 'f': force = B_TRUE; break; case 'g': name_flags |= VDEV_NAME_GUID; break; case 'L': name_flags |= VDEV_NAME_FOLLOW_LINKS; break; case 'n': dryrun = B_TRUE; break; case 'o': if ((propval = strchr(optarg, '=')) == NULL) { (void) fprintf(stderr, gettext("missing " "'=' for -o option\n")); usage(B_FALSE); } *propval = '\0'; propval++; if ((strcmp(optarg, ZPOOL_CONFIG_ASHIFT) != 0) || (add_prop_list(optarg, propval, &props, B_TRUE))) usage(B_FALSE); break; case 'P': name_flags |= VDEV_NAME_PATH; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing vdev specification\n")); usage(B_FALSE); } poolname = argv[0]; argc--; argv++; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); if ((config = zpool_get_config(zhp, NULL)) == NULL) { (void) fprintf(stderr, gettext("pool '%s' is unavailable\n"), poolname); zpool_close(zhp); return (1); } /* unless manually specified use "ashift" pool property (if set) */ if (!nvlist_exists(props, ZPOOL_CONFIG_ASHIFT)) { int intval; zprop_source_t src; char strval[ZPOOL_MAXPROPLEN]; intval = zpool_get_prop_int(zhp, ZPOOL_PROP_ASHIFT, &src); if (src != ZPROP_SRC_DEFAULT) { (void) sprintf(strval, "%" PRId32, intval); verify(add_prop_list(ZPOOL_CONFIG_ASHIFT, strval, &props, B_TRUE) == 0); } } /* pass off to make_root_vdev for processing */ nvroot = make_root_vdev(zhp, props, force, !force, B_FALSE, dryrun, argc, argv); if (nvroot == NULL) { zpool_close(zhp); return (1); } if (dryrun) { nvlist_t *poolnvroot; nvlist_t **l2child, **sparechild; uint_t l2children, sparechildren, c; char *vname; boolean_t hadcache = B_FALSE, hadspare = B_FALSE; verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &poolnvroot) == 0); (void) printf(gettext("would update '%s' to the following " "configuration:\n\n"), zpool_get_name(zhp)); /* print original main pool and new tree */ print_vdev_tree(zhp, poolname, poolnvroot, 0, "", name_flags | VDEV_NAME_TYPE_ID); print_vdev_tree(zhp, NULL, nvroot, 0, "", name_flags); /* print other classes: 'dedup', 'special', and 'log' */ if (zfs_special_devs(poolnvroot, VDEV_ALLOC_BIAS_DEDUP)) { print_vdev_tree(zhp, "dedup", poolnvroot, 0, VDEV_ALLOC_BIAS_DEDUP, name_flags); print_vdev_tree(zhp, NULL, nvroot, 0, VDEV_ALLOC_BIAS_DEDUP, name_flags); } else if (zfs_special_devs(nvroot, VDEV_ALLOC_BIAS_DEDUP)) { print_vdev_tree(zhp, "dedup", nvroot, 0, VDEV_ALLOC_BIAS_DEDUP, name_flags); } if (zfs_special_devs(poolnvroot, VDEV_ALLOC_BIAS_SPECIAL)) { print_vdev_tree(zhp, "special", poolnvroot, 0, VDEV_ALLOC_BIAS_SPECIAL, name_flags); print_vdev_tree(zhp, NULL, nvroot, 0, VDEV_ALLOC_BIAS_SPECIAL, name_flags); } else if (zfs_special_devs(nvroot, VDEV_ALLOC_BIAS_SPECIAL)) { print_vdev_tree(zhp, "special", nvroot, 0, VDEV_ALLOC_BIAS_SPECIAL, name_flags); } if (num_logs(poolnvroot) > 0) { print_vdev_tree(zhp, "logs", poolnvroot, 0, VDEV_ALLOC_BIAS_LOG, name_flags); print_vdev_tree(zhp, NULL, nvroot, 0, VDEV_ALLOC_BIAS_LOG, name_flags); } else if (num_logs(nvroot) > 0) { print_vdev_tree(zhp, "logs", nvroot, 0, VDEV_ALLOC_BIAS_LOG, name_flags); } /* Do the same for the caches */ if (nvlist_lookup_nvlist_array(poolnvroot, ZPOOL_CONFIG_L2CACHE, &l2child, &l2children) == 0 && l2children) { hadcache = B_TRUE; (void) printf(gettext("\tcache\n")); for (c = 0; c < l2children; c++) { vname = zpool_vdev_name(g_zfs, NULL, l2child[c], name_flags); (void) printf("\t %s\n", vname); free(vname); } } if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2child, &l2children) == 0 && l2children) { if (!hadcache) (void) printf(gettext("\tcache\n")); for (c = 0; c < l2children; c++) { vname = zpool_vdev_name(g_zfs, NULL, l2child[c], name_flags); (void) printf("\t %s\n", vname); free(vname); } } /* And finally the spares */ if (nvlist_lookup_nvlist_array(poolnvroot, ZPOOL_CONFIG_SPARES, &sparechild, &sparechildren) == 0 && sparechildren > 0) { hadspare = B_TRUE; (void) printf(gettext("\tspares\n")); for (c = 0; c < sparechildren; c++) { vname = zpool_vdev_name(g_zfs, NULL, sparechild[c], name_flags); (void) printf("\t %s\n", vname); free(vname); } } if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &sparechild, &sparechildren) == 0 && sparechildren > 0) { if (!hadspare) (void) printf(gettext("\tspares\n")); for (c = 0; c < sparechildren; c++) { vname = zpool_vdev_name(g_zfs, NULL, sparechild[c], name_flags); (void) printf("\t %s\n", vname); free(vname); } } ret = 0; } else { ret = (zpool_add(zhp, nvroot) != 0); } nvlist_free(props); nvlist_free(nvroot); zpool_close(zhp); return (ret); } /* * zpool remove [-npsw] ... * * Removes the given vdev from the pool. */ int zpool_do_remove(int argc, char **argv) { char *poolname; int i, ret = 0; zpool_handle_t *zhp = NULL; boolean_t stop = B_FALSE; int c; boolean_t noop = B_FALSE; boolean_t parsable = B_FALSE; boolean_t wait = B_FALSE; /* check options */ while ((c = getopt(argc, argv, "npsw")) != -1) { switch (c) { case 'n': noop = B_TRUE; break; case 'p': parsable = B_TRUE; break; case 's': stop = B_TRUE; break; case 'w': wait = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } poolname = argv[0]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); if (stop && noop) { (void) fprintf(stderr, gettext("stop request ignored\n")); return (0); } if (stop) { if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } if (zpool_vdev_remove_cancel(zhp) != 0) ret = 1; if (wait) { (void) fprintf(stderr, gettext("invalid option " "combination: -w cannot be used with -s\n")); usage(B_FALSE); } } else { if (argc < 2) { (void) fprintf(stderr, gettext("missing device\n")); usage(B_FALSE); } for (i = 1; i < argc; i++) { if (noop) { uint64_t size; if (zpool_vdev_indirect_size(zhp, argv[i], &size) != 0) { ret = 1; break; } if (parsable) { (void) printf("%s %llu\n", argv[i], (unsigned long long)size); } else { char valstr[32]; zfs_nicenum(size, valstr, sizeof (valstr)); (void) printf("Memory that will be " "used after removing %s: %s\n", argv[i], valstr); } } else { if (zpool_vdev_remove(zhp, argv[i]) != 0) ret = 1; } } if (ret == 0 && wait) ret = zpool_wait(zhp, ZPOOL_WAIT_REMOVE); } zpool_close(zhp); return (ret); } /* * Return 1 if a vdev is active (being used in a pool) * Return 0 if a vdev is inactive (offlined or faulted, or not in active pool) * * This is useful for checking if a disk in an active pool is offlined or * faulted. */ static int vdev_is_active(char *vdev_path) { int fd; fd = open(vdev_path, O_EXCL); if (fd < 0) { return (1); /* cant open O_EXCL - disk is active */ } close(fd); return (0); /* disk is inactive in the pool */ } /* * zpool labelclear [-f] * * -f Force clearing the label for the vdevs which are members of * the exported or foreign pools. * * Verifies that the vdev is not active and zeros out the label information * on the device. */ int zpool_do_labelclear(int argc, char **argv) { char vdev[MAXPATHLEN]; char *name = NULL; struct stat st; int c, fd = -1, ret = 0; nvlist_t *config; pool_state_t state; boolean_t inuse = B_FALSE; boolean_t force = B_FALSE; /* check options */ while ((c = getopt(argc, argv, "f")) != -1) { switch (c) { case 'f': force = B_TRUE; break; default: (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get vdev name */ if (argc < 1) { (void) fprintf(stderr, gettext("missing vdev name\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } /* * Check if we were given absolute path and use it as is. * Otherwise if the provided vdev name doesn't point to a file, * try prepending expected disk paths and partition numbers. */ (void) strlcpy(vdev, argv[0], sizeof (vdev)); if (vdev[0] != '/' && stat(vdev, &st) != 0) { int error; error = zfs_resolve_shortname(argv[0], vdev, MAXPATHLEN); if (error == 0 && zfs_dev_is_whole_disk(vdev)) { if (zfs_append_partition(vdev, MAXPATHLEN) == -1) error = ENOENT; } if (error || (stat(vdev, &st) != 0)) { (void) fprintf(stderr, gettext( "failed to find device %s, try specifying absolute " "path instead\n"), argv[0]); return (1); } } if ((fd = open(vdev, O_RDWR)) < 0) { (void) fprintf(stderr, gettext("failed to open %s: %s\n"), vdev, strerror(errno)); return (1); } /* * Flush all dirty pages for the block device. This should not be * fatal when the device does not support BLKFLSBUF as would be the * case for a file vdev. */ if ((zfs_dev_flush(fd) != 0) && (errno != ENOTTY)) (void) fprintf(stderr, gettext("failed to invalidate " "cache for %s: %s\n"), vdev, strerror(errno)); if (zpool_read_label(fd, &config, NULL) != 0) { (void) fprintf(stderr, gettext("failed to read label from %s\n"), vdev); ret = 1; goto errout; } nvlist_free(config); ret = zpool_in_use(g_zfs, fd, &state, &name, &inuse); if (ret != 0) { (void) fprintf(stderr, gettext("failed to check state for %s\n"), vdev); ret = 1; goto errout; } if (!inuse) goto wipe_label; switch (state) { default: case POOL_STATE_ACTIVE: case POOL_STATE_SPARE: case POOL_STATE_L2CACHE: /* * We allow the user to call 'zpool offline -f' * on an offlined disk in an active pool. We can check if * the disk is online by calling vdev_is_active(). */ if (force && !vdev_is_active(vdev)) break; (void) fprintf(stderr, gettext( "%s is a member (%s) of pool \"%s\""), vdev, zpool_pool_state_to_name(state), name); if (force) { (void) fprintf(stderr, gettext( ". Offline the disk first to clear its label.")); } printf("\n"); ret = 1; goto errout; case POOL_STATE_EXPORTED: if (force) break; (void) fprintf(stderr, gettext( "use '-f' to override the following error:\n" "%s is a member of exported pool \"%s\"\n"), vdev, name); ret = 1; goto errout; case POOL_STATE_POTENTIALLY_ACTIVE: if (force) break; (void) fprintf(stderr, gettext( "use '-f' to override the following error:\n" "%s is a member of potentially active pool \"%s\"\n"), vdev, name); ret = 1; goto errout; case POOL_STATE_DESTROYED: /* inuse should never be set for a destroyed pool */ assert(0); break; } wipe_label: ret = zpool_clear_label(fd); if (ret != 0) { (void) fprintf(stderr, gettext("failed to clear label for %s\n"), vdev); } errout: free(name); (void) close(fd); return (ret); } /* * zpool create [-fnd] [-o property=value] ... * [-O file-system-property=value] ... * [-R root] [-m mountpoint] ... * * -f Force creation, even if devices appear in use * -n Do not create the pool, but display the resulting layout if it * were to be created. * -R Create a pool under an alternate root * -m Set default mountpoint for the root dataset. By default it's * '/' * -o Set property=value. * -o Set feature@feature=enabled|disabled. * -d Don't automatically enable all supported pool features * (individual features can be enabled with -o). * -O Set fsproperty=value in the pool's root file system * * Creates the named pool according to the given vdev specification. The * bulk of the vdev processing is done in make_root_vdev() in zpool_vdev.c. * Once we get the nvlist back from make_root_vdev(), we either print out the * contents (if '-n' was specified), or pass it to libzfs to do the creation. */ int zpool_do_create(int argc, char **argv) { boolean_t force = B_FALSE; boolean_t dryrun = B_FALSE; boolean_t enable_pool_features = B_TRUE; int c; nvlist_t *nvroot = NULL; char *poolname; char *tname = NULL; int ret = 1; char *altroot = NULL; char *compat = NULL; char *mountpoint = NULL; nvlist_t *fsprops = NULL; nvlist_t *props = NULL; char *propval; /* check options */ while ((c = getopt(argc, argv, ":fndR:m:o:O:t:")) != -1) { switch (c) { case 'f': force = B_TRUE; break; case 'n': dryrun = B_TRUE; break; case 'd': enable_pool_features = B_FALSE; break; case 'R': altroot = optarg; if (add_prop_list(zpool_prop_to_name( ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE)) goto errout; if (add_prop_list_default(zpool_prop_to_name( ZPOOL_PROP_CACHEFILE), "none", &props, B_TRUE)) goto errout; break; case 'm': /* Equivalent to -O mountpoint=optarg */ mountpoint = optarg; break; case 'o': if ((propval = strchr(optarg, '=')) == NULL) { (void) fprintf(stderr, gettext("missing " "'=' for -o option\n")); goto errout; } *propval = '\0'; propval++; if (add_prop_list(optarg, propval, &props, B_TRUE)) goto errout; /* * If the user is creating a pool that doesn't support * feature flags, don't enable any features. */ if (zpool_name_to_prop(optarg) == ZPOOL_PROP_VERSION) { char *end; u_longlong_t ver; ver = strtoull(propval, &end, 10); if (*end == '\0' && ver < SPA_VERSION_FEATURES) { enable_pool_features = B_FALSE; } } if (zpool_name_to_prop(optarg) == ZPOOL_PROP_ALTROOT) altroot = propval; if (zpool_name_to_prop(optarg) == ZPOOL_PROP_COMPATIBILITY) compat = propval; break; case 'O': if ((propval = strchr(optarg, '=')) == NULL) { (void) fprintf(stderr, gettext("missing " "'=' for -O option\n")); goto errout; } *propval = '\0'; propval++; /* * Mountpoints are checked and then added later. * Uniquely among properties, they can be specified * more than once, to avoid conflict with -m. */ if (0 == strcmp(optarg, zfs_prop_to_name(ZFS_PROP_MOUNTPOINT))) { mountpoint = propval; } else if (add_prop_list(optarg, propval, &fsprops, B_FALSE)) { goto errout; } break; case 't': /* * Sanity check temporary pool name. */ if (strchr(optarg, '/') != NULL) { (void) fprintf(stderr, gettext("cannot create " "'%s': invalid character '/' in temporary " "name\n"), optarg); (void) fprintf(stderr, gettext("use 'zfs " "create' to create a dataset\n")); goto errout; } if (add_prop_list(zpool_prop_to_name( ZPOOL_PROP_TNAME), optarg, &props, B_TRUE)) goto errout; if (add_prop_list_default(zpool_prop_to_name( ZPOOL_PROP_CACHEFILE), "none", &props, B_TRUE)) goto errout; tname = optarg; break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); goto badusage; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); goto badusage; } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); goto badusage; } if (argc < 2) { (void) fprintf(stderr, gettext("missing vdev specification\n")); goto badusage; } poolname = argv[0]; /* * As a special case, check for use of '/' in the name, and direct the * user to use 'zfs create' instead. */ if (strchr(poolname, '/') != NULL) { (void) fprintf(stderr, gettext("cannot create '%s': invalid " "character '/' in pool name\n"), poolname); (void) fprintf(stderr, gettext("use 'zfs create' to " "create a dataset\n")); goto errout; } /* pass off to make_root_vdev for bulk processing */ nvroot = make_root_vdev(NULL, props, force, !force, B_FALSE, dryrun, argc - 1, argv + 1); if (nvroot == NULL) goto errout; /* make_root_vdev() allows 0 toplevel children if there are spares */ if (!zfs_allocatable_devs(nvroot)) { (void) fprintf(stderr, gettext("invalid vdev " "specification: at least one toplevel vdev must be " "specified\n")); goto errout; } if (altroot != NULL && altroot[0] != '/') { (void) fprintf(stderr, gettext("invalid alternate root '%s': " "must be an absolute path\n"), altroot); goto errout; } /* * Check the validity of the mountpoint and direct the user to use the * '-m' mountpoint option if it looks like its in use. */ if (mountpoint == NULL || (strcmp(mountpoint, ZFS_MOUNTPOINT_LEGACY) != 0 && strcmp(mountpoint, ZFS_MOUNTPOINT_NONE) != 0)) { char buf[MAXPATHLEN]; DIR *dirp; if (mountpoint && mountpoint[0] != '/') { (void) fprintf(stderr, gettext("invalid mountpoint " "'%s': must be an absolute path, 'legacy', or " "'none'\n"), mountpoint); goto errout; } if (mountpoint == NULL) { if (altroot != NULL) (void) snprintf(buf, sizeof (buf), "%s/%s", altroot, poolname); else (void) snprintf(buf, sizeof (buf), "/%s", poolname); } else { if (altroot != NULL) (void) snprintf(buf, sizeof (buf), "%s%s", altroot, mountpoint); else (void) snprintf(buf, sizeof (buf), "%s", mountpoint); } if ((dirp = opendir(buf)) == NULL && errno != ENOENT) { (void) fprintf(stderr, gettext("mountpoint '%s' : " "%s\n"), buf, strerror(errno)); (void) fprintf(stderr, gettext("use '-m' " "option to provide a different default\n")); goto errout; } else if (dirp) { int count = 0; while (count < 3 && readdir(dirp) != NULL) count++; (void) closedir(dirp); if (count > 2) { (void) fprintf(stderr, gettext("mountpoint " "'%s' exists and is not empty\n"), buf); (void) fprintf(stderr, gettext("use '-m' " "option to provide a " "different default\n")); goto errout; } } } /* * Now that the mountpoint's validity has been checked, ensure that * the property is set appropriately prior to creating the pool. */ if (mountpoint != NULL) { ret = add_prop_list(zfs_prop_to_name(ZFS_PROP_MOUNTPOINT), mountpoint, &fsprops, B_FALSE); if (ret != 0) goto errout; } ret = 1; if (dryrun) { /* * For a dry run invocation, print out a basic message and run * through all the vdevs in the list and print out in an * appropriate hierarchy. */ (void) printf(gettext("would create '%s' with the " "following layout:\n\n"), poolname); print_vdev_tree(NULL, poolname, nvroot, 0, "", 0); print_vdev_tree(NULL, "dedup", nvroot, 0, VDEV_ALLOC_BIAS_DEDUP, 0); print_vdev_tree(NULL, "special", nvroot, 0, VDEV_ALLOC_BIAS_SPECIAL, 0); print_vdev_tree(NULL, "logs", nvroot, 0, VDEV_ALLOC_BIAS_LOG, 0); print_cache_list(nvroot, 0); print_spare_list(nvroot, 0); ret = 0; } else { /* * Load in feature set. * Note: if compatibility property not given, we'll have * NULL, which means 'all features'. */ boolean_t requested_features[SPA_FEATURES]; if (zpool_do_load_compat(compat, requested_features) != ZPOOL_COMPATIBILITY_OK) goto errout; /* * props contains list of features to enable. * For each feature: * - remove it if feature@name=disabled * - leave it there if feature@name=enabled * - add it if: * - enable_pool_features (ie: no '-d' or '-o version') * - it's supported by the kernel module * - it's in the requested feature set * - warn if it's enabled but not in compat */ for (spa_feature_t i = 0; i < SPA_FEATURES; i++) { char propname[MAXPATHLEN]; char *propval; zfeature_info_t *feat = &spa_feature_table[i]; (void) snprintf(propname, sizeof (propname), "feature@%s", feat->fi_uname); if (!nvlist_lookup_string(props, propname, &propval)) { if (strcmp(propval, ZFS_FEATURE_DISABLED) == 0) (void) nvlist_remove_all(props, propname); if (strcmp(propval, ZFS_FEATURE_ENABLED) == 0 && !requested_features[i]) (void) fprintf(stderr, gettext( "Warning: feature \"%s\" enabled " "but is not in specified " "'compatibility' feature set.\n"), feat->fi_uname); } else if ( enable_pool_features && feat->fi_zfs_mod_supported && requested_features[i]) { ret = add_prop_list(propname, ZFS_FEATURE_ENABLED, &props, B_TRUE); if (ret != 0) goto errout; } } ret = 1; if (zpool_create(g_zfs, poolname, nvroot, props, fsprops) == 0) { zfs_handle_t *pool = zfs_open(g_zfs, tname ? tname : poolname, ZFS_TYPE_FILESYSTEM); if (pool != NULL) { if (zfs_mount(pool, NULL, 0) == 0) { ret = zfs_shareall(pool); zfs_commit_all_shares(); } zfs_close(pool); } } else if (libzfs_errno(g_zfs) == EZFS_INVALIDNAME) { (void) fprintf(stderr, gettext("pool name may have " "been omitted\n")); } } errout: nvlist_free(nvroot); nvlist_free(fsprops); nvlist_free(props); return (ret); badusage: nvlist_free(fsprops); nvlist_free(props); usage(B_FALSE); return (2); } /* * zpool destroy * * -f Forcefully unmount any datasets * * Destroy the given pool. Automatically unmounts any datasets in the pool. */ int zpool_do_destroy(int argc, char **argv) { boolean_t force = B_FALSE; int c; char *pool; zpool_handle_t *zhp; int ret; /* check options */ while ((c = getopt(argc, argv, "f")) != -1) { switch (c) { case 'f': force = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* check arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool argument\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } pool = argv[0]; if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) { /* * As a special case, check for use of '/' in the name, and * direct the user to use 'zfs destroy' instead. */ if (strchr(pool, '/') != NULL) (void) fprintf(stderr, gettext("use 'zfs destroy' to " "destroy a dataset\n")); return (1); } if (zpool_disable_datasets(zhp, force) != 0) { (void) fprintf(stderr, gettext("could not destroy '%s': " "could not unmount datasets\n"), zpool_get_name(zhp)); zpool_close(zhp); return (1); } /* The history must be logged as part of the export */ log_history = B_FALSE; ret = (zpool_destroy(zhp, history_str) != 0); zpool_close(zhp); return (ret); } typedef struct export_cbdata { boolean_t force; boolean_t hardforce; } export_cbdata_t; /* * Export one pool */ static int zpool_export_one(zpool_handle_t *zhp, void *data) { export_cbdata_t *cb = data; if (zpool_disable_datasets(zhp, cb->force) != 0) return (1); /* The history must be logged as part of the export */ log_history = B_FALSE; if (cb->hardforce) { if (zpool_export_force(zhp, history_str) != 0) return (1); } else if (zpool_export(zhp, cb->force, history_str) != 0) { return (1); } return (0); } /* * zpool export [-f] ... * * -a Export all pools * -f Forcefully unmount datasets * * Export the given pools. By default, the command will attempt to cleanly * unmount any active datasets within the pool. If the '-f' flag is specified, * then the datasets will be forcefully unmounted. */ int zpool_do_export(int argc, char **argv) { export_cbdata_t cb; boolean_t do_all = B_FALSE; boolean_t force = B_FALSE; boolean_t hardforce = B_FALSE; int c, ret; /* check options */ while ((c = getopt(argc, argv, "afF")) != -1) { switch (c) { case 'a': do_all = B_TRUE; break; case 'f': force = B_TRUE; break; case 'F': hardforce = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } cb.force = force; cb.hardforce = hardforce; argc -= optind; argv += optind; if (do_all) { if (argc != 0) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } return (for_each_pool(argc, argv, B_TRUE, NULL, B_FALSE, zpool_export_one, &cb)); } /* check arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool argument\n")); usage(B_FALSE); } ret = for_each_pool(argc, argv, B_TRUE, NULL, B_FALSE, zpool_export_one, &cb); return (ret); } /* * Given a vdev configuration, determine the maximum width needed for the device * name column. */ static int max_width(zpool_handle_t *zhp, nvlist_t *nv, int depth, int max, int name_flags) { char *name; nvlist_t **child; uint_t c, children; int ret; name = zpool_vdev_name(g_zfs, zhp, nv, name_flags); if (strlen(name) + depth > max) max = strlen(name) + depth; free(name); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0) { for (c = 0; c < children; c++) if ((ret = max_width(zhp, child[c], depth + 2, max, name_flags)) > max) max = ret; } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) == 0) { for (c = 0; c < children; c++) if ((ret = max_width(zhp, child[c], depth + 2, max, name_flags)) > max) max = ret; } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { for (c = 0; c < children; c++) if ((ret = max_width(zhp, child[c], depth + 2, max, name_flags)) > max) max = ret; } return (max); } typedef struct spare_cbdata { uint64_t cb_guid; zpool_handle_t *cb_zhp; } spare_cbdata_t; static boolean_t find_vdev(nvlist_t *nv, uint64_t search) { uint64_t guid; nvlist_t **child; uint_t c, children; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0 && search == guid) return (B_TRUE); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { for (c = 0; c < children; c++) if (find_vdev(child[c], search)) return (B_TRUE); } return (B_FALSE); } static int find_spare(zpool_handle_t *zhp, void *data) { spare_cbdata_t *cbp = data; nvlist_t *config, *nvroot; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); if (find_vdev(nvroot, cbp->cb_guid)) { cbp->cb_zhp = zhp; return (1); } zpool_close(zhp); return (0); } typedef struct status_cbdata { int cb_count; int cb_name_flags; int cb_namewidth; boolean_t cb_allpools; boolean_t cb_verbose; boolean_t cb_literal; boolean_t cb_explain; boolean_t cb_first; boolean_t cb_dedup_stats; boolean_t cb_print_status; boolean_t cb_print_slow_ios; boolean_t cb_print_vdev_init; boolean_t cb_print_vdev_trim; vdev_cmd_data_list_t *vcdl; } status_cbdata_t; /* Return 1 if string is NULL, empty, or whitespace; return 0 otherwise. */ static int is_blank_str(char *str) { while (str != NULL && *str != '\0') { if (!isblank(*str)) return (0); str++; } return (1); } /* Print command output lines for specific vdev in a specific pool */ static void zpool_print_cmd(vdev_cmd_data_list_t *vcdl, const char *pool, char *path) { vdev_cmd_data_t *data; int i, j; char *val; for (i = 0; i < vcdl->count; i++) { if ((strcmp(vcdl->data[i].path, path) != 0) || (strcmp(vcdl->data[i].pool, pool) != 0)) { /* Not the vdev we're looking for */ continue; } data = &vcdl->data[i]; /* Print out all the output values for this vdev */ for (j = 0; j < vcdl->uniq_cols_cnt; j++) { val = NULL; /* Does this vdev have values for this column? */ for (int k = 0; k < data->cols_cnt; k++) { if (strcmp(data->cols[k], vcdl->uniq_cols[j]) == 0) { /* yes it does, record the value */ val = data->lines[k]; break; } } /* * Mark empty values with dashes to make output * awk-able. */ if (val == NULL || is_blank_str(val)) val = "-"; printf("%*s", vcdl->uniq_cols_width[j], val); if (j < vcdl->uniq_cols_cnt - 1) printf(" "); } /* Print out any values that aren't in a column at the end */ for (j = data->cols_cnt; j < data->lines_cnt; j++) { /* Did we have any columns? If so print a spacer. */ if (vcdl->uniq_cols_cnt > 0) printf(" "); val = data->lines[j]; printf("%s", val ? val : ""); } break; } } /* * Print vdev initialization status for leaves */ static void print_status_initialize(vdev_stat_t *vs, boolean_t verbose) { if (verbose) { if ((vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE || vs->vs_initialize_state == VDEV_INITIALIZE_SUSPENDED || vs->vs_initialize_state == VDEV_INITIALIZE_COMPLETE) && !vs->vs_scan_removing) { char zbuf[1024]; char tbuf[256]; struct tm zaction_ts; time_t t = vs->vs_initialize_action_time; int initialize_pct = 100; if (vs->vs_initialize_state != VDEV_INITIALIZE_COMPLETE) { initialize_pct = (vs->vs_initialize_bytes_done * 100 / (vs->vs_initialize_bytes_est + 1)); } (void) localtime_r(&t, &zaction_ts); (void) strftime(tbuf, sizeof (tbuf), "%c", &zaction_ts); switch (vs->vs_initialize_state) { case VDEV_INITIALIZE_SUSPENDED: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("suspended, started at"), tbuf); break; case VDEV_INITIALIZE_ACTIVE: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("started at"), tbuf); break; case VDEV_INITIALIZE_COMPLETE: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("completed at"), tbuf); break; } (void) printf(gettext(" (%d%% initialized%s)"), initialize_pct, zbuf); } else { (void) printf(gettext(" (uninitialized)")); } } else if (vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE) { (void) printf(gettext(" (initializing)")); } } /* * Print vdev TRIM status for leaves */ static void print_status_trim(vdev_stat_t *vs, boolean_t verbose) { if (verbose) { if ((vs->vs_trim_state == VDEV_TRIM_ACTIVE || vs->vs_trim_state == VDEV_TRIM_SUSPENDED || vs->vs_trim_state == VDEV_TRIM_COMPLETE) && !vs->vs_scan_removing) { char zbuf[1024]; char tbuf[256]; struct tm zaction_ts; time_t t = vs->vs_trim_action_time; int trim_pct = 100; if (vs->vs_trim_state != VDEV_TRIM_COMPLETE) { trim_pct = (vs->vs_trim_bytes_done * 100 / (vs->vs_trim_bytes_est + 1)); } (void) localtime_r(&t, &zaction_ts); (void) strftime(tbuf, sizeof (tbuf), "%c", &zaction_ts); switch (vs->vs_trim_state) { case VDEV_TRIM_SUSPENDED: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("suspended, started at"), tbuf); break; case VDEV_TRIM_ACTIVE: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("started at"), tbuf); break; case VDEV_TRIM_COMPLETE: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("completed at"), tbuf); break; } (void) printf(gettext(" (%d%% trimmed%s)"), trim_pct, zbuf); } else if (vs->vs_trim_notsup) { (void) printf(gettext(" (trim unsupported)")); } else { (void) printf(gettext(" (untrimmed)")); } } else if (vs->vs_trim_state == VDEV_TRIM_ACTIVE) { (void) printf(gettext(" (trimming)")); } } /* * Return the color associated with a health string. This includes returning * NULL for no color change. */ static char * health_str_to_color(const char *health) { if (strcmp(health, gettext("FAULTED")) == 0 || strcmp(health, gettext("SUSPENDED")) == 0 || strcmp(health, gettext("UNAVAIL")) == 0) { return (ANSI_RED); } if (strcmp(health, gettext("OFFLINE")) == 0 || strcmp(health, gettext("DEGRADED")) == 0 || strcmp(health, gettext("REMOVED")) == 0) { return (ANSI_YELLOW); } return (NULL); } /* * Print out configuration state as requested by status_callback. */ static void print_status_config(zpool_handle_t *zhp, status_cbdata_t *cb, const char *name, nvlist_t *nv, int depth, boolean_t isspare, vdev_rebuild_stat_t *vrs) { nvlist_t **child, *root; uint_t c, i, vsc, children; pool_scan_stat_t *ps = NULL; vdev_stat_t *vs; char rbuf[6], wbuf[6], cbuf[6]; char *vname; uint64_t notpresent; spare_cbdata_t spare_cb; const char *state; char *type; char *path = NULL; char *rcolor = NULL, *wcolor = NULL, *ccolor = NULL; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc) == 0); verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0); if (strcmp(type, VDEV_TYPE_INDIRECT) == 0) return; state = zpool_state_to_name(vs->vs_state, vs->vs_aux); if (isspare) { /* * For hot spares, we use the terms 'INUSE' and 'AVAILABLE' for * online drives. */ if (vs->vs_aux == VDEV_AUX_SPARED) state = gettext("INUSE"); else if (vs->vs_state == VDEV_STATE_HEALTHY) state = gettext("AVAIL"); } printf_color(health_str_to_color(state), "\t%*s%-*s %-8s", depth, "", cb->cb_namewidth - depth, name, state); if (!isspare) { if (vs->vs_read_errors) rcolor = ANSI_RED; if (vs->vs_write_errors) wcolor = ANSI_RED; if (vs->vs_checksum_errors) ccolor = ANSI_RED; if (cb->cb_literal) { printf(" "); printf_color(rcolor, "%5llu", (u_longlong_t)vs->vs_read_errors); printf(" "); printf_color(wcolor, "%5llu", (u_longlong_t)vs->vs_write_errors); printf(" "); printf_color(ccolor, "%5llu", (u_longlong_t)vs->vs_checksum_errors); } else { zfs_nicenum(vs->vs_read_errors, rbuf, sizeof (rbuf)); zfs_nicenum(vs->vs_write_errors, wbuf, sizeof (wbuf)); zfs_nicenum(vs->vs_checksum_errors, cbuf, sizeof (cbuf)); printf(" "); printf_color(rcolor, "%5s", rbuf); printf(" "); printf_color(wcolor, "%5s", wbuf); printf(" "); printf_color(ccolor, "%5s", cbuf); } if (cb->cb_print_slow_ios) { if (children == 0) { /* Only leafs vdevs have slow IOs */ zfs_nicenum(vs->vs_slow_ios, rbuf, sizeof (rbuf)); } else { snprintf(rbuf, sizeof (rbuf), "-"); } if (cb->cb_literal) printf(" %5llu", (u_longlong_t)vs->vs_slow_ios); else printf(" %5s", rbuf); } } if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, ¬present) == 0) { verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0); (void) printf(" %s %s", gettext("was"), path); } else if (vs->vs_aux != 0) { (void) printf(" "); color_start(ANSI_RED); switch (vs->vs_aux) { case VDEV_AUX_OPEN_FAILED: (void) printf(gettext("cannot open")); break; case VDEV_AUX_BAD_GUID_SUM: (void) printf(gettext("missing device")); break; case VDEV_AUX_NO_REPLICAS: (void) printf(gettext("insufficient replicas")); break; case VDEV_AUX_VERSION_NEWER: (void) printf(gettext("newer version")); break; case VDEV_AUX_UNSUP_FEAT: (void) printf(gettext("unsupported feature(s)")); break; case VDEV_AUX_ASHIFT_TOO_BIG: (void) printf(gettext("unsupported minimum blocksize")); break; case VDEV_AUX_SPARED: verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &spare_cb.cb_guid) == 0); if (zpool_iter(g_zfs, find_spare, &spare_cb) == 1) { if (strcmp(zpool_get_name(spare_cb.cb_zhp), zpool_get_name(zhp)) == 0) (void) printf(gettext("currently in " "use")); else (void) printf(gettext("in use by " "pool '%s'"), zpool_get_name(spare_cb.cb_zhp)); zpool_close(spare_cb.cb_zhp); } else { (void) printf(gettext("currently in use")); } break; case VDEV_AUX_ERR_EXCEEDED: (void) printf(gettext("too many errors")); break; case VDEV_AUX_IO_FAILURE: (void) printf(gettext("experienced I/O failures")); break; case VDEV_AUX_BAD_LOG: (void) printf(gettext("bad intent log")); break; case VDEV_AUX_EXTERNAL: (void) printf(gettext("external device fault")); break; case VDEV_AUX_SPLIT_POOL: (void) printf(gettext("split into new pool")); break; case VDEV_AUX_ACTIVE: (void) printf(gettext("currently in use")); break; case VDEV_AUX_CHILDREN_OFFLINE: (void) printf(gettext("all children offline")); break; case VDEV_AUX_BAD_LABEL: (void) printf(gettext("invalid label")); break; default: (void) printf(gettext("corrupted data")); break; } color_end(); } else if (children == 0 && !isspare && getenv("ZPOOL_STATUS_NON_NATIVE_ASHIFT_IGNORE") == NULL && VDEV_STAT_VALID(vs_physical_ashift, vsc) && vs->vs_configured_ashift < vs->vs_physical_ashift) { (void) printf( gettext(" block size: %dB configured, %dB native"), 1 << vs->vs_configured_ashift, 1 << vs->vs_physical_ashift); } /* The root vdev has the scrub/resilver stats */ root = fnvlist_lookup_nvlist(zpool_get_config(zhp, NULL), ZPOOL_CONFIG_VDEV_TREE); (void) nvlist_lookup_uint64_array(root, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &c); /* * If you force fault a drive that's resilvering, its scan stats can * get frozen in time, giving the false impression that it's * being resilvered. That's why we check the state to see if the vdev * is healthy before reporting "resilvering" or "repairing". */ if (ps != NULL && ps->pss_state == DSS_SCANNING && children == 0 && vs->vs_state == VDEV_STATE_HEALTHY) { if (vs->vs_scan_processed != 0) { (void) printf(gettext(" (%s)"), (ps->pss_func == POOL_SCAN_RESILVER) ? "resilvering" : "repairing"); } else if (vs->vs_resilver_deferred) { (void) printf(gettext(" (awaiting resilver)")); } } /* The top-level vdevs have the rebuild stats */ if (vrs != NULL && vrs->vrs_state == VDEV_REBUILD_ACTIVE && children == 0 && vs->vs_state == VDEV_STATE_HEALTHY) { if (vs->vs_rebuild_processed != 0) { (void) printf(gettext(" (resilvering)")); } } if (cb->vcdl != NULL) { if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0) { printf(" "); zpool_print_cmd(cb->vcdl, zpool_get_name(zhp), path); } } /* Display vdev initialization and trim status for leaves. */ if (children == 0) { print_status_initialize(vs, cb->cb_print_vdev_init); print_status_trim(vs, cb->cb_print_vdev_trim); } (void) printf("\n"); for (c = 0; c < children; c++) { uint64_t islog = B_FALSE, ishole = B_FALSE; /* Don't print logs or holes here */ (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &islog); (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE, &ishole); if (islog || ishole) continue; /* Only print normal classes here */ if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS)) continue; /* Provide vdev_rebuild_stats to children if available */ if (vrs == NULL) { (void) nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i); } vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); print_status_config(zhp, cb, vname, child[c], depth + 2, isspare, vrs); free(vname); } } /* * Print the configuration of an exported pool. Iterate over all vdevs in the * pool, printing out the name and status for each one. */ static void print_import_config(status_cbdata_t *cb, const char *name, nvlist_t *nv, int depth) { nvlist_t **child; uint_t c, children; vdev_stat_t *vs; char *type, *vname; verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0); if (strcmp(type, VDEV_TYPE_MISSING) == 0 || strcmp(type, VDEV_TYPE_HOLE) == 0) return; verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c) == 0); (void) printf("\t%*s%-*s", depth, "", cb->cb_namewidth - depth, name); (void) printf(" %s", zpool_state_to_name(vs->vs_state, vs->vs_aux)); if (vs->vs_aux != 0) { (void) printf(" "); switch (vs->vs_aux) { case VDEV_AUX_OPEN_FAILED: (void) printf(gettext("cannot open")); break; case VDEV_AUX_BAD_GUID_SUM: (void) printf(gettext("missing device")); break; case VDEV_AUX_NO_REPLICAS: (void) printf(gettext("insufficient replicas")); break; case VDEV_AUX_VERSION_NEWER: (void) printf(gettext("newer version")); break; case VDEV_AUX_UNSUP_FEAT: (void) printf(gettext("unsupported feature(s)")); break; case VDEV_AUX_ERR_EXCEEDED: (void) printf(gettext("too many errors")); break; case VDEV_AUX_ACTIVE: (void) printf(gettext("currently in use")); break; case VDEV_AUX_CHILDREN_OFFLINE: (void) printf(gettext("all children offline")); break; case VDEV_AUX_BAD_LABEL: (void) printf(gettext("invalid label")); break; default: (void) printf(gettext("corrupted data")); break; } } (void) printf("\n"); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return; for (c = 0; c < children; c++) { uint64_t is_log = B_FALSE; (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log); if (is_log) continue; if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS)) continue; vname = zpool_vdev_name(g_zfs, NULL, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); print_import_config(cb, vname, child[c], depth + 2); free(vname); } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) == 0) { (void) printf(gettext("\tcache\n")); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, NULL, child[c], cb->cb_name_flags); (void) printf("\t %s\n", vname); free(vname); } } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0) { (void) printf(gettext("\tspares\n")); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, NULL, child[c], cb->cb_name_flags); (void) printf("\t %s\n", vname); free(vname); } } } /* * Print specialized class vdevs. * * These are recorded as top level vdevs in the main pool child array * but with "is_log" set to 1 or an "alloc_bias" string. We use either * print_status_config() or print_import_config() to print the top level * class vdevs then any of their children (eg mirrored slogs) are printed * recursively - which works because only the top level vdev is marked. */ static void print_class_vdevs(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t *nv, const char *class) { uint_t c, children; nvlist_t **child; boolean_t printed = B_FALSE; assert(zhp != NULL || !cb->cb_verbose); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return; for (c = 0; c < children; c++) { uint64_t is_log = B_FALSE; char *bias = NULL; char *type = NULL; (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log); if (is_log) { bias = VDEV_ALLOC_CLASS_LOGS; } else { (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &bias); (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_TYPE, &type); } if (bias == NULL || strcmp(bias, class) != 0) continue; if (!is_log && strcmp(type, VDEV_TYPE_INDIRECT) == 0) continue; if (!printed) { (void) printf("\t%s\t\n", gettext(class)); printed = B_TRUE; } char *name = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); if (cb->cb_print_status) print_status_config(zhp, cb, name, child[c], 2, B_FALSE, NULL); else print_import_config(cb, name, child[c], 2); free(name); } } /* * Display the status for the given pool. */ static int show_import(nvlist_t *config, boolean_t report_error) { uint64_t pool_state; vdev_stat_t *vs; char *name; uint64_t guid; uint64_t hostid = 0; char *msgid; char *hostname = "unknown"; nvlist_t *nvroot, *nvinfo; zpool_status_t reason; zpool_errata_t errata; const char *health; uint_t vsc; char *comment; status_cbdata_t cb = { 0 }; verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &name) == 0); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &guid) == 0); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, &pool_state) == 0); verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc) == 0); health = zpool_state_to_name(vs->vs_state, vs->vs_aux); reason = zpool_import_status(config, &msgid, &errata); /* * If we're importing using a cachefile, then we won't report any * errors unless we are in the scan phase of the import. */ if (reason != ZPOOL_STATUS_OK && !report_error) return (reason); (void) printf(gettext(" pool: %s\n"), name); (void) printf(gettext(" id: %llu\n"), (u_longlong_t)guid); (void) printf(gettext(" state: %s"), health); if (pool_state == POOL_STATE_DESTROYED) (void) printf(gettext(" (DESTROYED)")); (void) printf("\n"); switch (reason) { case ZPOOL_STATUS_MISSING_DEV_R: case ZPOOL_STATUS_MISSING_DEV_NR: case ZPOOL_STATUS_BAD_GUID_SUM: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "missing from the system.\n")); break; case ZPOOL_STATUS_CORRUPT_LABEL_R: case ZPOOL_STATUS_CORRUPT_LABEL_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices contains" " corrupted data.\n")); break; case ZPOOL_STATUS_CORRUPT_DATA: (void) printf( gettext(" status: The pool data is corrupted.\n")); break; case ZPOOL_STATUS_OFFLINE_DEV: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices " "are offlined.\n")); break; case ZPOOL_STATUS_CORRUPT_POOL: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool metadata is " "corrupted.\n")); break; case ZPOOL_STATUS_VERSION_OLDER: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is formatted using " "a legacy on-disk version.\n")); break; case ZPOOL_STATUS_VERSION_NEWER: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is formatted using " "an incompatible version.\n")); break; case ZPOOL_STATUS_FEAT_DISABLED: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Some supported " "features are not enabled on the pool.\n\t" "(Note that they may be intentionally disabled " "if the\n\t'compatibility' property is set.)\n")); break; case ZPOOL_STATUS_COMPATIBILITY_ERR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Error reading or parsing " "the file(s) indicated by the 'compatibility'\n" "property.\n")); break; case ZPOOL_STATUS_INCOMPATIBLE_FEAT: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more features " "are enabled on the pool despite not being\n" "requested by the 'compatibility' property.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_READ: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool uses the following " "feature(s) not supported on this system:\n")); color_start(ANSI_YELLOW); zpool_print_unsup_feat(config); color_end(); break; case ZPOOL_STATUS_UNSUP_FEAT_WRITE: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool can only be " "accessed in read-only mode on this system. It\n\tcannot be" " accessed in read-write mode because it uses the " "following\n\tfeature(s) not supported on this system:\n")); color_start(ANSI_YELLOW); zpool_print_unsup_feat(config); color_end(); break; case ZPOOL_STATUS_HOSTID_ACTIVE: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is currently " "imported by another system.\n")); break; case ZPOOL_STATUS_HOSTID_REQUIRED: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool has the " "multihost property on. It cannot\n\tbe safely imported " "when the system hostid is not set.\n")); break; case ZPOOL_STATUS_HOSTID_MISMATCH: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool was last accessed " "by another system.\n")); break; case ZPOOL_STATUS_FAULTED_DEV_R: case ZPOOL_STATUS_FAULTED_DEV_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "faulted.\n")); break; case ZPOOL_STATUS_BAD_LOG: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("An intent log record cannot " "be read.\n")); break; case ZPOOL_STATUS_RESILVERING: case ZPOOL_STATUS_REBUILDING: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices were " "being resilvered.\n")); break; case ZPOOL_STATUS_ERRATA: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Errata #%d detected.\n"), errata); break; case ZPOOL_STATUS_NON_NATIVE_ASHIFT: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "configured to use a non-native block size.\n" "\tExpect reduced performance.\n")); break; default: /* * No other status can be seen when importing pools. */ assert(reason == ZPOOL_STATUS_OK); } /* * Print out an action according to the overall state of the pool. */ if (vs->vs_state == VDEV_STATE_HEALTHY) { if (reason == ZPOOL_STATUS_VERSION_OLDER || reason == ZPOOL_STATUS_FEAT_DISABLED) { (void) printf(gettext(" action: The pool can be " "imported using its name or numeric identifier, " "though\n\tsome features will not be available " "without an explicit 'zpool upgrade'.\n")); } else if (reason == ZPOOL_STATUS_COMPATIBILITY_ERR) { (void) printf(gettext(" action: The pool can be " "imported using its name or numeric\n\tidentifier, " "though the file(s) indicated by its " "'compatibility'\n\tproperty cannot be parsed at " "this time.\n")); } else if (reason == ZPOOL_STATUS_HOSTID_MISMATCH) { (void) printf(gettext(" action: The pool can be " "imported using its name or numeric " "identifier and\n\tthe '-f' flag.\n")); } else if (reason == ZPOOL_STATUS_ERRATA) { switch (errata) { case ZPOOL_ERRATA_NONE: break; case ZPOOL_ERRATA_ZOL_2094_SCRUB: (void) printf(gettext(" action: The pool can " "be imported using its name or numeric " "identifier,\n\thowever there is a compat" "ibility issue which should be corrected" "\n\tby running 'zpool scrub'\n")); break; case ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY: (void) printf(gettext(" action: The pool can" "not be imported with this version of ZFS " "due to\n\tan active asynchronous destroy. " "Revert to an earlier version\n\tand " "allow the destroy to complete before " "updating.\n")); break; case ZPOOL_ERRATA_ZOL_6845_ENCRYPTION: (void) printf(gettext(" action: Existing " "encrypted datasets contain an on-disk " "incompatibility, which\n\tneeds to be " "corrected. Backup these datasets to new " "encrypted datasets\n\tand destroy the " "old ones.\n")); break; case ZPOOL_ERRATA_ZOL_8308_ENCRYPTION: (void) printf(gettext(" action: Existing " "encrypted snapshots and bookmarks contain " "an on-disk\n\tincompatibility. This may " "cause on-disk corruption if they are used" "\n\twith 'zfs recv'. To correct the " "issue, enable the bookmark_v2 feature.\n\t" "No additional action is needed if there " "are no encrypted snapshots or\n\t" "bookmarks. If preserving the encrypted " "snapshots and bookmarks is\n\trequired, " "use a non-raw send to backup and restore " "them. Alternately,\n\tthey may be removed" " to resolve the incompatibility.\n")); break; default: /* * All errata must contain an action message. */ assert(0); } } else { (void) printf(gettext(" action: The pool can be " "imported using its name or numeric " "identifier.\n")); } } else if (vs->vs_state == VDEV_STATE_DEGRADED) { (void) printf(gettext(" action: The pool can be imported " "despite missing or damaged devices. The\n\tfault " "tolerance of the pool may be compromised if imported.\n")); } else { switch (reason) { case ZPOOL_STATUS_VERSION_NEWER: (void) printf(gettext(" action: The pool cannot be " "imported. Access the pool on a system running " "newer\n\tsoftware, or recreate the pool from " "backup.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_READ: printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("The pool cannot be " "imported. Access the pool on a system that " "supports\n\tthe required feature(s), or recreate " "the pool from backup.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_WRITE: printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("The pool cannot be " "imported in read-write mode. Import the pool " "with\n" "\t\"-o readonly=on\", access the pool on a system " "that supports the\n\trequired feature(s), or " "recreate the pool from backup.\n")); break; case ZPOOL_STATUS_MISSING_DEV_R: case ZPOOL_STATUS_MISSING_DEV_NR: case ZPOOL_STATUS_BAD_GUID_SUM: (void) printf(gettext(" action: The pool cannot be " "imported. Attach the missing\n\tdevices and try " "again.\n")); break; case ZPOOL_STATUS_HOSTID_ACTIVE: VERIFY0(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, &nvinfo)); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME)) hostname = fnvlist_lookup_string(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTID)) hostid = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_HOSTID); (void) printf(gettext(" action: The pool must be " "exported from %s (hostid=%lx)\n\tbefore it " "can be safely imported.\n"), hostname, (unsigned long) hostid); break; case ZPOOL_STATUS_HOSTID_REQUIRED: (void) printf(gettext(" action: Set a unique system " "hostid with the zgenhostid(8) command.\n")); break; default: (void) printf(gettext(" action: The pool cannot be " "imported due to damaged devices or data.\n")); } } /* Print the comment attached to the pool. */ if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMMENT, &comment) == 0) (void) printf(gettext("comment: %s\n"), comment); /* * If the state is "closed" or "can't open", and the aux state * is "corrupt data": */ if (((vs->vs_state == VDEV_STATE_CLOSED) || (vs->vs_state == VDEV_STATE_CANT_OPEN)) && (vs->vs_aux == VDEV_AUX_CORRUPT_DATA)) { if (pool_state == POOL_STATE_DESTROYED) (void) printf(gettext("\tThe pool was destroyed, " "but can be imported using the '-Df' flags.\n")); else if (pool_state != POOL_STATE_EXPORTED) (void) printf(gettext("\tThe pool may be active on " "another system, but can be imported using\n\t" "the '-f' flag.\n")); } if (msgid != NULL) { (void) printf(gettext( " see: https://openzfs.github.io/openzfs-docs/msg/%s\n"), msgid); } (void) printf(gettext(" config:\n\n")); cb.cb_namewidth = max_width(NULL, nvroot, 0, strlen(name), VDEV_NAME_TYPE_ID); if (cb.cb_namewidth < 10) cb.cb_namewidth = 10; print_import_config(&cb, name, nvroot, 0); print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_BIAS_DEDUP); print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_BIAS_SPECIAL); print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_CLASS_LOGS); if (reason == ZPOOL_STATUS_BAD_GUID_SUM) { (void) printf(gettext("\n\tAdditional devices are known to " "be part of this pool, though their\n\texact " "configuration cannot be determined.\n")); } return (0); } static boolean_t zfs_force_import_required(nvlist_t *config) { uint64_t state; uint64_t hostid = 0; nvlist_t *nvinfo; state = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE); (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid); if (state != POOL_STATE_EXPORTED && hostid != get_system_hostid()) return (B_TRUE); nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_STATE)) { mmp_state_t mmp_state = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_STATE); if (mmp_state != MMP_STATE_INACTIVE) return (B_TRUE); } return (B_FALSE); } /* * Perform the import for the given configuration. This passes the heavy * lifting off to zpool_import_props(), and then mounts the datasets contained * within the pool. */ static int do_import(nvlist_t *config, const char *newname, const char *mntopts, nvlist_t *props, int flags) { int ret = 0; zpool_handle_t *zhp; char *name; uint64_t version; name = fnvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME); version = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION); if (!SPA_VERSION_IS_SUPPORTED(version)) { (void) fprintf(stderr, gettext("cannot import '%s': pool " "is formatted using an unsupported ZFS version\n"), name); return (1); } else if (zfs_force_import_required(config) && !(flags & ZFS_IMPORT_ANY_HOST)) { mmp_state_t mmp_state = MMP_STATE_INACTIVE; nvlist_t *nvinfo; nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_STATE)) mmp_state = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_STATE); if (mmp_state == MMP_STATE_ACTIVE) { char *hostname = ""; uint64_t hostid = 0; if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME)) hostname = fnvlist_lookup_string(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTID)) hostid = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_HOSTID); (void) fprintf(stderr, gettext("cannot import '%s': " "pool is imported on %s (hostid: " "0x%lx)\nExport the pool on the other system, " "then run 'zpool import'.\n"), name, hostname, (unsigned long) hostid); } else if (mmp_state == MMP_STATE_NO_HOSTID) { (void) fprintf(stderr, gettext("Cannot import '%s': " "pool has the multihost property on and the\n" "system's hostid is not set. Set a unique hostid " "with the zgenhostid(8) command.\n"), name); } else { char *hostname = ""; uint64_t timestamp = 0; uint64_t hostid = 0; if (nvlist_exists(config, ZPOOL_CONFIG_HOSTNAME)) hostname = fnvlist_lookup_string(config, ZPOOL_CONFIG_HOSTNAME); if (nvlist_exists(config, ZPOOL_CONFIG_TIMESTAMP)) timestamp = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_TIMESTAMP); if (nvlist_exists(config, ZPOOL_CONFIG_HOSTID)) hostid = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID); (void) fprintf(stderr, gettext("cannot import '%s': " "pool was previously in use from another system.\n" "Last accessed by %s (hostid=%lx) at %s" "The pool can be imported, use 'zpool import -f' " "to import the pool.\n"), name, hostname, (unsigned long)hostid, ctime((time_t *)×tamp)); } return (1); } if (zpool_import_props(g_zfs, config, newname, props, flags) != 0) return (1); if (newname != NULL) name = (char *)newname; if ((zhp = zpool_open_canfail(g_zfs, name)) == NULL) return (1); /* * Loading keys is best effort. We don't want to return immediately * if it fails but we do want to give the error to the caller. */ if (flags & ZFS_IMPORT_LOAD_KEYS) { ret = zfs_crypto_attempt_load_keys(g_zfs, name); if (ret != 0) ret = 1; } if (zpool_get_state(zhp) != POOL_STATE_UNAVAIL && !(flags & ZFS_IMPORT_ONLY) && zpool_enable_datasets(zhp, mntopts, 0) != 0) { zpool_close(zhp); return (1); } zpool_close(zhp); return (ret); } static int import_pools(nvlist_t *pools, nvlist_t *props, char *mntopts, int flags, char *orig_name, char *new_name, boolean_t do_destroyed, boolean_t pool_specified, boolean_t do_all, importargs_t *import) { nvlist_t *config = NULL; nvlist_t *found_config = NULL; uint64_t pool_state; /* * At this point we have a list of import candidate configs. Even if * we were searching by pool name or guid, we still need to * post-process the list to deal with pool state and possible * duplicate names. */ int err = 0; nvpair_t *elem = NULL; boolean_t first = B_TRUE; while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) { verify(nvpair_value_nvlist(elem, &config) == 0); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, &pool_state) == 0); if (!do_destroyed && pool_state == POOL_STATE_DESTROYED) continue; if (do_destroyed && pool_state != POOL_STATE_DESTROYED) continue; verify(nvlist_add_nvlist(config, ZPOOL_LOAD_POLICY, import->policy) == 0); if (!pool_specified) { if (first) first = B_FALSE; else if (!do_all) (void) printf("\n"); if (do_all) { err |= do_import(config, NULL, mntopts, props, flags); } else { /* * If we're importing from cachefile, then * we don't want to report errors until we * are in the scan phase of the import. If * we get an error, then we return that error * to invoke the scan phase. */ if (import->cachefile && !import->scan) err = show_import(config, B_FALSE); else (void) show_import(config, B_TRUE); } } else if (import->poolname != NULL) { char *name; /* * We are searching for a pool based on name. */ verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &name) == 0); if (strcmp(name, import->poolname) == 0) { if (found_config != NULL) { (void) fprintf(stderr, gettext( "cannot import '%s': more than " "one matching pool\n"), import->poolname); (void) fprintf(stderr, gettext( "import by numeric ID instead\n")); err = B_TRUE; } found_config = config; } } else { uint64_t guid; /* * Search for a pool by guid. */ verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &guid) == 0); if (guid == import->guid) found_config = config; } } /* * If we were searching for a specific pool, verify that we found a * pool, and then do the import. */ if (pool_specified && err == 0) { if (found_config == NULL) { (void) fprintf(stderr, gettext("cannot import '%s': " "no such pool available\n"), orig_name); err = B_TRUE; } else { err |= do_import(found_config, new_name, mntopts, props, flags); } } /* * If we were just looking for pools, report an error if none were * found. */ if (!pool_specified && first) (void) fprintf(stderr, gettext("no pools available to import\n")); return (err); } typedef struct target_exists_args { const char *poolname; uint64_t poolguid; } target_exists_args_t; static int name_or_guid_exists(zpool_handle_t *zhp, void *data) { target_exists_args_t *args = data; nvlist_t *config = zpool_get_config(zhp, NULL); int found = 0; if (config == NULL) return (0); if (args->poolname != NULL) { char *pool_name; verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &pool_name) == 0); if (strcmp(pool_name, args->poolname) == 0) found = 1; } else { uint64_t pool_guid; verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &pool_guid) == 0); if (pool_guid == args->poolguid) found = 1; } zpool_close(zhp); return (found); } /* * zpool checkpoint * checkpoint --discard * * -d Discard the checkpoint from a checkpointed * --discard pool. * * -w Wait for discarding a checkpoint to complete. * --wait * * Checkpoints the specified pool, by taking a "snapshot" of its * current state. A pool can only have one checkpoint at a time. */ int zpool_do_checkpoint(int argc, char **argv) { boolean_t discard, wait; char *pool; zpool_handle_t *zhp; int c, err; struct option long_options[] = { {"discard", no_argument, NULL, 'd'}, {"wait", no_argument, NULL, 'w'}, {0, 0, 0, 0} }; discard = B_FALSE; wait = B_FALSE; while ((c = getopt_long(argc, argv, ":dw", long_options, NULL)) != -1) { switch (c) { case 'd': discard = B_TRUE; break; case 'w': wait = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } if (wait && !discard) { (void) fprintf(stderr, gettext("--wait only valid when " "--discard also specified\n")); usage(B_FALSE); } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool argument\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } pool = argv[0]; if ((zhp = zpool_open(g_zfs, pool)) == NULL) { /* As a special case, check for use of '/' in the name */ if (strchr(pool, '/') != NULL) (void) fprintf(stderr, gettext("'zpool checkpoint' " "doesn't work on datasets. To save the state " "of a dataset from a specific point in time " "please use 'zfs snapshot'\n")); return (1); } if (discard) { err = (zpool_discard_checkpoint(zhp) != 0); if (err == 0 && wait) err = zpool_wait(zhp, ZPOOL_WAIT_CKPT_DISCARD); } else { err = (zpool_checkpoint(zhp) != 0); } zpool_close(zhp); return (err); } #define CHECKPOINT_OPT 1024 /* * zpool import [-d dir] [-D] * import [-o mntopts] [-o prop=value] ... [-R root] [-D] [-l] * [-d dir | -c cachefile | -s] [-f] -a * import [-o mntopts] [-o prop=value] ... [-R root] [-D] [-l] * [-d dir | -c cachefile | -s] [-f] [-n] [-F] * [newpool] * * -c Read pool information from a cachefile instead of searching * devices. If importing from a cachefile config fails, then * fallback to searching for devices only in the directories that * exist in the cachefile. * * -d Scan in a specific directory, other than /dev/. More than * one directory can be specified using multiple '-d' options. * * -D Scan for previously destroyed pools or import all or only * specified destroyed pools. * * -R Temporarily import the pool, with all mountpoints relative to * the given root. The pool will remain exported when the machine * is rebooted. * * -V Import even in the presence of faulted vdevs. This is an * intentionally undocumented option for testing purposes, and * treats the pool configuration as complete, leaving any bad * vdevs in the FAULTED state. In other words, it does verbatim * import. * * -f Force import, even if it appears that the pool is active. * * -F Attempt rewind if necessary. * * -n See if rewind would work, but don't actually rewind. * * -N Import the pool but don't mount datasets. * * -T Specify a starting txg to use for import. This option is * intentionally undocumented option for testing purposes. * * -a Import all pools found. * * -l Load encryption keys while importing. * * -o Set property=value and/or temporary mount options (without '='). * * -s Scan using the default search path, the libblkid cache will * not be consulted. * * --rewind-to-checkpoint * Import the pool and revert back to the checkpoint. * * The import command scans for pools to import, and import pools based on pool * name and GUID. The pool can also be renamed as part of the import process. */ int zpool_do_import(int argc, char **argv) { char **searchdirs = NULL; char *env, *envdup = NULL; int nsearch = 0; int c; int err = 0; nvlist_t *pools = NULL; boolean_t do_all = B_FALSE; boolean_t do_destroyed = B_FALSE; char *mntopts = NULL; uint64_t searchguid = 0; char *searchname = NULL; char *propval; nvlist_t *policy = NULL; nvlist_t *props = NULL; int flags = ZFS_IMPORT_NORMAL; uint32_t rewind_policy = ZPOOL_NO_REWIND; boolean_t dryrun = B_FALSE; boolean_t do_rewind = B_FALSE; boolean_t xtreme_rewind = B_FALSE; boolean_t do_scan = B_FALSE; boolean_t pool_exists = B_FALSE; boolean_t pool_specified = B_FALSE; uint64_t txg = -1ULL; char *cachefile = NULL; importargs_t idata = { 0 }; char *endptr; struct option long_options[] = { {"rewind-to-checkpoint", no_argument, NULL, CHECKPOINT_OPT}, {0, 0, 0, 0} }; /* check options */ while ((c = getopt_long(argc, argv, ":aCc:d:DEfFlmnNo:R:stT:VX", long_options, NULL)) != -1) { switch (c) { case 'a': do_all = B_TRUE; break; case 'c': cachefile = optarg; break; case 'd': if (searchdirs == NULL) { searchdirs = safe_malloc(sizeof (char *)); } else { char **tmp = safe_malloc((nsearch + 1) * sizeof (char *)); bcopy(searchdirs, tmp, nsearch * sizeof (char *)); free(searchdirs); searchdirs = tmp; } searchdirs[nsearch++] = optarg; break; case 'D': do_destroyed = B_TRUE; break; case 'f': flags |= ZFS_IMPORT_ANY_HOST; break; case 'F': do_rewind = B_TRUE; break; case 'l': flags |= ZFS_IMPORT_LOAD_KEYS; break; case 'm': flags |= ZFS_IMPORT_MISSING_LOG; break; case 'n': dryrun = B_TRUE; break; case 'N': flags |= ZFS_IMPORT_ONLY; break; case 'o': if ((propval = strchr(optarg, '=')) != NULL) { *propval = '\0'; propval++; if (add_prop_list(optarg, propval, &props, B_TRUE)) goto error; } else { mntopts = optarg; } break; case 'R': if (add_prop_list(zpool_prop_to_name( ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE)) goto error; if (add_prop_list_default(zpool_prop_to_name( ZPOOL_PROP_CACHEFILE), "none", &props, B_TRUE)) goto error; break; case 's': do_scan = B_TRUE; break; case 't': flags |= ZFS_IMPORT_TEMP_NAME; if (add_prop_list_default(zpool_prop_to_name( ZPOOL_PROP_CACHEFILE), "none", &props, B_TRUE)) goto error; break; case 'T': errno = 0; txg = strtoull(optarg, &endptr, 0); if (errno != 0 || *endptr != '\0') { (void) fprintf(stderr, gettext("invalid txg value\n")); usage(B_FALSE); } rewind_policy = ZPOOL_DO_REWIND | ZPOOL_EXTREME_REWIND; break; case 'V': flags |= ZFS_IMPORT_VERBATIM; break; case 'X': xtreme_rewind = B_TRUE; break; case CHECKPOINT_OPT: flags |= ZFS_IMPORT_CHECKPOINT; break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); usage(B_FALSE); break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (cachefile && nsearch != 0) { (void) fprintf(stderr, gettext("-c is incompatible with -d\n")); usage(B_FALSE); } if (cachefile && do_scan) { (void) fprintf(stderr, gettext("-c is incompatible with -s\n")); usage(B_FALSE); } if ((flags & ZFS_IMPORT_LOAD_KEYS) && (flags & ZFS_IMPORT_ONLY)) { (void) fprintf(stderr, gettext("-l is incompatible with -N\n")); usage(B_FALSE); } if ((flags & ZFS_IMPORT_LOAD_KEYS) && !do_all && argc == 0) { (void) fprintf(stderr, gettext("-l is only meaningful during " "an import\n")); usage(B_FALSE); } if ((dryrun || xtreme_rewind) && !do_rewind) { (void) fprintf(stderr, gettext("-n or -X only meaningful with -F\n")); usage(B_FALSE); } if (dryrun) rewind_policy = ZPOOL_TRY_REWIND; else if (do_rewind) rewind_policy = ZPOOL_DO_REWIND; if (xtreme_rewind) rewind_policy |= ZPOOL_EXTREME_REWIND; /* In the future, we can capture further policy and include it here */ if (nvlist_alloc(&policy, NV_UNIQUE_NAME, 0) != 0 || nvlist_add_uint64(policy, ZPOOL_LOAD_REQUEST_TXG, txg) != 0 || nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY, rewind_policy) != 0) goto error; /* check argument count */ if (do_all) { if (argc != 0) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } } else { if (argc > 2) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } } /* * Check for the effective uid. We do this explicitly here because * otherwise any attempt to discover pools will silently fail. */ if (argc == 0 && geteuid() != 0) { (void) fprintf(stderr, gettext("cannot " "discover pools: permission denied\n")); if (searchdirs != NULL) free(searchdirs); nvlist_free(props); nvlist_free(policy); return (1); } /* * Depending on the arguments given, we do one of the following: * * Iterate through all pools and display information about * each one. * * -a Iterate through all pools and try to import each one. * * Find the pool that corresponds to the given GUID/pool * name and import that one. * * -D Above options applies only to destroyed pools. */ if (argc != 0) { char *endptr; errno = 0; searchguid = strtoull(argv[0], &endptr, 10); if (errno != 0 || *endptr != '\0') { searchname = argv[0]; searchguid = 0; } pool_specified = B_TRUE; /* * User specified a name or guid. Ensure it's unique. */ target_exists_args_t search = {searchname, searchguid}; pool_exists = zpool_iter(g_zfs, name_or_guid_exists, &search); } /* * Check the environment for the preferred search path. */ if ((searchdirs == NULL) && (env = getenv("ZPOOL_IMPORT_PATH"))) { char *dir; envdup = strdup(env); dir = strtok(envdup, ":"); while (dir != NULL) { if (searchdirs == NULL) { searchdirs = safe_malloc(sizeof (char *)); } else { char **tmp = safe_malloc((nsearch + 1) * sizeof (char *)); bcopy(searchdirs, tmp, nsearch * sizeof (char *)); free(searchdirs); searchdirs = tmp; } searchdirs[nsearch++] = dir; dir = strtok(NULL, ":"); } } idata.path = searchdirs; idata.paths = nsearch; idata.poolname = searchname; idata.guid = searchguid; idata.cachefile = cachefile; idata.scan = do_scan; idata.policy = policy; pools = zpool_search_import(g_zfs, &idata, &libzfs_config_ops); if (pools != NULL && pool_exists && (argc == 1 || strcmp(argv[0], argv[1]) == 0)) { (void) fprintf(stderr, gettext("cannot import '%s': " "a pool with that name already exists\n"), argv[0]); (void) fprintf(stderr, gettext("use the form '%s " " ' to give it a new name\n"), "zpool import"); err = 1; } else if (pools == NULL && pool_exists) { (void) fprintf(stderr, gettext("cannot import '%s': " "a pool with that name is already created/imported,\n"), argv[0]); (void) fprintf(stderr, gettext("and no additional pools " "with that name were found\n")); err = 1; } else if (pools == NULL) { if (argc != 0) { (void) fprintf(stderr, gettext("cannot import '%s': " "no such pool available\n"), argv[0]); } err = 1; } if (err == 1) { if (searchdirs != NULL) free(searchdirs); if (envdup != NULL) free(envdup); nvlist_free(policy); nvlist_free(pools); nvlist_free(props); return (1); } err = import_pools(pools, props, mntopts, flags, argc >= 1 ? argv[0] : NULL, argc >= 2 ? argv[1] : NULL, do_destroyed, pool_specified, do_all, &idata); /* * If we're using the cachefile and we failed to import, then * fallback to scanning the directory for pools that match * those in the cachefile. */ if (err != 0 && cachefile != NULL) { (void) printf(gettext("cachefile import failed, retrying\n")); /* * We use the scan flag to gather the directories that exist * in the cachefile. If we need to fallback to searching for * the pool config, we will only search devices in these * directories. */ idata.scan = B_TRUE; nvlist_free(pools); pools = zpool_search_import(g_zfs, &idata, &libzfs_config_ops); err = import_pools(pools, props, mntopts, flags, argc >= 1 ? argv[0] : NULL, argc >= 2 ? argv[1] : NULL, do_destroyed, pool_specified, do_all, &idata); } error: nvlist_free(props); nvlist_free(pools); nvlist_free(policy); if (searchdirs != NULL) free(searchdirs); if (envdup != NULL) free(envdup); return (err ? 1 : 0); } /* * zpool sync [-f] [pool] ... * * -f (undocumented) force uberblock (and config including zpool cache file) * update. * * Sync the specified pool(s). * Without arguments "zpool sync" will sync all pools. * This command initiates TXG sync(s) and will return after the TXG(s) commit. * */ static int zpool_do_sync(int argc, char **argv) { int ret; boolean_t force = B_FALSE; /* check options */ while ((ret = getopt(argc, argv, "f")) != -1) { switch (ret) { case 'f': force = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* if argc == 0 we will execute zpool_sync_one on all pools */ ret = for_each_pool(argc, argv, B_FALSE, NULL, B_FALSE, zpool_sync_one, &force); return (ret); } typedef struct iostat_cbdata { uint64_t cb_flags; int cb_name_flags; int cb_namewidth; int cb_iteration; char **cb_vdev_names; /* Only show these vdevs */ unsigned int cb_vdev_names_count; boolean_t cb_verbose; boolean_t cb_literal; boolean_t cb_scripted; zpool_list_t *cb_list; vdev_cmd_data_list_t *vcdl; } iostat_cbdata_t; /* iostat labels */ typedef struct name_and_columns { const char *name; /* Column name */ unsigned int columns; /* Center name to this number of columns */ } name_and_columns_t; #define IOSTAT_MAX_LABELS 13 /* Max number of labels on one line */ static const name_and_columns_t iostat_top_labels[][IOSTAT_MAX_LABELS] = { [IOS_DEFAULT] = {{"capacity", 2}, {"operations", 2}, {"bandwidth", 2}, {NULL}}, [IOS_LATENCY] = {{"total_wait", 2}, {"disk_wait", 2}, {"syncq_wait", 2}, {"asyncq_wait", 2}, {"scrub", 1}, {"trim", 1}, {NULL}}, [IOS_QUEUES] = {{"syncq_read", 2}, {"syncq_write", 2}, {"asyncq_read", 2}, {"asyncq_write", 2}, {"scrubq_read", 2}, {"trimq_write", 2}, {NULL}}, [IOS_L_HISTO] = {{"total_wait", 2}, {"disk_wait", 2}, {"syncq_wait", 2}, {"asyncq_wait", 2}, {NULL}}, [IOS_RQ_HISTO] = {{"sync_read", 2}, {"sync_write", 2}, {"async_read", 2}, {"async_write", 2}, {"scrub", 2}, {"trim", 2}, {NULL}}, }; /* Shorthand - if "columns" field not set, default to 1 column */ static const name_and_columns_t iostat_bottom_labels[][IOSTAT_MAX_LABELS] = { [IOS_DEFAULT] = {{"alloc"}, {"free"}, {"read"}, {"write"}, {"read"}, {"write"}, {NULL}}, [IOS_LATENCY] = {{"read"}, {"write"}, {"read"}, {"write"}, {"read"}, {"write"}, {"read"}, {"write"}, {"wait"}, {"wait"}, {NULL}}, [IOS_QUEUES] = {{"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"}, {NULL}}, [IOS_L_HISTO] = {{"read"}, {"write"}, {"read"}, {"write"}, {"read"}, {"write"}, {"read"}, {"write"}, {"scrub"}, {"trim"}, {NULL}}, [IOS_RQ_HISTO] = {{"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"}, {NULL}}, }; static const char *histo_to_title[] = { [IOS_L_HISTO] = "latency", [IOS_RQ_HISTO] = "req_size", }; /* * Return the number of labels in a null-terminated name_and_columns_t * array. * */ static unsigned int label_array_len(const name_and_columns_t *labels) { int i = 0; while (labels[i].name) i++; return (i); } /* * Return the number of strings in a null-terminated string array. * For example: * * const char foo[] = {"bar", "baz", NULL} * * returns 2 */ static uint64_t str_array_len(const char *array[]) { uint64_t i = 0; while (array[i]) i++; return (i); } /* * Return a default column width for default/latency/queue columns. This does * not include histograms, which have their columns autosized. */ static unsigned int default_column_width(iostat_cbdata_t *cb, enum iostat_type type) { unsigned long column_width = 5; /* Normal niceprint */ static unsigned long widths[] = { /* * Choose some sane default column sizes for printing the * raw numbers. */ [IOS_DEFAULT] = 15, /* 1PB capacity */ [IOS_LATENCY] = 10, /* 1B ns = 10sec */ [IOS_QUEUES] = 6, /* 1M queue entries */ [IOS_L_HISTO] = 10, /* 1B ns = 10sec */ [IOS_RQ_HISTO] = 6, /* 1M queue entries */ }; if (cb->cb_literal) column_width = widths[type]; return (column_width); } /* * Print the column labels, i.e: * * capacity operations bandwidth * alloc free read write read write ... * * If force_column_width is set, use it for the column width. If not set, use * the default column width. */ static void print_iostat_labels(iostat_cbdata_t *cb, unsigned int force_column_width, const name_and_columns_t labels[][IOSTAT_MAX_LABELS]) { int i, idx, s; int text_start, rw_column_width, spaces_to_end; uint64_t flags = cb->cb_flags; uint64_t f; unsigned int column_width = force_column_width; /* For each bit set in flags */ for (f = flags; f; f &= ~(1ULL << idx)) { idx = lowbit64(f) - 1; if (!force_column_width) column_width = default_column_width(cb, idx); /* Print our top labels centered over "read write" label. */ for (i = 0; i < label_array_len(labels[idx]); i++) { const char *name = labels[idx][i].name; /* * We treat labels[][].columns == 0 as shorthand * for one column. It makes writing out the label * tables more concise. */ unsigned int columns = MAX(1, labels[idx][i].columns); unsigned int slen = strlen(name); rw_column_width = (column_width * columns) + (2 * (columns - 1)); text_start = (int)((rw_column_width) / columns - slen / columns); if (text_start < 0) text_start = 0; printf(" "); /* Two spaces between columns */ /* Space from beginning of column to label */ for (s = 0; s < text_start; s++) printf(" "); printf("%s", name); /* Print space after label to end of column */ spaces_to_end = rw_column_width - text_start - slen; if (spaces_to_end < 0) spaces_to_end = 0; for (s = 0; s < spaces_to_end; s++) printf(" "); } } } /* * print_cmd_columns - Print custom column titles from -c * * If the user specified the "zpool status|iostat -c" then print their custom * column titles in the header. For example, print_cmd_columns() would print * the " col1 col2" part of this: * * $ zpool iostat -vc 'echo col1=val1; echo col2=val2' * ... * capacity operations bandwidth * pool alloc free read write read write col1 col2 * ---------- ----- ----- ----- ----- ----- ----- ---- ---- * mypool 269K 1008M 0 0 107 946 * mirror 269K 1008M 0 0 107 946 * sdb - - 0 0 102 473 val1 val2 * sdc - - 0 0 5 473 val1 val2 * ---------- ----- ----- ----- ----- ----- ----- ---- ---- */ static void print_cmd_columns(vdev_cmd_data_list_t *vcdl, int use_dashes) { int i, j; vdev_cmd_data_t *data = &vcdl->data[0]; if (vcdl->count == 0 || data == NULL) return; /* * Each vdev cmd should have the same column names unless the user did * something weird with their cmd. Just take the column names from the * first vdev and assume it works for all of them. */ for (i = 0; i < vcdl->uniq_cols_cnt; i++) { printf(" "); if (use_dashes) { for (j = 0; j < vcdl->uniq_cols_width[i]; j++) printf("-"); } else { printf_color(ANSI_BOLD, "%*s", vcdl->uniq_cols_width[i], vcdl->uniq_cols[i]); } } } /* * Utility function to print out a line of dashes like: * * -------------------------------- ----- ----- ----- ----- ----- * * ...or a dashed named-row line like: * * logs - - - - - * * @cb: iostat data * * @force_column_width If non-zero, use the value as the column width. * Otherwise use the default column widths. * * @name: Print a dashed named-row line starting * with @name. Otherwise, print a regular * dashed line. */ static void print_iostat_dashes(iostat_cbdata_t *cb, unsigned int force_column_width, const char *name) { int i; unsigned int namewidth; uint64_t flags = cb->cb_flags; uint64_t f; int idx; const name_and_columns_t *labels; const char *title; if (cb->cb_flags & IOS_ANYHISTO_M) { title = histo_to_title[IOS_HISTO_IDX(cb->cb_flags)]; } else if (cb->cb_vdev_names_count) { title = "vdev"; } else { title = "pool"; } namewidth = MAX(MAX(strlen(title), cb->cb_namewidth), name ? strlen(name) : 0); if (name) { printf("%-*s", namewidth, name); } else { for (i = 0; i < namewidth; i++) (void) printf("-"); } /* For each bit in flags */ for (f = flags; f; f &= ~(1ULL << idx)) { unsigned int column_width; idx = lowbit64(f) - 1; if (force_column_width) column_width = force_column_width; else column_width = default_column_width(cb, idx); labels = iostat_bottom_labels[idx]; for (i = 0; i < label_array_len(labels); i++) { if (name) printf(" %*s-", column_width - 1, " "); else printf(" %.*s", column_width, "--------------------"); } } } static void print_iostat_separator_impl(iostat_cbdata_t *cb, unsigned int force_column_width) { print_iostat_dashes(cb, force_column_width, NULL); } static void print_iostat_separator(iostat_cbdata_t *cb) { print_iostat_separator_impl(cb, 0); } static void print_iostat_header_impl(iostat_cbdata_t *cb, unsigned int force_column_width, const char *histo_vdev_name) { unsigned int namewidth; const char *title; color_start(ANSI_BOLD); if (cb->cb_flags & IOS_ANYHISTO_M) { title = histo_to_title[IOS_HISTO_IDX(cb->cb_flags)]; } else if (cb->cb_vdev_names_count) { title = "vdev"; } else { title = "pool"; } namewidth = MAX(MAX(strlen(title), cb->cb_namewidth), histo_vdev_name ? strlen(histo_vdev_name) : 0); if (histo_vdev_name) printf("%-*s", namewidth, histo_vdev_name); else printf("%*s", namewidth, ""); print_iostat_labels(cb, force_column_width, iostat_top_labels); printf("\n"); printf("%-*s", namewidth, title); print_iostat_labels(cb, force_column_width, iostat_bottom_labels); if (cb->vcdl != NULL) print_cmd_columns(cb->vcdl, 0); printf("\n"); print_iostat_separator_impl(cb, force_column_width); if (cb->vcdl != NULL) print_cmd_columns(cb->vcdl, 1); color_end(); printf("\n"); } static void print_iostat_header(iostat_cbdata_t *cb) { print_iostat_header_impl(cb, 0, NULL); } /* * Prints a size string (i.e. 120M) with the suffix ("M") colored * by order of magnitude. Uses column_size to add padding. */ static void print_stat_color(const char *statbuf, unsigned int column_size) { fputs(" ", stdout); size_t len = strlen(statbuf); while (len < column_size) { fputc(' ', stdout); column_size--; } if (*statbuf == '0') { color_start(ANSI_GRAY); fputc('0', stdout); } else { for (; *statbuf; statbuf++) { if (*statbuf == 'K') color_start(ANSI_GREEN); else if (*statbuf == 'M') color_start(ANSI_YELLOW); else if (*statbuf == 'G') color_start(ANSI_RED); else if (*statbuf == 'T') color_start(ANSI_BOLD_BLUE); else if (*statbuf == 'P') color_start(ANSI_MAGENTA); else if (*statbuf == 'E') color_start(ANSI_CYAN); fputc(*statbuf, stdout); if (--column_size <= 0) break; } } color_end(); } /* * Display a single statistic. */ static void print_one_stat(uint64_t value, enum zfs_nicenum_format format, unsigned int column_size, boolean_t scripted) { char buf[64]; zfs_nicenum_format(value, buf, sizeof (buf), format); if (scripted) printf("\t%s", buf); else print_stat_color(buf, column_size); } /* * Calculate the default vdev stats * * Subtract oldvs from newvs, apply a scaling factor, and save the resulting * stats into calcvs. */ static void calc_default_iostats(vdev_stat_t *oldvs, vdev_stat_t *newvs, vdev_stat_t *calcvs) { int i; memcpy(calcvs, newvs, sizeof (*calcvs)); for (i = 0; i < ARRAY_SIZE(calcvs->vs_ops); i++) calcvs->vs_ops[i] = (newvs->vs_ops[i] - oldvs->vs_ops[i]); for (i = 0; i < ARRAY_SIZE(calcvs->vs_bytes); i++) calcvs->vs_bytes[i] = (newvs->vs_bytes[i] - oldvs->vs_bytes[i]); } /* * Internal representation of the extended iostats data. * * The extended iostat stats are exported in nvlists as either uint64_t arrays * or single uint64_t's. We make both look like arrays to make them easier * to process. In order to make single uint64_t's look like arrays, we set * __data to the stat data, and then set *data = &__data with count = 1. Then, * we can just use *data and count. */ struct stat_array { uint64_t *data; uint_t count; /* Number of entries in data[] */ uint64_t __data; /* Only used when data is a single uint64_t */ }; static uint64_t stat_histo_max(struct stat_array *nva, unsigned int len) { uint64_t max = 0; int i; for (i = 0; i < len; i++) max = MAX(max, array64_max(nva[i].data, nva[i].count)); return (max); } /* * Helper function to lookup a uint64_t array or uint64_t value and store its * data as a stat_array. If the nvpair is a single uint64_t value, then we make * it look like a one element array to make it easier to process. */ static int nvpair64_to_stat_array(nvlist_t *nvl, const char *name, struct stat_array *nva) { nvpair_t *tmp; int ret; verify(nvlist_lookup_nvpair(nvl, name, &tmp) == 0); switch (nvpair_type(tmp)) { case DATA_TYPE_UINT64_ARRAY: ret = nvpair_value_uint64_array(tmp, &nva->data, &nva->count); break; case DATA_TYPE_UINT64: ret = nvpair_value_uint64(tmp, &nva->__data); nva->data = &nva->__data; nva->count = 1; break; default: /* Not a uint64_t */ ret = EINVAL; break; } return (ret); } /* * Given a list of nvlist names, look up the extended stats in newnv and oldnv, * subtract them, and return the results in a newly allocated stat_array. * You must free the returned array after you are done with it with * free_calc_stats(). * * Additionally, you can set "oldnv" to NULL if you simply want the newnv * values. */ static struct stat_array * calc_and_alloc_stats_ex(const char **names, unsigned int len, nvlist_t *oldnv, nvlist_t *newnv) { nvlist_t *oldnvx = NULL, *newnvx; struct stat_array *oldnva, *newnva, *calcnva; int i, j; unsigned int alloc_size = (sizeof (struct stat_array)) * len; /* Extract our extended stats nvlist from the main list */ verify(nvlist_lookup_nvlist(newnv, ZPOOL_CONFIG_VDEV_STATS_EX, &newnvx) == 0); if (oldnv) { verify(nvlist_lookup_nvlist(oldnv, ZPOOL_CONFIG_VDEV_STATS_EX, &oldnvx) == 0); } newnva = safe_malloc(alloc_size); oldnva = safe_malloc(alloc_size); calcnva = safe_malloc(alloc_size); for (j = 0; j < len; j++) { verify(nvpair64_to_stat_array(newnvx, names[j], &newnva[j]) == 0); calcnva[j].count = newnva[j].count; alloc_size = calcnva[j].count * sizeof (calcnva[j].data[0]); calcnva[j].data = safe_malloc(alloc_size); memcpy(calcnva[j].data, newnva[j].data, alloc_size); if (oldnvx) { verify(nvpair64_to_stat_array(oldnvx, names[j], &oldnva[j]) == 0); for (i = 0; i < oldnva[j].count; i++) calcnva[j].data[i] -= oldnva[j].data[i]; } } free(newnva); free(oldnva); return (calcnva); } static void free_calc_stats(struct stat_array *nva, unsigned int len) { int i; for (i = 0; i < len; i++) free(nva[i].data); free(nva); } static void print_iostat_histo(struct stat_array *nva, unsigned int len, iostat_cbdata_t *cb, unsigned int column_width, unsigned int namewidth, double scale) { int i, j; char buf[6]; uint64_t val; enum zfs_nicenum_format format; unsigned int buckets; unsigned int start_bucket; if (cb->cb_literal) format = ZFS_NICENUM_RAW; else format = ZFS_NICENUM_1024; /* All these histos are the same size, so just use nva[0].count */ buckets = nva[0].count; if (cb->cb_flags & IOS_RQ_HISTO_M) { /* Start at 512 - req size should never be lower than this */ start_bucket = 9; } else { start_bucket = 0; } for (j = start_bucket; j < buckets; j++) { /* Print histogram bucket label */ if (cb->cb_flags & IOS_L_HISTO_M) { /* Ending range of this bucket */ val = (1UL << (j + 1)) - 1; zfs_nicetime(val, buf, sizeof (buf)); } else { /* Request size (starting range of bucket) */ val = (1UL << j); zfs_nicenum(val, buf, sizeof (buf)); } if (cb->cb_scripted) printf("%llu", (u_longlong_t)val); else printf("%-*s", namewidth, buf); /* Print the values on the line */ for (i = 0; i < len; i++) { print_one_stat(nva[i].data[j] * scale, format, column_width, cb->cb_scripted); } printf("\n"); } } static void print_solid_separator(unsigned int length) { while (length--) printf("-"); printf("\n"); } static void print_iostat_histos(iostat_cbdata_t *cb, nvlist_t *oldnv, nvlist_t *newnv, double scale, const char *name) { unsigned int column_width; unsigned int namewidth; unsigned int entire_width; enum iostat_type type; struct stat_array *nva; const char **names; unsigned int names_len; /* What type of histo are we? */ type = IOS_HISTO_IDX(cb->cb_flags); /* Get NULL-terminated array of nvlist names for our histo */ names = vsx_type_to_nvlist[type]; names_len = str_array_len(names); /* num of names */ nva = calc_and_alloc_stats_ex(names, names_len, oldnv, newnv); if (cb->cb_literal) { column_width = MAX(5, (unsigned int) log10(stat_histo_max(nva, names_len)) + 1); } else { column_width = 5; } namewidth = MAX(cb->cb_namewidth, strlen(histo_to_title[IOS_HISTO_IDX(cb->cb_flags)])); /* * Calculate the entire line width of what we're printing. The * +2 is for the two spaces between columns: */ /* read write */ /* ----- ----- */ /* |___| <---------- column_width */ /* */ /* |__________| <--- entire_width */ /* */ entire_width = namewidth + (column_width + 2) * label_array_len(iostat_bottom_labels[type]); if (cb->cb_scripted) printf("%s\n", name); else print_iostat_header_impl(cb, column_width, name); print_iostat_histo(nva, names_len, cb, column_width, namewidth, scale); free_calc_stats(nva, names_len); if (!cb->cb_scripted) print_solid_separator(entire_width); } /* * Calculate the average latency of a power-of-two latency histogram */ static uint64_t single_histo_average(uint64_t *histo, unsigned int buckets) { int i; uint64_t count = 0, total = 0; for (i = 0; i < buckets; i++) { /* * Our buckets are power-of-two latency ranges. Use the * midpoint latency of each bucket to calculate the average. * For example: * * Bucket Midpoint * 8ns-15ns: 12ns * 16ns-31ns: 24ns * ... */ if (histo[i] != 0) { total += histo[i] * (((1UL << i) + ((1UL << i)/2))); count += histo[i]; } } /* Prevent divide by zero */ return (count == 0 ? 0 : total / count); } static void print_iostat_queues(iostat_cbdata_t *cb, nvlist_t *oldnv, nvlist_t *newnv) { int i; uint64_t val; const char *names[] = { ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE, ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE, ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE, ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_TRIM_PEND_QUEUE, ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE, }; struct stat_array *nva; unsigned int column_width = default_column_width(cb, IOS_QUEUES); enum zfs_nicenum_format format; nva = calc_and_alloc_stats_ex(names, ARRAY_SIZE(names), NULL, newnv); if (cb->cb_literal) format = ZFS_NICENUM_RAW; else format = ZFS_NICENUM_1024; for (i = 0; i < ARRAY_SIZE(names); i++) { val = nva[i].data[0]; print_one_stat(val, format, column_width, cb->cb_scripted); } free_calc_stats(nva, ARRAY_SIZE(names)); } static void print_iostat_latency(iostat_cbdata_t *cb, nvlist_t *oldnv, nvlist_t *newnv) { int i; uint64_t val; const char *names[] = { ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO, ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO, }; struct stat_array *nva; unsigned int column_width = default_column_width(cb, IOS_LATENCY); enum zfs_nicenum_format format; nva = calc_and_alloc_stats_ex(names, ARRAY_SIZE(names), oldnv, newnv); if (cb->cb_literal) format = ZFS_NICENUM_RAWTIME; else format = ZFS_NICENUM_TIME; /* Print our avg latencies on the line */ for (i = 0; i < ARRAY_SIZE(names); i++) { /* Compute average latency for a latency histo */ val = single_histo_average(nva[i].data, nva[i].count); print_one_stat(val, format, column_width, cb->cb_scripted); } free_calc_stats(nva, ARRAY_SIZE(names)); } /* * Print default statistics (capacity/operations/bandwidth) */ static void print_iostat_default(vdev_stat_t *vs, iostat_cbdata_t *cb, double scale) { unsigned int column_width = default_column_width(cb, IOS_DEFAULT); enum zfs_nicenum_format format; char na; /* char to print for "not applicable" values */ if (cb->cb_literal) { format = ZFS_NICENUM_RAW; na = '0'; } else { format = ZFS_NICENUM_1024; na = '-'; } /* only toplevel vdevs have capacity stats */ if (vs->vs_space == 0) { if (cb->cb_scripted) printf("\t%c\t%c", na, na); else printf(" %*c %*c", column_width, na, column_width, na); } else { print_one_stat(vs->vs_alloc, format, column_width, cb->cb_scripted); print_one_stat(vs->vs_space - vs->vs_alloc, format, column_width, cb->cb_scripted); } print_one_stat((uint64_t)(vs->vs_ops[ZIO_TYPE_READ] * scale), format, column_width, cb->cb_scripted); print_one_stat((uint64_t)(vs->vs_ops[ZIO_TYPE_WRITE] * scale), format, column_width, cb->cb_scripted); print_one_stat((uint64_t)(vs->vs_bytes[ZIO_TYPE_READ] * scale), format, column_width, cb->cb_scripted); print_one_stat((uint64_t)(vs->vs_bytes[ZIO_TYPE_WRITE] * scale), format, column_width, cb->cb_scripted); } static const char *class_name[] = { VDEV_ALLOC_BIAS_DEDUP, VDEV_ALLOC_BIAS_SPECIAL, VDEV_ALLOC_CLASS_LOGS }; /* * Print out all the statistics for the given vdev. This can either be the * toplevel configuration, or called recursively. If 'name' is NULL, then this * is a verbose output, and we don't want to display the toplevel pool stats. * * Returns the number of stat lines printed. */ static unsigned int print_vdev_stats(zpool_handle_t *zhp, const char *name, nvlist_t *oldnv, nvlist_t *newnv, iostat_cbdata_t *cb, int depth) { nvlist_t **oldchild, **newchild; uint_t c, children, oldchildren; vdev_stat_t *oldvs, *newvs, *calcvs; vdev_stat_t zerovs = { 0 }; char *vname; int i; int ret = 0; uint64_t tdelta; double scale; if (strcmp(name, VDEV_TYPE_INDIRECT) == 0) return (ret); calcvs = safe_malloc(sizeof (*calcvs)); if (oldnv != NULL) { verify(nvlist_lookup_uint64_array(oldnv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&oldvs, &c) == 0); } else { oldvs = &zerovs; } /* Do we only want to see a specific vdev? */ for (i = 0; i < cb->cb_vdev_names_count; i++) { /* Yes we do. Is this the vdev? */ if (strcmp(name, cb->cb_vdev_names[i]) == 0) { /* * This is our vdev. Since it is the only vdev we * will be displaying, make depth = 0 so that it * doesn't get indented. */ depth = 0; break; } } if (cb->cb_vdev_names_count && (i == cb->cb_vdev_names_count)) { /* Couldn't match the name */ goto children; } verify(nvlist_lookup_uint64_array(newnv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&newvs, &c) == 0); /* * Print the vdev name unless it's is a histogram. Histograms * display the vdev name in the header itself. */ if (!(cb->cb_flags & IOS_ANYHISTO_M)) { if (cb->cb_scripted) { printf("%s", name); } else { if (strlen(name) + depth > cb->cb_namewidth) (void) printf("%*s%s", depth, "", name); else (void) printf("%*s%s%*s", depth, "", name, (int)(cb->cb_namewidth - strlen(name) - depth), ""); } } /* Calculate our scaling factor */ tdelta = newvs->vs_timestamp - oldvs->vs_timestamp; if ((oldvs->vs_timestamp == 0) && (cb->cb_flags & IOS_ANYHISTO_M)) { /* * If we specify printing histograms with no time interval, then * print the histogram numbers over the entire lifetime of the * vdev. */ scale = 1; } else { if (tdelta == 0) scale = 1.0; else scale = (double)NANOSEC / tdelta; } if (cb->cb_flags & IOS_DEFAULT_M) { calc_default_iostats(oldvs, newvs, calcvs); print_iostat_default(calcvs, cb, scale); } if (cb->cb_flags & IOS_LATENCY_M) print_iostat_latency(cb, oldnv, newnv); if (cb->cb_flags & IOS_QUEUES_M) print_iostat_queues(cb, oldnv, newnv); if (cb->cb_flags & IOS_ANYHISTO_M) { printf("\n"); print_iostat_histos(cb, oldnv, newnv, scale, name); } if (cb->vcdl != NULL) { char *path; if (nvlist_lookup_string(newnv, ZPOOL_CONFIG_PATH, &path) == 0) { printf(" "); zpool_print_cmd(cb->vcdl, zpool_get_name(zhp), path); } } if (!(cb->cb_flags & IOS_ANYHISTO_M)) printf("\n"); ret++; children: free(calcvs); if (!cb->cb_verbose) return (ret); if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_CHILDREN, &newchild, &children) != 0) return (ret); if (oldnv) { if (nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_CHILDREN, &oldchild, &oldchildren) != 0) return (ret); children = MIN(oldchildren, children); } /* * print normal top-level devices */ for (c = 0; c < children; c++) { uint64_t ishole = B_FALSE, islog = B_FALSE; (void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_HOLE, &ishole); (void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_LOG, &islog); if (ishole || islog) continue; if (nvlist_exists(newchild[c], ZPOOL_CONFIG_ALLOCATION_BIAS)) continue; vname = zpool_vdev_name(g_zfs, zhp, newchild[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c] : NULL, newchild[c], cb, depth + 2); free(vname); } /* * print all other top-level devices */ for (uint_t n = 0; n < 3; n++) { boolean_t printed = B_FALSE; for (c = 0; c < children; c++) { uint64_t islog = B_FALSE; char *bias = NULL; char *type = NULL; (void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_LOG, &islog); if (islog) { bias = VDEV_ALLOC_CLASS_LOGS; } else { (void) nvlist_lookup_string(newchild[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &bias); (void) nvlist_lookup_string(newchild[c], ZPOOL_CONFIG_TYPE, &type); } if (bias == NULL || strcmp(bias, class_name[n]) != 0) continue; if (!islog && strcmp(type, VDEV_TYPE_INDIRECT) == 0) continue; if (!printed) { if ((!(cb->cb_flags & IOS_ANYHISTO_M)) && !cb->cb_scripted && !cb->cb_vdev_names) { print_iostat_dashes(cb, 0, class_name[n]); } printf("\n"); printed = B_TRUE; } vname = zpool_vdev_name(g_zfs, zhp, newchild[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c] : NULL, newchild[c], cb, depth + 2); free(vname); } } /* * Include level 2 ARC devices in iostat output */ if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_L2CACHE, &newchild, &children) != 0) return (ret); if (oldnv) { if (nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_L2CACHE, &oldchild, &oldchildren) != 0) return (ret); children = MIN(oldchildren, children); } if (children > 0) { if ((!(cb->cb_flags & IOS_ANYHISTO_M)) && !cb->cb_scripted && !cb->cb_vdev_names) { print_iostat_dashes(cb, 0, "cache"); } printf("\n"); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, zhp, newchild[c], cb->cb_name_flags); ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c] : NULL, newchild[c], cb, depth + 2); free(vname); } } return (ret); } static int refresh_iostat(zpool_handle_t *zhp, void *data) { iostat_cbdata_t *cb = data; boolean_t missing; /* * If the pool has disappeared, remove it from the list and continue. */ if (zpool_refresh_stats(zhp, &missing) != 0) return (-1); if (missing) pool_list_remove(cb->cb_list, zhp); return (0); } /* * Callback to print out the iostats for the given pool. */ static int print_iostat(zpool_handle_t *zhp, void *data) { iostat_cbdata_t *cb = data; nvlist_t *oldconfig, *newconfig; nvlist_t *oldnvroot, *newnvroot; int ret; newconfig = zpool_get_config(zhp, &oldconfig); if (cb->cb_iteration == 1) oldconfig = NULL; verify(nvlist_lookup_nvlist(newconfig, ZPOOL_CONFIG_VDEV_TREE, &newnvroot) == 0); if (oldconfig == NULL) oldnvroot = NULL; else verify(nvlist_lookup_nvlist(oldconfig, ZPOOL_CONFIG_VDEV_TREE, &oldnvroot) == 0); ret = print_vdev_stats(zhp, zpool_get_name(zhp), oldnvroot, newnvroot, cb, 0); if ((ret != 0) && !(cb->cb_flags & IOS_ANYHISTO_M) && !cb->cb_scripted && cb->cb_verbose && !cb->cb_vdev_names_count) { print_iostat_separator(cb); if (cb->vcdl != NULL) { print_cmd_columns(cb->vcdl, 1); } printf("\n"); } return (ret); } static int get_columns(void) { struct winsize ws; int columns = 80; int error; if (isatty(STDOUT_FILENO)) { error = ioctl(STDOUT_FILENO, TIOCGWINSZ, &ws); if (error == 0) columns = ws.ws_col; } else { columns = 999; } return (columns); } /* * Return the required length of the pool/vdev name column. The minimum * allowed width and output formatting flags must be provided. */ static int get_namewidth(zpool_handle_t *zhp, int min_width, int flags, boolean_t verbose) { nvlist_t *config, *nvroot; int width = min_width; if ((config = zpool_get_config(zhp, NULL)) != NULL) { verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); unsigned int poolname_len = strlen(zpool_get_name(zhp)); if (verbose == B_FALSE) { width = MAX(poolname_len, min_width); } else { width = MAX(poolname_len, max_width(zhp, nvroot, 0, min_width, flags)); } } return (width); } /* * Parse the input string, get the 'interval' and 'count' value if there is one. */ static void get_interval_count(int *argcp, char **argv, float *iv, unsigned long *cnt) { float interval = 0; unsigned long count = 0; int argc = *argcp; /* * Determine if the last argument is an integer or a pool name */ if (argc > 0 && zfs_isnumber(argv[argc - 1])) { char *end; errno = 0; interval = strtof(argv[argc - 1], &end); if (*end == '\0' && errno == 0) { if (interval == 0) { (void) fprintf(stderr, gettext( "interval cannot be zero\n")); usage(B_FALSE); } /* * Ignore the last parameter */ argc--; } else { /* * If this is not a valid number, just plow on. The * user will get a more informative error message later * on. */ interval = 0; } } /* * If the last argument is also an integer, then we have both a count * and an interval. */ if (argc > 0 && zfs_isnumber(argv[argc - 1])) { char *end; errno = 0; count = interval; interval = strtof(argv[argc - 1], &end); if (*end == '\0' && errno == 0) { if (interval == 0) { (void) fprintf(stderr, gettext( "interval cannot be zero\n")); usage(B_FALSE); } /* * Ignore the last parameter */ argc--; } else { interval = 0; } } *iv = interval; *cnt = count; *argcp = argc; } static void get_timestamp_arg(char c) { if (c == 'u') timestamp_fmt = UDATE; else if (c == 'd') timestamp_fmt = DDATE; else usage(B_FALSE); } /* * Return stat flags that are supported by all pools by both the module and * zpool iostat. "*data" should be initialized to all 0xFFs before running. * It will get ANDed down until only the flags that are supported on all pools * remain. */ static int get_stat_flags_cb(zpool_handle_t *zhp, void *data) { uint64_t *mask = data; nvlist_t *config, *nvroot, *nvx; uint64_t flags = 0; int i, j; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); /* Default stats are always supported, but for completeness.. */ if (nvlist_exists(nvroot, ZPOOL_CONFIG_VDEV_STATS)) flags |= IOS_DEFAULT_M; /* Get our extended stats nvlist from the main list */ if (nvlist_lookup_nvlist(nvroot, ZPOOL_CONFIG_VDEV_STATS_EX, &nvx) != 0) { /* * No extended stats; they're probably running an older * module. No big deal, we support that too. */ goto end; } /* For each extended stat, make sure all its nvpairs are supported */ for (j = 0; j < ARRAY_SIZE(vsx_type_to_nvlist); j++) { if (!vsx_type_to_nvlist[j][0]) continue; /* Start off by assuming the flag is supported, then check */ flags |= (1ULL << j); for (i = 0; vsx_type_to_nvlist[j][i]; i++) { if (!nvlist_exists(nvx, vsx_type_to_nvlist[j][i])) { /* flag isn't supported */ flags = flags & ~(1ULL << j); break; } } } end: *mask = *mask & flags; return (0); } /* * Return a bitmask of stats that are supported on all pools by both the module * and zpool iostat. */ static uint64_t get_stat_flags(zpool_list_t *list) { uint64_t mask = -1; /* * get_stat_flags_cb() will lop off bits from "mask" until only the * flags that are supported on all pools remain. */ pool_list_iter(list, B_FALSE, get_stat_flags_cb, &mask); return (mask); } /* * Return 1 if cb_data->cb_vdev_names[0] is this vdev's name, 0 otherwise. */ static int is_vdev_cb(void *zhp_data, nvlist_t *nv, void *cb_data) { iostat_cbdata_t *cb = cb_data; char *name = NULL; int ret = 0; zpool_handle_t *zhp = zhp_data; name = zpool_vdev_name(g_zfs, zhp, nv, cb->cb_name_flags); if (strcmp(name, cb->cb_vdev_names[0]) == 0) ret = 1; /* match */ free(name); return (ret); } /* * Returns 1 if cb_data->cb_vdev_names[0] is a vdev name, 0 otherwise. */ static int is_vdev(zpool_handle_t *zhp, void *cb_data) { return (for_each_vdev(zhp, is_vdev_cb, cb_data)); } /* * Check if vdevs are in a pool * * Return 1 if all argv[] strings are vdev names in pool "pool_name". Otherwise * return 0. If pool_name is NULL, then search all pools. */ static int are_vdevs_in_pool(int argc, char **argv, char *pool_name, iostat_cbdata_t *cb) { char **tmp_name; int ret = 0; int i; int pool_count = 0; if ((argc == 0) || !*argv) return (0); if (pool_name) pool_count = 1; /* Temporarily hijack cb_vdev_names for a second... */ tmp_name = cb->cb_vdev_names; /* Go though our list of prospective vdev names */ for (i = 0; i < argc; i++) { cb->cb_vdev_names = argv + i; /* Is this name a vdev in our pools? */ ret = for_each_pool(pool_count, &pool_name, B_TRUE, NULL, B_FALSE, is_vdev, cb); if (!ret) { /* No match */ break; } } cb->cb_vdev_names = tmp_name; return (ret); } static int is_pool_cb(zpool_handle_t *zhp, void *data) { char *name = data; if (strcmp(name, zpool_get_name(zhp)) == 0) return (1); return (0); } /* * Do we have a pool named *name? If so, return 1, otherwise 0. */ static int is_pool(char *name) { return (for_each_pool(0, NULL, B_TRUE, NULL, B_FALSE, is_pool_cb, name)); } /* Are all our argv[] strings pool names? If so return 1, 0 otherwise. */ static int are_all_pools(int argc, char **argv) { if ((argc == 0) || !*argv) return (0); while (--argc >= 0) if (!is_pool(argv[argc])) return (0); return (1); } /* * Helper function to print out vdev/pool names we can't resolve. Used for an * error message. */ static void error_list_unresolved_vdevs(int argc, char **argv, char *pool_name, iostat_cbdata_t *cb) { int i; char *name; char *str; for (i = 0; i < argc; i++) { name = argv[i]; if (is_pool(name)) str = gettext("pool"); else if (are_vdevs_in_pool(1, &name, pool_name, cb)) str = gettext("vdev in this pool"); else if (are_vdevs_in_pool(1, &name, NULL, cb)) str = gettext("vdev in another pool"); else str = gettext("unknown"); fprintf(stderr, "\t%s (%s)\n", name, str); } } /* * Same as get_interval_count(), but with additional checks to not misinterpret * guids as interval/count values. Assumes VDEV_NAME_GUID is set in * cb.cb_name_flags. */ static void get_interval_count_filter_guids(int *argc, char **argv, float *interval, unsigned long *count, iostat_cbdata_t *cb) { char **tmpargv = argv; int argc_for_interval = 0; /* Is the last arg an interval value? Or a guid? */ if (*argc >= 1 && !are_vdevs_in_pool(1, &argv[*argc - 1], NULL, cb)) { /* * The last arg is not a guid, so it's probably an * interval value. */ argc_for_interval++; if (*argc >= 2 && !are_vdevs_in_pool(1, &argv[*argc - 2], NULL, cb)) { /* * The 2nd to last arg is not a guid, so it's probably * an interval value. */ argc_for_interval++; } } /* Point to our list of possible intervals */ tmpargv = &argv[*argc - argc_for_interval]; *argc = *argc - argc_for_interval; get_interval_count(&argc_for_interval, tmpargv, interval, count); } /* * Floating point sleep(). Allows you to pass in a floating point value for * seconds. */ static void fsleep(float sec) { struct timespec req; req.tv_sec = floor(sec); req.tv_nsec = (sec - (float)req.tv_sec) * NANOSEC; nanosleep(&req, NULL); } /* * Terminal height, in rows. Returns -1 if stdout is not connected to a TTY or * if we were unable to determine its size. */ static int terminal_height(void) { struct winsize win; if (isatty(STDOUT_FILENO) == 0) return (-1); if (ioctl(STDOUT_FILENO, TIOCGWINSZ, &win) != -1 && win.ws_row > 0) return (win.ws_row); return (-1); } /* * Run one of the zpool status/iostat -c scripts with the help (-h) option and * print the result. * * name: Short name of the script ('iostat'). * path: Full path to the script ('/usr/local/etc/zfs/zpool.d/iostat'); */ static void print_zpool_script_help(char *name, char *path) { char *argv[] = {path, "-h", NULL}; char **lines = NULL; int lines_cnt = 0; int rc; rc = libzfs_run_process_get_stdout_nopath(path, argv, NULL, &lines, &lines_cnt); if (rc != 0 || lines == NULL || lines_cnt <= 0) { if (lines != NULL) libzfs_free_str_array(lines, lines_cnt); return; } for (int i = 0; i < lines_cnt; i++) if (!is_blank_str(lines[i])) printf(" %-14s %s\n", name, lines[i]); libzfs_free_str_array(lines, lines_cnt); } /* * Go though the zpool status/iostat -c scripts in the user's path, run their * help option (-h), and print out the results. */ static void print_zpool_dir_scripts(char *dirpath) { DIR *dir; struct dirent *ent; char fullpath[MAXPATHLEN]; struct stat dir_stat; if ((dir = opendir(dirpath)) != NULL) { /* print all the files and directories within directory */ while ((ent = readdir(dir)) != NULL) { if (snprintf(fullpath, sizeof (fullpath), "%s/%s", dirpath, ent->d_name) >= sizeof (fullpath)) { (void) fprintf(stderr, gettext("internal error: " "ZPOOL_SCRIPTS_PATH too large.\n")); exit(1); } /* Print the scripts */ if (stat(fullpath, &dir_stat) == 0) if (dir_stat.st_mode & S_IXUSR && S_ISREG(dir_stat.st_mode)) print_zpool_script_help(ent->d_name, fullpath); } closedir(dir); } } /* * Print out help text for all zpool status/iostat -c scripts. */ static void print_zpool_script_list(char *subcommand) { char *dir, *sp; printf(gettext("Available 'zpool %s -c' commands:\n"), subcommand); sp = zpool_get_cmd_search_path(); if (sp == NULL) return; dir = strtok(sp, ":"); while (dir != NULL) { print_zpool_dir_scripts(dir); dir = strtok(NULL, ":"); } free(sp); } /* * Set the minimum pool/vdev name column width. The width must be at least 10, * but may be as large as the column width - 42 so it still fits on one line. * NOTE: 42 is the width of the default capacity/operations/bandwidth output */ static int get_namewidth_iostat(zpool_handle_t *zhp, void *data) { iostat_cbdata_t *cb = data; int width, available_width; /* * get_namewidth() returns the maximum width of any name in that column * for any pool/vdev/device line that will be output. */ width = get_namewidth(zhp, cb->cb_namewidth, cb->cb_name_flags | VDEV_NAME_TYPE_ID, cb->cb_verbose); /* * The width we are calculating is the width of the header and also the * padding width for names that are less than maximum width. The stats * take up 42 characters, so the width available for names is: */ available_width = get_columns() - 42; /* * If the maximum width fits on a screen, then great! Make everything * line up by justifying all lines to the same width. If that max * width is larger than what's available, the name plus stats won't fit * on one line, and justifying to that width would cause every line to * wrap on the screen. We only want lines with long names to wrap. * Limit the padding to what won't wrap. */ if (width > available_width) width = available_width; /* * And regardless of whatever the screen width is (get_columns can * return 0 if the width is not known or less than 42 for a narrow * terminal) have the width be a minimum of 10. */ if (width < 10) width = 10; /* Save the calculated width */ cb->cb_namewidth = width; return (0); } /* * zpool iostat [[-c [script1,script2,...]] [-lq]|[-rw]] [-ghHLpPvy] [-n name] * [-T d|u] [[ pool ...]|[pool vdev ...]|[vdev ...]] * [interval [count]] * * -c CMD For each vdev, run command CMD * -g Display guid for individual vdev name. * -L Follow links when resolving vdev path name. * -P Display full path for vdev name. * -v Display statistics for individual vdevs * -h Display help * -p Display values in parsable (exact) format. * -H Scripted mode. Don't display headers, and separate properties * by a single tab. * -l Display average latency * -q Display queue depths * -w Display latency histograms * -r Display request size histogram * -T Display a timestamp in date(1) or Unix format * -n Only print headers once * * This command can be tricky because we want to be able to deal with pool * creation/destruction as well as vdev configuration changes. The bulk of this * processing is handled by the pool_list_* routines in zpool_iter.c. We rely * on pool_list_update() to detect the addition of new pools. Configuration * changes are all handled within libzfs. */ int zpool_do_iostat(int argc, char **argv) { int c; int ret; int npools; float interval = 0; unsigned long count = 0; int winheight = 24; zpool_list_t *list; boolean_t verbose = B_FALSE; boolean_t latency = B_FALSE, l_histo = B_FALSE, rq_histo = B_FALSE; boolean_t queues = B_FALSE, parsable = B_FALSE, scripted = B_FALSE; boolean_t omit_since_boot = B_FALSE; boolean_t guid = B_FALSE; boolean_t follow_links = B_FALSE; boolean_t full_name = B_FALSE; boolean_t headers_once = B_FALSE; iostat_cbdata_t cb = { 0 }; char *cmd = NULL; /* Used for printing error message */ const char flag_to_arg[] = {[IOS_LATENCY] = 'l', [IOS_QUEUES] = 'q', [IOS_L_HISTO] = 'w', [IOS_RQ_HISTO] = 'r'}; uint64_t unsupported_flags; /* check options */ while ((c = getopt(argc, argv, "c:gLPT:vyhplqrwnH")) != -1) { switch (c) { case 'c': if (cmd != NULL) { fprintf(stderr, gettext("Can't set -c flag twice\n")); exit(1); } if (getenv("ZPOOL_SCRIPTS_ENABLED") != NULL && !libzfs_envvar_is_set("ZPOOL_SCRIPTS_ENABLED")) { fprintf(stderr, gettext( "Can't run -c, disabled by " "ZPOOL_SCRIPTS_ENABLED.\n")); exit(1); } if ((getuid() <= 0 || geteuid() <= 0) && !libzfs_envvar_is_set("ZPOOL_SCRIPTS_AS_ROOT")) { fprintf(stderr, gettext( "Can't run -c with root privileges " "unless ZPOOL_SCRIPTS_AS_ROOT is set.\n")); exit(1); } cmd = optarg; verbose = B_TRUE; break; case 'g': guid = B_TRUE; break; case 'L': follow_links = B_TRUE; break; case 'P': full_name = B_TRUE; break; case 'T': get_timestamp_arg(*optarg); break; case 'v': verbose = B_TRUE; break; case 'p': parsable = B_TRUE; break; case 'l': latency = B_TRUE; break; case 'q': queues = B_TRUE; break; case 'H': scripted = B_TRUE; break; case 'w': l_histo = B_TRUE; break; case 'r': rq_histo = B_TRUE; break; case 'y': omit_since_boot = B_TRUE; break; case 'n': headers_once = B_TRUE; break; case 'h': usage(B_FALSE); break; case '?': if (optopt == 'c') { print_zpool_script_list("iostat"); exit(0); } else { fprintf(stderr, gettext("invalid option '%c'\n"), optopt); } usage(B_FALSE); } } argc -= optind; argv += optind; cb.cb_literal = parsable; cb.cb_scripted = scripted; if (guid) cb.cb_name_flags |= VDEV_NAME_GUID; if (follow_links) cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS; if (full_name) cb.cb_name_flags |= VDEV_NAME_PATH; cb.cb_iteration = 0; cb.cb_namewidth = 0; cb.cb_verbose = verbose; /* Get our interval and count values (if any) */ if (guid) { get_interval_count_filter_guids(&argc, argv, &interval, &count, &cb); } else { get_interval_count(&argc, argv, &interval, &count); } if (argc == 0) { /* No args, so just print the defaults. */ } else if (are_all_pools(argc, argv)) { /* All the args are pool names */ } else if (are_vdevs_in_pool(argc, argv, NULL, &cb)) { /* All the args are vdevs */ cb.cb_vdev_names = argv; cb.cb_vdev_names_count = argc; argc = 0; /* No pools to process */ } else if (are_all_pools(1, argv)) { /* The first arg is a pool name */ if (are_vdevs_in_pool(argc - 1, argv + 1, argv[0], &cb)) { /* ...and the rest are vdev names */ cb.cb_vdev_names = argv + 1; cb.cb_vdev_names_count = argc - 1; argc = 1; /* One pool to process */ } else { fprintf(stderr, gettext("Expected either a list of ")); fprintf(stderr, gettext("pools, or list of vdevs in")); fprintf(stderr, " \"%s\", ", argv[0]); fprintf(stderr, gettext("but got:\n")); error_list_unresolved_vdevs(argc - 1, argv + 1, argv[0], &cb); fprintf(stderr, "\n"); usage(B_FALSE); return (1); } } else { /* * The args don't make sense. The first arg isn't a pool name, * nor are all the args vdevs. */ fprintf(stderr, gettext("Unable to parse pools/vdevs list.\n")); fprintf(stderr, "\n"); return (1); } if (cb.cb_vdev_names_count != 0) { /* * If user specified vdevs, it implies verbose. */ cb.cb_verbose = B_TRUE; } /* * Construct the list of all interesting pools. */ ret = 0; if ((list = pool_list_get(argc, argv, NULL, parsable, &ret)) == NULL) return (1); if (pool_list_count(list) == 0 && argc != 0) { pool_list_free(list); return (1); } if (pool_list_count(list) == 0 && interval == 0) { pool_list_free(list); (void) fprintf(stderr, gettext("no pools available\n")); return (1); } if ((l_histo || rq_histo) && (cmd != NULL || latency || queues)) { pool_list_free(list); (void) fprintf(stderr, gettext("[-r|-w] isn't allowed with [-c|-l|-q]\n")); usage(B_FALSE); return (1); } if (l_histo && rq_histo) { pool_list_free(list); (void) fprintf(stderr, gettext("Only one of [-r|-w] can be passed at a time\n")); usage(B_FALSE); return (1); } /* * Enter the main iostat loop. */ cb.cb_list = list; if (l_histo) { /* * Histograms tables look out of place when you try to display * them with the other stats, so make a rule that you can only * print histograms by themselves. */ cb.cb_flags = IOS_L_HISTO_M; } else if (rq_histo) { cb.cb_flags = IOS_RQ_HISTO_M; } else { cb.cb_flags = IOS_DEFAULT_M; if (latency) cb.cb_flags |= IOS_LATENCY_M; if (queues) cb.cb_flags |= IOS_QUEUES_M; } /* * See if the module supports all the stats we want to display. */ unsupported_flags = cb.cb_flags & ~get_stat_flags(list); if (unsupported_flags) { uint64_t f; int idx; fprintf(stderr, gettext("The loaded zfs module doesn't support:")); /* for each bit set in unsupported_flags */ for (f = unsupported_flags; f; f &= ~(1ULL << idx)) { idx = lowbit64(f) - 1; fprintf(stderr, " -%c", flag_to_arg[idx]); } fprintf(stderr, ". Try running a newer module.\n"); pool_list_free(list); return (1); } for (;;) { if ((npools = pool_list_count(list)) == 0) (void) fprintf(stderr, gettext("no pools available\n")); else { /* * If this is the first iteration and -y was supplied * we skip any printing. */ boolean_t skip = (omit_since_boot && cb.cb_iteration == 0); /* * Refresh all statistics. This is done as an * explicit step before calculating the maximum name * width, so that any * configuration changes are * properly accounted for. */ (void) pool_list_iter(list, B_FALSE, refresh_iostat, &cb); /* * Iterate over all pools to determine the maximum width * for the pool / device name column across all pools. */ cb.cb_namewidth = 0; (void) pool_list_iter(list, B_FALSE, get_namewidth_iostat, &cb); if (timestamp_fmt != NODATE) print_timestamp(timestamp_fmt); if (cmd != NULL && cb.cb_verbose && !(cb.cb_flags & IOS_ANYHISTO_M)) { cb.vcdl = all_pools_for_each_vdev_run(argc, argv, cmd, g_zfs, cb.cb_vdev_names, cb.cb_vdev_names_count, cb.cb_name_flags); } else { cb.vcdl = NULL; } /* * Check terminal size so we can print headers * even when terminal window has its height * changed. */ winheight = terminal_height(); /* * Are we connected to TTY? If not, headers_once * should be true, to avoid breaking scripts. */ if (winheight < 0) headers_once = B_TRUE; /* * If it's the first time and we're not skipping it, * or either skip or verbose mode, print the header. * * The histogram code explicitly prints its header on * every vdev, so skip this for histograms. */ if (((++cb.cb_iteration == 1 && !skip) || (skip != verbose) || (!headers_once && (cb.cb_iteration % winheight) == 0)) && (!(cb.cb_flags & IOS_ANYHISTO_M)) && !cb.cb_scripted) print_iostat_header(&cb); if (skip) { (void) fsleep(interval); continue; } pool_list_iter(list, B_FALSE, print_iostat, &cb); /* * If there's more than one pool, and we're not in * verbose mode (which prints a separator for us), * then print a separator. * * In addition, if we're printing specific vdevs then * we also want an ending separator. */ if (((npools > 1 && !verbose && !(cb.cb_flags & IOS_ANYHISTO_M)) || (!(cb.cb_flags & IOS_ANYHISTO_M) && cb.cb_vdev_names_count)) && !cb.cb_scripted) { print_iostat_separator(&cb); if (cb.vcdl != NULL) print_cmd_columns(cb.vcdl, 1); printf("\n"); } if (cb.vcdl != NULL) free_vdev_cmd_data_list(cb.vcdl); } /* * Flush the output so that redirection to a file isn't buffered * indefinitely. */ (void) fflush(stdout); if (interval == 0) break; if (count != 0 && --count == 0) break; (void) fsleep(interval); } pool_list_free(list); return (ret); } typedef struct list_cbdata { boolean_t cb_verbose; int cb_name_flags; int cb_namewidth; boolean_t cb_scripted; zprop_list_t *cb_proplist; boolean_t cb_literal; } list_cbdata_t; /* * Given a list of columns to display, output appropriate headers for each one. */ static void print_header(list_cbdata_t *cb) { zprop_list_t *pl = cb->cb_proplist; char headerbuf[ZPOOL_MAXPROPLEN]; const char *header; boolean_t first = B_TRUE; boolean_t right_justify; size_t width = 0; for (; pl != NULL; pl = pl->pl_next) { width = pl->pl_width; if (first && cb->cb_verbose) { /* * Reset the width to accommodate the verbose listing * of devices. */ width = cb->cb_namewidth; } if (!first) (void) printf(" "); else first = B_FALSE; right_justify = B_FALSE; if (pl->pl_prop != ZPROP_INVAL) { header = zpool_prop_column_name(pl->pl_prop); right_justify = zpool_prop_align_right(pl->pl_prop); } else { int i; for (i = 0; pl->pl_user_prop[i] != '\0'; i++) headerbuf[i] = toupper(pl->pl_user_prop[i]); headerbuf[i] = '\0'; header = headerbuf; } if (pl->pl_next == NULL && !right_justify) (void) printf("%s", header); else if (right_justify) (void) printf("%*s", (int)width, header); else (void) printf("%-*s", (int)width, header); } (void) printf("\n"); } /* * Given a pool and a list of properties, print out all the properties according * to the described layout. Used by zpool_do_list(). */ static void print_pool(zpool_handle_t *zhp, list_cbdata_t *cb) { zprop_list_t *pl = cb->cb_proplist; boolean_t first = B_TRUE; char property[ZPOOL_MAXPROPLEN]; char *propstr; boolean_t right_justify; size_t width; for (; pl != NULL; pl = pl->pl_next) { width = pl->pl_width; if (first && cb->cb_verbose) { /* * Reset the width to accommodate the verbose listing * of devices. */ width = cb->cb_namewidth; } if (!first) { if (cb->cb_scripted) (void) printf("\t"); else (void) printf(" "); } else { first = B_FALSE; } right_justify = B_FALSE; if (pl->pl_prop != ZPROP_INVAL) { if (zpool_get_prop(zhp, pl->pl_prop, property, sizeof (property), NULL, cb->cb_literal) != 0) propstr = "-"; else propstr = property; right_justify = zpool_prop_align_right(pl->pl_prop); } else if ((zpool_prop_feature(pl->pl_user_prop) || zpool_prop_unsupported(pl->pl_user_prop)) && zpool_prop_get_feature(zhp, pl->pl_user_prop, property, sizeof (property)) == 0) { propstr = property; } else { propstr = "-"; } /* * If this is being called in scripted mode, or if this is the * last column and it is left-justified, don't include a width * format specifier. */ if (cb->cb_scripted || (pl->pl_next == NULL && !right_justify)) (void) printf("%s", propstr); else if (right_justify) (void) printf("%*s", (int)width, propstr); else (void) printf("%-*s", (int)width, propstr); } (void) printf("\n"); } static void print_one_column(zpool_prop_t prop, uint64_t value, const char *str, boolean_t scripted, boolean_t valid, enum zfs_nicenum_format format) { char propval[64]; boolean_t fixed; size_t width = zprop_width(prop, &fixed, ZFS_TYPE_POOL); switch (prop) { case ZPOOL_PROP_SIZE: case ZPOOL_PROP_EXPANDSZ: case ZPOOL_PROP_CHECKPOINT: case ZPOOL_PROP_DEDUPRATIO: if (value == 0) (void) strlcpy(propval, "-", sizeof (propval)); else zfs_nicenum_format(value, propval, sizeof (propval), format); break; case ZPOOL_PROP_FRAGMENTATION: if (value == ZFS_FRAG_INVALID) { (void) strlcpy(propval, "-", sizeof (propval)); } else if (format == ZFS_NICENUM_RAW) { (void) snprintf(propval, sizeof (propval), "%llu", (unsigned long long)value); } else { (void) snprintf(propval, sizeof (propval), "%llu%%", (unsigned long long)value); } break; case ZPOOL_PROP_CAPACITY: /* capacity value is in parts-per-10,000 (aka permyriad) */ if (format == ZFS_NICENUM_RAW) (void) snprintf(propval, sizeof (propval), "%llu", (unsigned long long)value / 100); else (void) snprintf(propval, sizeof (propval), value < 1000 ? "%1.2f%%" : value < 10000 ? "%2.1f%%" : "%3.0f%%", value / 100.0); break; case ZPOOL_PROP_HEALTH: width = 8; (void) strlcpy(propval, str, sizeof (propval)); break; default: zfs_nicenum_format(value, propval, sizeof (propval), format); } if (!valid) (void) strlcpy(propval, "-", sizeof (propval)); if (scripted) (void) printf("\t%s", propval); else (void) printf(" %*s", (int)width, propval); } /* * print static default line per vdev * not compatible with '-o' option */ static void print_list_stats(zpool_handle_t *zhp, const char *name, nvlist_t *nv, list_cbdata_t *cb, int depth, boolean_t isspare) { nvlist_t **child; vdev_stat_t *vs; uint_t c, children; char *vname; boolean_t scripted = cb->cb_scripted; uint64_t islog = B_FALSE; char *dashes = "%-*s - - - - " "- - - - -\n"; verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c) == 0); if (name != NULL) { boolean_t toplevel = (vs->vs_space != 0); uint64_t cap; enum zfs_nicenum_format format; const char *state; if (cb->cb_literal) format = ZFS_NICENUM_RAW; else format = ZFS_NICENUM_1024; if (strcmp(name, VDEV_TYPE_INDIRECT) == 0) return; if (scripted) (void) printf("\t%s", name); else if (strlen(name) + depth > cb->cb_namewidth) (void) printf("%*s%s", depth, "", name); else (void) printf("%*s%s%*s", depth, "", name, (int)(cb->cb_namewidth - strlen(name) - depth), ""); /* * Print the properties for the individual vdevs. Some * properties are only applicable to toplevel vdevs. The * 'toplevel' boolean value is passed to the print_one_column() * to indicate that the value is valid. */ if (vs->vs_pspace) print_one_column(ZPOOL_PROP_SIZE, vs->vs_pspace, NULL, scripted, B_TRUE, format); else print_one_column(ZPOOL_PROP_SIZE, vs->vs_space, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_ALLOCATED, vs->vs_alloc, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_FREE, vs->vs_space - vs->vs_alloc, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_CHECKPOINT, vs->vs_checkpoint_space, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_EXPANDSZ, vs->vs_esize, NULL, scripted, B_TRUE, format); print_one_column(ZPOOL_PROP_FRAGMENTATION, vs->vs_fragmentation, NULL, scripted, (vs->vs_fragmentation != ZFS_FRAG_INVALID && toplevel), format); cap = (vs->vs_space == 0) ? 0 : (vs->vs_alloc * 10000 / vs->vs_space); print_one_column(ZPOOL_PROP_CAPACITY, cap, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_DEDUPRATIO, 0, NULL, scripted, toplevel, format); state = zpool_state_to_name(vs->vs_state, vs->vs_aux); if (isspare) { if (vs->vs_aux == VDEV_AUX_SPARED) state = "INUSE"; else if (vs->vs_state == VDEV_STATE_HEALTHY) state = "AVAIL"; } print_one_column(ZPOOL_PROP_HEALTH, 0, state, scripted, B_TRUE, format); (void) printf("\n"); } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return; /* list the normal vdevs first */ for (c = 0; c < children; c++) { uint64_t ishole = B_FALSE; if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE, &ishole) == 0 && ishole) continue; if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &islog) == 0 && islog) continue; if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS)) continue; vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); print_list_stats(zhp, vname, child[c], cb, depth + 2, B_FALSE); free(vname); } /* list the classes: 'logs', 'dedup', and 'special' */ for (uint_t n = 0; n < 3; n++) { boolean_t printed = B_FALSE; for (c = 0; c < children; c++) { char *bias = NULL; char *type = NULL; if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &islog) == 0 && islog) { bias = VDEV_ALLOC_CLASS_LOGS; } else { (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &bias); (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_TYPE, &type); } if (bias == NULL || strcmp(bias, class_name[n]) != 0) continue; if (!islog && strcmp(type, VDEV_TYPE_INDIRECT) == 0) continue; if (!printed) { /* LINTED E_SEC_PRINTF_VAR_FMT */ (void) printf(dashes, cb->cb_namewidth, class_name[n]); printed = B_TRUE; } vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); print_list_stats(zhp, vname, child[c], cb, depth + 2, B_FALSE); free(vname); } } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) == 0 && children > 0) { /* LINTED E_SEC_PRINTF_VAR_FMT */ (void) printf(dashes, cb->cb_namewidth, "cache"); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags); print_list_stats(zhp, vname, child[c], cb, depth + 2, B_FALSE); free(vname); } } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0 && children > 0) { /* LINTED E_SEC_PRINTF_VAR_FMT */ (void) printf(dashes, cb->cb_namewidth, "spare"); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags); print_list_stats(zhp, vname, child[c], cb, depth + 2, B_TRUE); free(vname); } } } /* * Generic callback function to list a pool. */ static int list_callback(zpool_handle_t *zhp, void *data) { list_cbdata_t *cbp = data; print_pool(zhp, cbp); if (cbp->cb_verbose) { nvlist_t *config, *nvroot; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); print_list_stats(zhp, NULL, nvroot, cbp, 0, B_FALSE); } return (0); } /* * Set the minimum pool/vdev name column width. The width must be at least 9, * but may be as large as needed. */ static int get_namewidth_list(zpool_handle_t *zhp, void *data) { list_cbdata_t *cb = data; int width; width = get_namewidth(zhp, cb->cb_namewidth, cb->cb_name_flags | VDEV_NAME_TYPE_ID, cb->cb_verbose); if (width < 9) width = 9; cb->cb_namewidth = width; return (0); } /* * zpool list [-gHLpP] [-o prop[,prop]*] [-T d|u] [pool] ... [interval [count]] * * -g Display guid for individual vdev name. * -H Scripted mode. Don't display headers, and separate properties * by a single tab. * -L Follow links when resolving vdev path name. * -o List of properties to display. Defaults to * "name,size,allocated,free,expandsize,fragmentation,capacity," * "dedupratio,health,altroot" * -p Display values in parsable (exact) format. * -P Display full path for vdev name. * -T Display a timestamp in date(1) or Unix format * * List all pools in the system, whether or not they're healthy. Output space * statistics for each one, as well as health status summary. */ int zpool_do_list(int argc, char **argv) { int c; int ret = 0; list_cbdata_t cb = { 0 }; static char default_props[] = "name,size,allocated,free,checkpoint,expandsize,fragmentation," "capacity,dedupratio,health,altroot"; char *props = default_props; float interval = 0; unsigned long count = 0; zpool_list_t *list; boolean_t first = B_TRUE; /* check options */ while ((c = getopt(argc, argv, ":gHLo:pPT:v")) != -1) { switch (c) { case 'g': cb.cb_name_flags |= VDEV_NAME_GUID; break; case 'H': cb.cb_scripted = B_TRUE; break; case 'L': cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS; break; case 'o': props = optarg; break; case 'P': cb.cb_name_flags |= VDEV_NAME_PATH; break; case 'p': cb.cb_literal = B_TRUE; break; case 'T': get_timestamp_arg(*optarg); break; case 'v': cb.cb_verbose = B_TRUE; cb.cb_namewidth = 8; /* 8 until precalc is avail */ break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); usage(B_FALSE); break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; get_interval_count(&argc, argv, &interval, &count); if (zprop_get_list(g_zfs, props, &cb.cb_proplist, ZFS_TYPE_POOL) != 0) usage(B_FALSE); for (;;) { if ((list = pool_list_get(argc, argv, &cb.cb_proplist, cb.cb_literal, &ret)) == NULL) return (1); if (pool_list_count(list) == 0) break; cb.cb_namewidth = 0; (void) pool_list_iter(list, B_FALSE, get_namewidth_list, &cb); if (timestamp_fmt != NODATE) print_timestamp(timestamp_fmt); if (!cb.cb_scripted && (first || cb.cb_verbose)) { print_header(&cb); first = B_FALSE; } ret = pool_list_iter(list, B_TRUE, list_callback, &cb); if (interval == 0) break; if (count != 0 && --count == 0) break; pool_list_free(list); (void) fsleep(interval); } if (argc == 0 && !cb.cb_scripted && pool_list_count(list) == 0) { (void) printf(gettext("no pools available\n")); ret = 0; } pool_list_free(list); zprop_free_list(cb.cb_proplist); return (ret); } static int zpool_do_attach_or_replace(int argc, char **argv, int replacing) { boolean_t force = B_FALSE; boolean_t rebuild = B_FALSE; boolean_t wait = B_FALSE; int c; nvlist_t *nvroot; char *poolname, *old_disk, *new_disk; zpool_handle_t *zhp; nvlist_t *props = NULL; char *propval; int ret; /* check options */ while ((c = getopt(argc, argv, "fo:sw")) != -1) { switch (c) { case 'f': force = B_TRUE; break; case 'o': if ((propval = strchr(optarg, '=')) == NULL) { (void) fprintf(stderr, gettext("missing " "'=' for -o option\n")); usage(B_FALSE); } *propval = '\0'; propval++; if ((strcmp(optarg, ZPOOL_CONFIG_ASHIFT) != 0) || (add_prop_list(optarg, propval, &props, B_TRUE))) usage(B_FALSE); break; case 's': rebuild = B_TRUE; break; case 'w': wait = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } poolname = argv[0]; if (argc < 2) { (void) fprintf(stderr, gettext("missing specification\n")); usage(B_FALSE); } old_disk = argv[1]; if (argc < 3) { if (!replacing) { (void) fprintf(stderr, gettext("missing specification\n")); usage(B_FALSE); } new_disk = old_disk; argc -= 1; argv += 1; } else { new_disk = argv[2]; argc -= 2; argv += 2; } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } if ((zhp = zpool_open(g_zfs, poolname)) == NULL) { nvlist_free(props); return (1); } if (zpool_get_config(zhp, NULL) == NULL) { (void) fprintf(stderr, gettext("pool '%s' is unavailable\n"), poolname); zpool_close(zhp); nvlist_free(props); return (1); } /* unless manually specified use "ashift" pool property (if set) */ if (!nvlist_exists(props, ZPOOL_CONFIG_ASHIFT)) { int intval; zprop_source_t src; char strval[ZPOOL_MAXPROPLEN]; intval = zpool_get_prop_int(zhp, ZPOOL_PROP_ASHIFT, &src); if (src != ZPROP_SRC_DEFAULT) { (void) sprintf(strval, "%" PRId32, intval); verify(add_prop_list(ZPOOL_CONFIG_ASHIFT, strval, &props, B_TRUE) == 0); } } nvroot = make_root_vdev(zhp, props, force, B_FALSE, replacing, B_FALSE, argc, argv); if (nvroot == NULL) { zpool_close(zhp); nvlist_free(props); return (1); } ret = zpool_vdev_attach(zhp, old_disk, new_disk, nvroot, replacing, rebuild); if (ret == 0 && wait) ret = zpool_wait(zhp, replacing ? ZPOOL_WAIT_REPLACE : ZPOOL_WAIT_RESILVER); nvlist_free(props); nvlist_free(nvroot); zpool_close(zhp); return (ret); } /* * zpool replace [-fsw] [-o property=value] * * -f Force attach, even if appears to be in use. * -s Use sequential instead of healing reconstruction for resilver. * -o Set property=value. * -w Wait for replacing to complete before returning * * Replace with . */ /* ARGSUSED */ int zpool_do_replace(int argc, char **argv) { return (zpool_do_attach_or_replace(argc, argv, B_TRUE)); } /* * zpool attach [-fsw] [-o property=value] * * -f Force attach, even if appears to be in use. * -s Use sequential instead of healing reconstruction for resilver. * -o Set property=value. * -w Wait for resilvering to complete before returning * * Attach to the mirror containing . If is not * part of a mirror, then will be transformed into a mirror of * and . In either case, will begin life * with a DTL of [0, now], and will immediately begin to resilver itself. */ int zpool_do_attach(int argc, char **argv) { return (zpool_do_attach_or_replace(argc, argv, B_FALSE)); } /* * zpool detach [-f] * * -f Force detach of , even if DTLs argue against it * (not supported yet) * * Detach a device from a mirror. The operation will be refused if * is the last device in the mirror, or if the DTLs indicate that this device * has the only valid copy of some data. */ /* ARGSUSED */ int zpool_do_detach(int argc, char **argv) { int c; char *poolname, *path; zpool_handle_t *zhp; int ret; /* check options */ while ((c = getopt(argc, argv, "")) != -1) { switch (c) { case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing specification\n")); usage(B_FALSE); } poolname = argv[0]; path = argv[1]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); ret = zpool_vdev_detach(zhp, path); zpool_close(zhp); return (ret); } /* * zpool split [-gLnP] [-o prop=val] ... * [-o mntopt] ... * [-R altroot] [ ...] * * -g Display guid for individual vdev name. * -L Follow links when resolving vdev path name. * -n Do not split the pool, but display the resulting layout if * it were to be split. * -o Set property=value, or set mount options. * -P Display full path for vdev name. * -R Mount the split-off pool under an alternate root. * -l Load encryption keys while importing. * * Splits the named pool and gives it the new pool name. Devices to be split * off may be listed, provided that no more than one device is specified * per top-level vdev mirror. The newly split pool is left in an exported * state unless -R is specified. * * Restrictions: the top-level of the pool pool must only be made up of * mirrors; all devices in the pool must be healthy; no device may be * undergoing a resilvering operation. */ int zpool_do_split(int argc, char **argv) { char *srcpool, *newpool, *propval; char *mntopts = NULL; splitflags_t flags; int c, ret = 0; boolean_t loadkeys = B_FALSE; zpool_handle_t *zhp; nvlist_t *config, *props = NULL; flags.dryrun = B_FALSE; flags.import = B_FALSE; flags.name_flags = 0; /* check options */ while ((c = getopt(argc, argv, ":gLR:lno:P")) != -1) { switch (c) { case 'g': flags.name_flags |= VDEV_NAME_GUID; break; case 'L': flags.name_flags |= VDEV_NAME_FOLLOW_LINKS; break; case 'R': flags.import = B_TRUE; if (add_prop_list( zpool_prop_to_name(ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE) != 0) { nvlist_free(props); usage(B_FALSE); } break; case 'l': loadkeys = B_TRUE; break; case 'n': flags.dryrun = B_TRUE; break; case 'o': if ((propval = strchr(optarg, '=')) != NULL) { *propval = '\0'; propval++; if (add_prop_list(optarg, propval, &props, B_TRUE) != 0) { nvlist_free(props); usage(B_FALSE); } } else { mntopts = optarg; } break; case 'P': flags.name_flags |= VDEV_NAME_PATH; break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); usage(B_FALSE); break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); break; } } if (!flags.import && mntopts != NULL) { (void) fprintf(stderr, gettext("setting mntopts is only " "valid when importing the pool\n")); usage(B_FALSE); } if (!flags.import && loadkeys) { (void) fprintf(stderr, gettext("loading keys is only " "valid when importing the pool\n")); usage(B_FALSE); } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("Missing pool name\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("Missing new pool name\n")); usage(B_FALSE); } srcpool = argv[0]; newpool = argv[1]; argc -= 2; argv += 2; if ((zhp = zpool_open(g_zfs, srcpool)) == NULL) { nvlist_free(props); return (1); } config = split_mirror_vdev(zhp, newpool, props, flags, argc, argv); if (config == NULL) { ret = 1; } else { if (flags.dryrun) { (void) printf(gettext("would create '%s' with the " "following layout:\n\n"), newpool); print_vdev_tree(NULL, newpool, config, 0, "", flags.name_flags); print_vdev_tree(NULL, "dedup", config, 0, VDEV_ALLOC_BIAS_DEDUP, 0); print_vdev_tree(NULL, "special", config, 0, VDEV_ALLOC_BIAS_SPECIAL, 0); } } zpool_close(zhp); if (ret != 0 || flags.dryrun || !flags.import) { nvlist_free(config); nvlist_free(props); return (ret); } /* * The split was successful. Now we need to open the new * pool and import it. */ if ((zhp = zpool_open_canfail(g_zfs, newpool)) == NULL) { nvlist_free(config); nvlist_free(props); return (1); } if (loadkeys) { ret = zfs_crypto_attempt_load_keys(g_zfs, newpool); if (ret != 0) ret = 1; } if (zpool_get_state(zhp) != POOL_STATE_UNAVAIL && zpool_enable_datasets(zhp, mntopts, 0) != 0) { ret = 1; (void) fprintf(stderr, gettext("Split was successful, but " "the datasets could not all be mounted\n")); (void) fprintf(stderr, gettext("Try doing '%s' with a " "different altroot\n"), "zpool import"); } zpool_close(zhp); nvlist_free(config); nvlist_free(props); return (ret); } /* * zpool online ... */ int zpool_do_online(int argc, char **argv) { int c, i; char *poolname; zpool_handle_t *zhp; int ret = 0; vdev_state_t newstate; int flags = 0; /* check options */ while ((c = getopt(argc, argv, "e")) != -1) { switch (c) { case 'e': flags |= ZFS_ONLINE_EXPAND; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing device name\n")); usage(B_FALSE); } poolname = argv[0]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); for (i = 1; i < argc; i++) { if (zpool_vdev_online(zhp, argv[i], flags, &newstate) == 0) { if (newstate != VDEV_STATE_HEALTHY) { (void) printf(gettext("warning: device '%s' " "onlined, but remains in faulted state\n"), argv[i]); if (newstate == VDEV_STATE_FAULTED) (void) printf(gettext("use 'zpool " "clear' to restore a faulted " "device\n")); else (void) printf(gettext("use 'zpool " "replace' to replace devices " "that are no longer present\n")); } } else { ret = 1; } } zpool_close(zhp); return (ret); } /* * zpool offline [-ft] ... * * -f Force the device into a faulted state. * * -t Only take the device off-line temporarily. The offline/faulted * state will not be persistent across reboots. */ /* ARGSUSED */ int zpool_do_offline(int argc, char **argv) { int c, i; char *poolname; zpool_handle_t *zhp; int ret = 0; boolean_t istmp = B_FALSE; boolean_t fault = B_FALSE; /* check options */ while ((c = getopt(argc, argv, "ft")) != -1) { switch (c) { case 'f': fault = B_TRUE; break; case 't': istmp = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing device name\n")); usage(B_FALSE); } poolname = argv[0]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); for (i = 1; i < argc; i++) { if (fault) { uint64_t guid = zpool_vdev_path_to_guid(zhp, argv[i]); vdev_aux_t aux; if (istmp == B_FALSE) { /* Force the fault to persist across imports */ aux = VDEV_AUX_EXTERNAL_PERSIST; } else { aux = VDEV_AUX_EXTERNAL; } if (guid == 0 || zpool_vdev_fault(zhp, guid, aux) != 0) ret = 1; } else { if (zpool_vdev_offline(zhp, argv[i], istmp) != 0) ret = 1; } } zpool_close(zhp); return (ret); } /* * zpool clear [device] * * Clear all errors associated with a pool or a particular device. */ int zpool_do_clear(int argc, char **argv) { int c; int ret = 0; boolean_t dryrun = B_FALSE; boolean_t do_rewind = B_FALSE; boolean_t xtreme_rewind = B_FALSE; uint32_t rewind_policy = ZPOOL_NO_REWIND; nvlist_t *policy = NULL; zpool_handle_t *zhp; char *pool, *device; /* check options */ while ((c = getopt(argc, argv, "FnX")) != -1) { switch (c) { case 'F': do_rewind = B_TRUE; break; case 'n': dryrun = B_TRUE; break; case 'X': xtreme_rewind = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc > 2) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } if ((dryrun || xtreme_rewind) && !do_rewind) { (void) fprintf(stderr, gettext("-n or -X only meaningful with -F\n")); usage(B_FALSE); } if (dryrun) rewind_policy = ZPOOL_TRY_REWIND; else if (do_rewind) rewind_policy = ZPOOL_DO_REWIND; if (xtreme_rewind) rewind_policy |= ZPOOL_EXTREME_REWIND; /* In future, further rewind policy choices can be passed along here */ if (nvlist_alloc(&policy, NV_UNIQUE_NAME, 0) != 0 || nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY, rewind_policy) != 0) { return (1); } pool = argv[0]; device = argc == 2 ? argv[1] : NULL; if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) { nvlist_free(policy); return (1); } if (zpool_clear(zhp, device, policy) != 0) ret = 1; zpool_close(zhp); nvlist_free(policy); return (ret); } /* * zpool reguid */ int zpool_do_reguid(int argc, char **argv) { int c; char *poolname; zpool_handle_t *zhp; int ret = 0; /* check options */ while ((c = getopt(argc, argv, "")) != -1) { switch (c) { case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } poolname = argv[0]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); ret = zpool_reguid(zhp); zpool_close(zhp); return (ret); } /* * zpool reopen * * Reopen the pool so that the kernel can update the sizes of all vdevs. */ int zpool_do_reopen(int argc, char **argv) { int c; int ret = 0; boolean_t scrub_restart = B_TRUE; /* check options */ while ((c = getopt(argc, argv, "n")) != -1) { switch (c) { case 'n': scrub_restart = B_FALSE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* if argc == 0 we will execute zpool_reopen_one on all pools */ ret = for_each_pool(argc, argv, B_TRUE, NULL, B_FALSE, zpool_reopen_one, &scrub_restart); return (ret); } typedef struct scrub_cbdata { int cb_type; pool_scrub_cmd_t cb_scrub_cmd; } scrub_cbdata_t; static boolean_t zpool_has_checkpoint(zpool_handle_t *zhp) { nvlist_t *config, *nvroot; config = zpool_get_config(zhp, NULL); if (config != NULL) { pool_checkpoint_stat_t *pcs = NULL; uint_t c; nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c); if (pcs == NULL || pcs->pcs_state == CS_NONE) return (B_FALSE); assert(pcs->pcs_state == CS_CHECKPOINT_EXISTS || pcs->pcs_state == CS_CHECKPOINT_DISCARDING); return (B_TRUE); } return (B_FALSE); } static int scrub_callback(zpool_handle_t *zhp, void *data) { scrub_cbdata_t *cb = data; int err; /* * Ignore faulted pools. */ if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) { (void) fprintf(stderr, gettext("cannot scan '%s': pool is " "currently unavailable\n"), zpool_get_name(zhp)); return (1); } err = zpool_scan(zhp, cb->cb_type, cb->cb_scrub_cmd); if (err == 0 && zpool_has_checkpoint(zhp) && cb->cb_type == POOL_SCAN_SCRUB) { (void) printf(gettext("warning: will not scrub state that " "belongs to the checkpoint of pool '%s'\n"), zpool_get_name(zhp)); } return (err != 0); } static int wait_callback(zpool_handle_t *zhp, void *data) { zpool_wait_activity_t *act = data; return (zpool_wait(zhp, *act)); } /* * zpool scrub [-s | -p] [-w] ... * * -s Stop. Stops any in-progress scrub. * -p Pause. Pause in-progress scrub. * -w Wait. Blocks until scrub has completed. */ int zpool_do_scrub(int argc, char **argv) { int c; scrub_cbdata_t cb; boolean_t wait = B_FALSE; int error; cb.cb_type = POOL_SCAN_SCRUB; cb.cb_scrub_cmd = POOL_SCRUB_NORMAL; /* check options */ while ((c = getopt(argc, argv, "spw")) != -1) { switch (c) { case 's': cb.cb_type = POOL_SCAN_NONE; break; case 'p': cb.cb_scrub_cmd = POOL_SCRUB_PAUSE; break; case 'w': wait = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } if (cb.cb_type == POOL_SCAN_NONE && cb.cb_scrub_cmd == POOL_SCRUB_PAUSE) { (void) fprintf(stderr, gettext("invalid option combination: " "-s and -p are mutually exclusive\n")); usage(B_FALSE); } if (wait && (cb.cb_type == POOL_SCAN_NONE || cb.cb_scrub_cmd == POOL_SCRUB_PAUSE)) { (void) fprintf(stderr, gettext("invalid option combination: " "-w cannot be used with -p or -s\n")); usage(B_FALSE); } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } error = for_each_pool(argc, argv, B_TRUE, NULL, B_FALSE, scrub_callback, &cb); if (wait && !error) { zpool_wait_activity_t act = ZPOOL_WAIT_SCRUB; error = for_each_pool(argc, argv, B_TRUE, NULL, B_FALSE, wait_callback, &act); } return (error); } /* * zpool resilver ... * * Restarts any in-progress resilver */ int zpool_do_resilver(int argc, char **argv) { int c; scrub_cbdata_t cb; cb.cb_type = POOL_SCAN_RESILVER; cb.cb_scrub_cmd = POOL_SCRUB_NORMAL; /* check options */ while ((c = getopt(argc, argv, "")) != -1) { switch (c) { case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } return (for_each_pool(argc, argv, B_TRUE, NULL, B_FALSE, scrub_callback, &cb)); } /* * zpool trim [-d] [-r ] [-c | -s] [ ...] * * -c Cancel. Ends any in-progress trim. * -d Secure trim. Requires kernel and device support. * -r Sets the TRIM rate in bytes (per second). Supports * adding a multiplier suffix such as 'k' or 'm'. * -s Suspend. TRIM can then be restarted with no flags. * -w Wait. Blocks until trimming has completed. */ int zpool_do_trim(int argc, char **argv) { struct option long_options[] = { {"cancel", no_argument, NULL, 'c'}, {"secure", no_argument, NULL, 'd'}, {"rate", required_argument, NULL, 'r'}, {"suspend", no_argument, NULL, 's'}, {"wait", no_argument, NULL, 'w'}, {0, 0, 0, 0} }; pool_trim_func_t cmd_type = POOL_TRIM_START; uint64_t rate = 0; boolean_t secure = B_FALSE; boolean_t wait = B_FALSE; int c; while ((c = getopt_long(argc, argv, "cdr:sw", long_options, NULL)) != -1) { switch (c) { case 'c': if (cmd_type != POOL_TRIM_START && cmd_type != POOL_TRIM_CANCEL) { (void) fprintf(stderr, gettext("-c cannot be " "combined with other options\n")); usage(B_FALSE); } cmd_type = POOL_TRIM_CANCEL; break; case 'd': if (cmd_type != POOL_TRIM_START) { (void) fprintf(stderr, gettext("-d cannot be " "combined with the -c or -s options\n")); usage(B_FALSE); } secure = B_TRUE; break; case 'r': if (cmd_type != POOL_TRIM_START) { (void) fprintf(stderr, gettext("-r cannot be " "combined with the -c or -s options\n")); usage(B_FALSE); } if (zfs_nicestrtonum(NULL, optarg, &rate) == -1) { (void) fprintf(stderr, gettext("invalid value for rate\n")); usage(B_FALSE); } break; case 's': if (cmd_type != POOL_TRIM_START && cmd_type != POOL_TRIM_SUSPEND) { (void) fprintf(stderr, gettext("-s cannot be " "combined with other options\n")); usage(B_FALSE); } cmd_type = POOL_TRIM_SUSPEND; break; case 'w': wait = B_TRUE; break; case '?': if (optopt != 0) { (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); } else { (void) fprintf(stderr, gettext("invalid option '%s'\n"), argv[optind - 1]); } usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); return (-1); } if (wait && (cmd_type != POOL_TRIM_START)) { (void) fprintf(stderr, gettext("-w cannot be used with -c or " "-s\n")); usage(B_FALSE); } char *poolname = argv[0]; zpool_handle_t *zhp = zpool_open(g_zfs, poolname); if (zhp == NULL) return (-1); trimflags_t trim_flags = { .secure = secure, .rate = rate, .wait = wait, }; nvlist_t *vdevs = fnvlist_alloc(); if (argc == 1) { /* no individual leaf vdevs specified, so add them all */ nvlist_t *config = zpool_get_config(zhp, NULL); nvlist_t *nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); zpool_collect_leaves(zhp, nvroot, vdevs); trim_flags.fullpool = B_TRUE; } else { trim_flags.fullpool = B_FALSE; for (int i = 1; i < argc; i++) { fnvlist_add_boolean(vdevs, argv[i]); } } int error = zpool_trim(zhp, cmd_type, vdevs, &trim_flags); fnvlist_free(vdevs); zpool_close(zhp); return (error); } /* * Converts a total number of seconds to a human readable string broken * down in to days/hours/minutes/seconds. */ static void secs_to_dhms(uint64_t total, char *buf) { uint64_t days = total / 60 / 60 / 24; uint64_t hours = (total / 60 / 60) % 24; uint64_t mins = (total / 60) % 60; uint64_t secs = (total % 60); if (days > 0) { (void) sprintf(buf, "%llu days %02llu:%02llu:%02llu", (u_longlong_t)days, (u_longlong_t)hours, (u_longlong_t)mins, (u_longlong_t)secs); } else { (void) sprintf(buf, "%02llu:%02llu:%02llu", (u_longlong_t)hours, (u_longlong_t)mins, (u_longlong_t)secs); } } /* * Print out detailed scrub status. */ static void print_scan_scrub_resilver_status(pool_scan_stat_t *ps) { time_t start, end, pause; uint64_t pass_scanned, scanned, pass_issued, issued, total; uint64_t elapsed, scan_rate, issue_rate; double fraction_done; char processed_buf[7], scanned_buf[7], issued_buf[7], total_buf[7]; char srate_buf[7], irate_buf[7], time_buf[32]; printf(" "); printf_color(ANSI_BOLD, gettext("scan:")); printf(" "); /* If there's never been a scan, there's not much to say. */ if (ps == NULL || ps->pss_func == POOL_SCAN_NONE || ps->pss_func >= POOL_SCAN_FUNCS) { (void) printf(gettext("none requested\n")); return; } start = ps->pss_start_time; end = ps->pss_end_time; pause = ps->pss_pass_scrub_pause; zfs_nicebytes(ps->pss_processed, processed_buf, sizeof (processed_buf)); - assert(ps->pss_func == POOL_SCAN_SCRUB || - ps->pss_func == POOL_SCAN_RESILVER); + int is_resilver = ps->pss_func == POOL_SCAN_RESILVER; + int is_scrub = ps->pss_func == POOL_SCAN_SCRUB; + assert(is_resilver || is_scrub); /* Scan is finished or canceled. */ if (ps->pss_state == DSS_FINISHED) { secs_to_dhms(end - start, time_buf); - if (ps->pss_func == POOL_SCAN_SCRUB) { + if (is_scrub) { (void) printf(gettext("scrub repaired %s " "in %s with %llu errors on %s"), processed_buf, time_buf, (u_longlong_t)ps->pss_errors, ctime(&end)); - } else if (ps->pss_func == POOL_SCAN_RESILVER) { + } else if (is_resilver) { (void) printf(gettext("resilvered %s " "in %s with %llu errors on %s"), processed_buf, time_buf, (u_longlong_t)ps->pss_errors, ctime(&end)); } return; } else if (ps->pss_state == DSS_CANCELED) { - if (ps->pss_func == POOL_SCAN_SCRUB) { + if (is_scrub) { (void) printf(gettext("scrub canceled on %s"), ctime(&end)); - } else if (ps->pss_func == POOL_SCAN_RESILVER) { + } else if (is_resilver) { (void) printf(gettext("resilver canceled on %s"), ctime(&end)); } return; } assert(ps->pss_state == DSS_SCANNING); /* Scan is in progress. Resilvers can't be paused. */ - if (ps->pss_func == POOL_SCAN_SCRUB) { + if (is_scrub) { if (pause == 0) { (void) printf(gettext("scrub in progress since %s"), ctime(&start)); } else { (void) printf(gettext("scrub paused since %s"), ctime(&pause)); (void) printf(gettext("\tscrub started on %s"), ctime(&start)); } - } else if (ps->pss_func == POOL_SCAN_RESILVER) { + } else if (is_resilver) { (void) printf(gettext("resilver in progress since %s"), ctime(&start)); } scanned = ps->pss_examined; pass_scanned = ps->pss_pass_exam; issued = ps->pss_issued; pass_issued = ps->pss_pass_issued; total = ps->pss_to_examine; /* we are only done with a block once we have issued the IO for it */ fraction_done = (double)issued / total; /* elapsed time for this pass, rounding up to 1 if it's 0 */ elapsed = time(NULL) - ps->pss_pass_start; elapsed -= ps->pss_pass_scrub_spent_paused; elapsed = (elapsed != 0) ? elapsed : 1; scan_rate = pass_scanned / elapsed; issue_rate = pass_issued / elapsed; uint64_t total_secs_left = (issue_rate != 0 && total >= issued) ? ((total - issued) / issue_rate) : UINT64_MAX; secs_to_dhms(total_secs_left, time_buf); /* format all of the numbers we will be reporting */ zfs_nicebytes(scanned, scanned_buf, sizeof (scanned_buf)); zfs_nicebytes(issued, issued_buf, sizeof (issued_buf)); zfs_nicebytes(total, total_buf, sizeof (total_buf)); zfs_nicebytes(scan_rate, srate_buf, sizeof (srate_buf)); zfs_nicebytes(issue_rate, irate_buf, sizeof (irate_buf)); /* do not print estimated time if we have a paused scrub */ if (pause == 0) { (void) printf(gettext("\t%s scanned at %s/s, " "%s issued at %s/s, %s total\n"), scanned_buf, srate_buf, issued_buf, irate_buf, total_buf); } else { (void) printf(gettext("\t%s scanned, %s issued, %s total\n"), scanned_buf, issued_buf, total_buf); } - if (ps->pss_func == POOL_SCAN_RESILVER) { + if (is_resilver) { (void) printf(gettext("\t%s resilvered, %.2f%% done"), processed_buf, 100 * fraction_done); - } else if (ps->pss_func == POOL_SCAN_SCRUB) { + } else if (is_scrub) { (void) printf(gettext("\t%s repaired, %.2f%% done"), processed_buf, 100 * fraction_done); } if (pause == 0) { + /* + * Only provide an estimate iff: + * 1) the time remaining is valid, and + * 2) the issue rate exceeds 10 MB/s, and + * 3) it's either: + * a) a resilver which has started repairs, or + * b) a scrub which has entered the issue phase. + */ if (total_secs_left != UINT64_MAX && - issue_rate >= 10 * 1024 * 1024) { + issue_rate >= 10 * 1024 * 1024 && + ((is_resilver && ps->pss_processed > 0) || + (is_scrub && issued > 0))) { (void) printf(gettext(", %s to go\n"), time_buf); } else { (void) printf(gettext(", no estimated " "completion time\n")); } } else { (void) printf(gettext("\n")); } } static void print_rebuild_status_impl(vdev_rebuild_stat_t *vrs, char *vdev_name) { if (vrs == NULL || vrs->vrs_state == VDEV_REBUILD_NONE) return; printf(" "); printf_color(ANSI_BOLD, gettext("scan:")); printf(" "); uint64_t bytes_scanned = vrs->vrs_bytes_scanned; uint64_t bytes_issued = vrs->vrs_bytes_issued; uint64_t bytes_rebuilt = vrs->vrs_bytes_rebuilt; uint64_t bytes_est = vrs->vrs_bytes_est; uint64_t scan_rate = (vrs->vrs_pass_bytes_scanned / (vrs->vrs_pass_time_ms + 1)) * 1000; uint64_t issue_rate = (vrs->vrs_pass_bytes_issued / (vrs->vrs_pass_time_ms + 1)) * 1000; double scan_pct = MIN((double)bytes_scanned * 100 / (bytes_est + 1), 100); /* Format all of the numbers we will be reporting */ char bytes_scanned_buf[7], bytes_issued_buf[7]; char bytes_rebuilt_buf[7], bytes_est_buf[7]; char scan_rate_buf[7], issue_rate_buf[7], time_buf[32]; zfs_nicebytes(bytes_scanned, bytes_scanned_buf, sizeof (bytes_scanned_buf)); zfs_nicebytes(bytes_issued, bytes_issued_buf, sizeof (bytes_issued_buf)); zfs_nicebytes(bytes_rebuilt, bytes_rebuilt_buf, sizeof (bytes_rebuilt_buf)); zfs_nicebytes(bytes_est, bytes_est_buf, sizeof (bytes_est_buf)); zfs_nicebytes(scan_rate, scan_rate_buf, sizeof (scan_rate_buf)); zfs_nicebytes(issue_rate, issue_rate_buf, sizeof (issue_rate_buf)); time_t start = vrs->vrs_start_time; time_t end = vrs->vrs_end_time; /* Rebuild is finished or canceled. */ if (vrs->vrs_state == VDEV_REBUILD_COMPLETE) { secs_to_dhms(vrs->vrs_scan_time_ms / 1000, time_buf); (void) printf(gettext("resilvered (%s) %s in %s " "with %llu errors on %s"), vdev_name, bytes_rebuilt_buf, time_buf, (u_longlong_t)vrs->vrs_errors, ctime(&end)); return; } else if (vrs->vrs_state == VDEV_REBUILD_CANCELED) { (void) printf(gettext("resilver (%s) canceled on %s"), vdev_name, ctime(&end)); return; } else if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) { (void) printf(gettext("resilver (%s) in progress since %s"), vdev_name, ctime(&start)); } assert(vrs->vrs_state == VDEV_REBUILD_ACTIVE); secs_to_dhms(MAX((int64_t)bytes_est - (int64_t)bytes_scanned, 0) / MAX(scan_rate, 1), time_buf); (void) printf(gettext("\t%s scanned at %s/s, %s issued %s/s, " "%s total\n"), bytes_scanned_buf, scan_rate_buf, bytes_issued_buf, issue_rate_buf, bytes_est_buf); (void) printf(gettext("\t%s resilvered, %.2f%% done"), bytes_rebuilt_buf, scan_pct); if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) { if (scan_rate >= 10 * 1024 * 1024) { (void) printf(gettext(", %s to go\n"), time_buf); } else { (void) printf(gettext(", no estimated " "completion time\n")); } } else { (void) printf(gettext("\n")); } } /* * Print rebuild status for top-level vdevs. */ static void print_rebuild_status(zpool_handle_t *zhp, nvlist_t *nvroot) { nvlist_t **child; uint_t children; if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (uint_t c = 0; c < children; c++) { vdev_rebuild_stat_t *vrs; uint_t i; if (nvlist_lookup_uint64_array(child[c], ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i) == 0) { char *name = zpool_vdev_name(g_zfs, zhp, child[c], VDEV_NAME_TYPE_ID); print_rebuild_status_impl(vrs, name); free(name); } } } /* * As we don't scrub checkpointed blocks, we want to warn the user that we * skipped scanning some blocks if a checkpoint exists or existed at any * time during the scan. If a sequential instead of healing reconstruction * was performed then the blocks were reconstructed. However, their checksums * have not been verified so we still print the warning. */ static void print_checkpoint_scan_warning(pool_scan_stat_t *ps, pool_checkpoint_stat_t *pcs) { if (ps == NULL || pcs == NULL) return; if (pcs->pcs_state == CS_NONE || pcs->pcs_state == CS_CHECKPOINT_DISCARDING) return; assert(pcs->pcs_state == CS_CHECKPOINT_EXISTS); if (ps->pss_state == DSS_NONE) return; if ((ps->pss_state == DSS_FINISHED || ps->pss_state == DSS_CANCELED) && ps->pss_end_time < pcs->pcs_start_time) return; if (ps->pss_state == DSS_FINISHED || ps->pss_state == DSS_CANCELED) { (void) printf(gettext(" scan warning: skipped blocks " "that are only referenced by the checkpoint.\n")); } else { assert(ps->pss_state == DSS_SCANNING); (void) printf(gettext(" scan warning: skipping blocks " "that are only referenced by the checkpoint.\n")); } } /* * Returns B_TRUE if there is an active rebuild in progress. Otherwise, * B_FALSE is returned and 'rebuild_end_time' is set to the end time for * the last completed (or cancelled) rebuild. */ static boolean_t check_rebuilding(nvlist_t *nvroot, uint64_t *rebuild_end_time) { nvlist_t **child; uint_t children; boolean_t rebuilding = B_FALSE; uint64_t end_time = 0; if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (uint_t c = 0; c < children; c++) { vdev_rebuild_stat_t *vrs; uint_t i; if (nvlist_lookup_uint64_array(child[c], ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i) == 0) { if (vrs->vrs_end_time > end_time) end_time = vrs->vrs_end_time; if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) { rebuilding = B_TRUE; end_time = 0; break; } } } if (rebuild_end_time != NULL) *rebuild_end_time = end_time; return (rebuilding); } /* * Print the scan status. */ static void print_scan_status(zpool_handle_t *zhp, nvlist_t *nvroot) { uint64_t rebuild_end_time = 0, resilver_end_time = 0; boolean_t have_resilver = B_FALSE, have_scrub = B_FALSE; boolean_t active_resilver = B_FALSE; pool_checkpoint_stat_t *pcs = NULL; pool_scan_stat_t *ps = NULL; uint_t c; if (nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &c) == 0) { if (ps->pss_func == POOL_SCAN_RESILVER) { resilver_end_time = ps->pss_end_time; active_resilver = (ps->pss_state == DSS_SCANNING); } have_resilver = (ps->pss_func == POOL_SCAN_RESILVER); have_scrub = (ps->pss_func == POOL_SCAN_SCRUB); } boolean_t active_rebuild = check_rebuilding(nvroot, &rebuild_end_time); boolean_t have_rebuild = (active_rebuild || (rebuild_end_time > 0)); /* Always print the scrub status when available. */ if (have_scrub) print_scan_scrub_resilver_status(ps); /* * When there is an active resilver or rebuild print its status. * Otherwise print the status of the last resilver or rebuild. */ if (active_resilver || (!active_rebuild && have_resilver && resilver_end_time && resilver_end_time > rebuild_end_time)) { print_scan_scrub_resilver_status(ps); } else if (active_rebuild || (!active_resilver && have_rebuild && rebuild_end_time && rebuild_end_time > resilver_end_time)) { print_rebuild_status(zhp, nvroot); } (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c); print_checkpoint_scan_warning(ps, pcs); } /* * Print out detailed removal status. */ static void print_removal_status(zpool_handle_t *zhp, pool_removal_stat_t *prs) { char copied_buf[7], examined_buf[7], total_buf[7], rate_buf[7]; time_t start, end; nvlist_t *config, *nvroot; nvlist_t **child; uint_t children; char *vdev_name; if (prs == NULL || prs->prs_state == DSS_NONE) return; /* * Determine name of vdev. */ config = zpool_get_config(zhp, NULL); nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); verify(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0); assert(prs->prs_removing_vdev < children); vdev_name = zpool_vdev_name(g_zfs, zhp, child[prs->prs_removing_vdev], B_TRUE); printf_color(ANSI_BOLD, gettext("remove: ")); start = prs->prs_start_time; end = prs->prs_end_time; zfs_nicenum(prs->prs_copied, copied_buf, sizeof (copied_buf)); /* * Removal is finished or canceled. */ if (prs->prs_state == DSS_FINISHED) { uint64_t minutes_taken = (end - start) / 60; (void) printf(gettext("Removal of vdev %llu copied %s " "in %lluh%um, completed on %s"), (longlong_t)prs->prs_removing_vdev, copied_buf, (u_longlong_t)(minutes_taken / 60), (uint_t)(minutes_taken % 60), ctime((time_t *)&end)); } else if (prs->prs_state == DSS_CANCELED) { (void) printf(gettext("Removal of %s canceled on %s"), vdev_name, ctime(&end)); } else { uint64_t copied, total, elapsed, mins_left, hours_left; double fraction_done; uint_t rate; assert(prs->prs_state == DSS_SCANNING); /* * Removal is in progress. */ (void) printf(gettext( "Evacuation of %s in progress since %s"), vdev_name, ctime(&start)); copied = prs->prs_copied > 0 ? prs->prs_copied : 1; total = prs->prs_to_copy; fraction_done = (double)copied / total; /* elapsed time for this pass */ elapsed = time(NULL) - prs->prs_start_time; elapsed = elapsed > 0 ? elapsed : 1; rate = copied / elapsed; rate = rate > 0 ? rate : 1; mins_left = ((total - copied) / rate) / 60; hours_left = mins_left / 60; zfs_nicenum(copied, examined_buf, sizeof (examined_buf)); zfs_nicenum(total, total_buf, sizeof (total_buf)); zfs_nicenum(rate, rate_buf, sizeof (rate_buf)); /* * do not print estimated time if hours_left is more than * 30 days */ (void) printf(gettext( "\t%s copied out of %s at %s/s, %.2f%% done"), examined_buf, total_buf, rate_buf, 100 * fraction_done); if (hours_left < (30 * 24)) { (void) printf(gettext(", %lluh%um to go\n"), (u_longlong_t)hours_left, (uint_t)(mins_left % 60)); } else { (void) printf(gettext( ", (copy is slow, no estimated time)\n")); } } free(vdev_name); if (prs->prs_mapping_memory > 0) { char mem_buf[7]; zfs_nicenum(prs->prs_mapping_memory, mem_buf, sizeof (mem_buf)); (void) printf(gettext( "\t%s memory used for removed device mappings\n"), mem_buf); } } static void print_checkpoint_status(pool_checkpoint_stat_t *pcs) { time_t start; char space_buf[7]; if (pcs == NULL || pcs->pcs_state == CS_NONE) return; (void) printf(gettext("checkpoint: ")); start = pcs->pcs_start_time; zfs_nicenum(pcs->pcs_space, space_buf, sizeof (space_buf)); if (pcs->pcs_state == CS_CHECKPOINT_EXISTS) { char *date = ctime(&start); /* * ctime() adds a newline at the end of the generated * string, thus the weird format specifier and the * strlen() call used to chop it off from the output. */ (void) printf(gettext("created %.*s, consumes %s\n"), (int)(strlen(date) - 1), date, space_buf); return; } assert(pcs->pcs_state == CS_CHECKPOINT_DISCARDING); (void) printf(gettext("discarding, %s remaining.\n"), space_buf); } static void print_error_log(zpool_handle_t *zhp) { nvlist_t *nverrlist = NULL; nvpair_t *elem; char *pathname; size_t len = MAXPATHLEN * 2; if (zpool_get_errlog(zhp, &nverrlist) != 0) return; (void) printf("errors: Permanent errors have been " "detected in the following files:\n\n"); pathname = safe_malloc(len); elem = NULL; while ((elem = nvlist_next_nvpair(nverrlist, elem)) != NULL) { nvlist_t *nv; uint64_t dsobj, obj; verify(nvpair_value_nvlist(elem, &nv) == 0); verify(nvlist_lookup_uint64(nv, ZPOOL_ERR_DATASET, &dsobj) == 0); verify(nvlist_lookup_uint64(nv, ZPOOL_ERR_OBJECT, &obj) == 0); zpool_obj_to_path(zhp, dsobj, obj, pathname, len); (void) printf("%7s %s\n", "", pathname); } free(pathname); nvlist_free(nverrlist); } static void print_spares(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t **spares, uint_t nspares) { uint_t i; char *name; if (nspares == 0) return; (void) printf(gettext("\tspares\n")); for (i = 0; i < nspares; i++) { name = zpool_vdev_name(g_zfs, zhp, spares[i], cb->cb_name_flags); print_status_config(zhp, cb, name, spares[i], 2, B_TRUE, NULL); free(name); } } static void print_l2cache(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t **l2cache, uint_t nl2cache) { uint_t i; char *name; if (nl2cache == 0) return; (void) printf(gettext("\tcache\n")); for (i = 0; i < nl2cache; i++) { name = zpool_vdev_name(g_zfs, zhp, l2cache[i], cb->cb_name_flags); print_status_config(zhp, cb, name, l2cache[i], 2, B_FALSE, NULL); free(name); } } static void print_dedup_stats(nvlist_t *config) { ddt_histogram_t *ddh; ddt_stat_t *dds; ddt_object_t *ddo; uint_t c; char dspace[6], mspace[6]; /* * If the pool was faulted then we may not have been able to * obtain the config. Otherwise, if we have anything in the dedup * table continue processing the stats. */ if (nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_OBJ_STATS, (uint64_t **)&ddo, &c) != 0) return; (void) printf("\n"); (void) printf(gettext(" dedup: ")); if (ddo->ddo_count == 0) { (void) printf(gettext("no DDT entries\n")); return; } zfs_nicebytes(ddo->ddo_dspace, dspace, sizeof (dspace)); zfs_nicebytes(ddo->ddo_mspace, mspace, sizeof (mspace)); (void) printf("DDT entries %llu, size %s on disk, %s in core\n", (u_longlong_t)ddo->ddo_count, dspace, mspace); verify(nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_STATS, (uint64_t **)&dds, &c) == 0); verify(nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_HISTOGRAM, (uint64_t **)&ddh, &c) == 0); zpool_dump_ddt(dds, ddh); } /* * Display a summary of pool status. Displays a summary such as: * * pool: tank * status: DEGRADED * reason: One or more devices ... * see: https://openzfs.github.io/openzfs-docs/msg/ZFS-xxxx-01 * config: * mirror DEGRADED * c1t0d0 OK * c2t0d0 UNAVAIL * * When given the '-v' option, we print out the complete config. If the '-e' * option is specified, then we print out error rate information as well. */ static int status_callback(zpool_handle_t *zhp, void *data) { status_cbdata_t *cbp = data; nvlist_t *config, *nvroot; char *msgid; zpool_status_t reason; zpool_errata_t errata; const char *health; uint_t c; vdev_stat_t *vs; config = zpool_get_config(zhp, NULL); reason = zpool_get_status(zhp, &msgid, &errata); cbp->cb_count++; /* * If we were given 'zpool status -x', only report those pools with * problems. */ if (cbp->cb_explain && (reason == ZPOOL_STATUS_OK || reason == ZPOOL_STATUS_VERSION_OLDER || reason == ZPOOL_STATUS_FEAT_DISABLED || reason == ZPOOL_STATUS_COMPATIBILITY_ERR || reason == ZPOOL_STATUS_INCOMPATIBLE_FEAT)) { if (!cbp->cb_allpools) { (void) printf(gettext("pool '%s' is healthy\n"), zpool_get_name(zhp)); if (cbp->cb_first) cbp->cb_first = B_FALSE; } return (0); } if (cbp->cb_first) cbp->cb_first = B_FALSE; else (void) printf("\n"); nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c) == 0); health = zpool_get_state_str(zhp); printf(" "); printf_color(ANSI_BOLD, gettext("pool:")); printf(" %s\n", zpool_get_name(zhp)); printf(" "); printf_color(ANSI_BOLD, gettext("state: ")); printf_color(health_str_to_color(health), "%s", health); printf("\n"); switch (reason) { case ZPOOL_STATUS_MISSING_DEV_R: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices could " "not be opened. Sufficient replicas exist for\n\tthe pool " "to continue functioning in a degraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Attach the missing device " "and online it using 'zpool online'.\n")); break; case ZPOOL_STATUS_MISSING_DEV_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices could " "not be opened. There are insufficient\n\treplicas for the" " pool to continue functioning.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Attach the missing device " "and online it using 'zpool online'.\n")); break; case ZPOOL_STATUS_CORRUPT_LABEL_R: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices could " "not be used because the label is missing or\n\tinvalid. " "Sufficient replicas exist for the pool to continue\n\t" "functioning in a degraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Replace the device using " "'zpool replace'.\n")); break; case ZPOOL_STATUS_CORRUPT_LABEL_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices could " "not be used because the label is missing \n\tor invalid. " "There are insufficient replicas for the pool to " "continue\n\tfunctioning.\n")); zpool_explain_recover(zpool_get_handle(zhp), zpool_get_name(zhp), reason, config); break; case ZPOOL_STATUS_FAILING_DEV: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices has " "experienced an unrecoverable error. An\n\tattempt was " "made to correct the error. Applications are " "unaffected.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Determine if the " "device needs to be replaced, and clear the errors\n\tusing" " 'zpool clear' or replace the device with 'zpool " "replace'.\n")); break; case ZPOOL_STATUS_OFFLINE_DEV: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices has " "been taken offline by the administrator.\n\tSufficient " "replicas exist for the pool to continue functioning in " "a\n\tdegraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Online the device " "using 'zpool online' or replace the device with\n\t'zpool " "replace'.\n")); break; case ZPOOL_STATUS_REMOVED_DEV: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices has " "been removed by the administrator.\n\tSufficient " "replicas exist for the pool to continue functioning in " "a\n\tdegraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Online the device " "using zpool online' or replace the device with\n\t'zpool " "replace'.\n")); break; case ZPOOL_STATUS_RESILVERING: case ZPOOL_STATUS_REBUILDING: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices is " "currently being resilvered. The pool will\n\tcontinue " "to function, possibly in a degraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Wait for the resilver to " "complete.\n")); break; case ZPOOL_STATUS_REBUILD_SCRUB: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices have " "been sequentially resilvered, scrubbing\n\tthe pool " "is recommended.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Use 'zpool scrub' to " "verify all data checksums.\n")); break; case ZPOOL_STATUS_CORRUPT_DATA: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices has " "experienced an error resulting in data\n\tcorruption. " "Applications may be affected.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Restore the file in question" " if possible. Otherwise restore the\n\tentire pool from " "backup.\n")); break; case ZPOOL_STATUS_CORRUPT_POOL: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool metadata is " "corrupted and the pool cannot be opened.\n")); zpool_explain_recover(zpool_get_handle(zhp), zpool_get_name(zhp), reason, config); break; case ZPOOL_STATUS_VERSION_OLDER: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is formatted using " "a legacy on-disk format. The pool can\n\tstill be used, " "but some features are unavailable.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Upgrade the pool using " "'zpool upgrade'. Once this is done, the\n\tpool will no " "longer be accessible on software that does not support\n\t" "feature flags.\n")); break; case ZPOOL_STATUS_VERSION_NEWER: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool has been upgraded " "to a newer, incompatible on-disk version.\n\tThe pool " "cannot be accessed on this system.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Access the pool from a " "system running more recent software, or\n\trestore the " "pool from backup.\n")); break; case ZPOOL_STATUS_FEAT_DISABLED: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Some supported and " "requested features are not enabled on the pool.\n\t" "The pool can still be used, but some features are " "unavailable.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Enable all features using " "'zpool upgrade'. Once this is done,\n\tthe pool may no " "longer be accessible by software that does not support\n\t" "the features. See zpool-features(7) for details.\n")); break; case ZPOOL_STATUS_COMPATIBILITY_ERR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("This pool has a " "compatibility list specified, but it could not be\n\t" "read/parsed at this time. The pool can still be used, " "but this\n\tshould be investigated.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Check the value of the " "'compatibility' property against the\n\t" "appropriate file in " ZPOOL_SYSCONF_COMPAT_D " or " ZPOOL_DATA_COMPAT_D ".\n")); break; case ZPOOL_STATUS_INCOMPATIBLE_FEAT: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more features " "are enabled on the pool despite not being\n\t" "requested by the 'compatibility' property.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Consider setting " "'compatibility' to an appropriate value, or\n\t" "adding needed features to the relevant file in\n\t" ZPOOL_SYSCONF_COMPAT_D " or " ZPOOL_DATA_COMPAT_D ".\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_READ: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool cannot be accessed " "on this system because it uses the\n\tfollowing feature(s)" " not supported on this system:\n")); zpool_print_unsup_feat(config); (void) printf("\n"); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Access the pool from a " "system that supports the required feature(s),\n\tor " "restore the pool from backup.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_WRITE: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool can only be " "accessed in read-only mode on this system. It\n\tcannot be" " accessed in read-write mode because it uses the " "following\n\tfeature(s) not supported on this system:\n")); zpool_print_unsup_feat(config); (void) printf("\n"); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("The pool cannot be accessed " "in read-write mode. Import the pool with\n" "\t\"-o readonly=on\", access the pool from a system that " "supports the\n\trequired feature(s), or restore the " "pool from backup.\n")); break; case ZPOOL_STATUS_FAULTED_DEV_R: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "faulted in response to persistent errors.\n\tSufficient " "replicas exist for the pool to continue functioning " "in a\n\tdegraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Replace the faulted device, " "or use 'zpool clear' to mark the device\n\trepaired.\n")); break; case ZPOOL_STATUS_FAULTED_DEV_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "faulted in response to persistent errors. There are " "insufficient replicas for the pool to\n\tcontinue " "functioning.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Destroy and re-create the " "pool from a backup source. Manually marking the device\n" "\trepaired using 'zpool clear' may allow some data " "to be recovered.\n")); break; case ZPOOL_STATUS_IO_FAILURE_MMP: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is suspended " "because multihost writes failed or were delayed;\n\t" "another system could import the pool undetected.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Make sure the pool's devices" " are connected, then reboot your system and\n\timport the " "pool.\n")); break; case ZPOOL_STATUS_IO_FAILURE_WAIT: case ZPOOL_STATUS_IO_FAILURE_CONTINUE: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "faulted in response to IO failures.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Make sure the affected " "devices are connected, then run 'zpool clear'.\n")); break; case ZPOOL_STATUS_BAD_LOG: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("An intent log record " "could not be read.\n" "\tWaiting for administrator intervention to fix the " "faulted pool.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Either restore the affected " "device(s) and run 'zpool online',\n" "\tor ignore the intent log records by running " "'zpool clear'.\n")); break; case ZPOOL_STATUS_NON_NATIVE_ASHIFT: (void) printf(gettext("status: One or more devices are " "configured to use a non-native block size.\n" "\tExpect reduced performance.\n")); (void) printf(gettext("action: Replace affected devices with " "devices that support the\n\tconfigured block size, or " "migrate data to a properly configured\n\tpool.\n")); break; case ZPOOL_STATUS_HOSTID_MISMATCH: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Mismatch between pool hostid" " and system hostid on imported pool.\n\tThis pool was " "previously imported into a system with a different " "hostid,\n\tand then was verbatim imported into this " "system.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Export this pool on all " "systems on which it is imported.\n" "\tThen import it to correct the mismatch.\n")); break; case ZPOOL_STATUS_ERRATA: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Errata #%d detected.\n"), errata); switch (errata) { case ZPOOL_ERRATA_NONE: break; case ZPOOL_ERRATA_ZOL_2094_SCRUB: printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("To correct the issue" " run 'zpool scrub'.\n")); break; case ZPOOL_ERRATA_ZOL_6845_ENCRYPTION: (void) printf(gettext("\tExisting encrypted datasets " "contain an on-disk incompatibility\n\twhich " "needs to be corrected.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("To correct the issue" " backup existing encrypted datasets to new\n\t" "encrypted datasets and destroy the old ones. " "'zfs mount -o ro' can\n\tbe used to temporarily " "mount existing encrypted datasets readonly.\n")); break; case ZPOOL_ERRATA_ZOL_8308_ENCRYPTION: (void) printf(gettext("\tExisting encrypted snapshots " "and bookmarks contain an on-disk\n\tincompat" "ibility. This may cause on-disk corruption if " "they are used\n\twith 'zfs recv'.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("To correct the" "issue, enable the bookmark_v2 feature. No " "additional\n\taction is needed if there are no " "encrypted snapshots or bookmarks.\n\tIf preserving" "the encrypted snapshots and bookmarks is required," " use\n\ta non-raw send to backup and restore them." " Alternately, they may be\n\tremoved to resolve " "the incompatibility.\n")); break; default: /* * All errata which allow the pool to be imported * must contain an action message. */ assert(0); } break; default: /* * The remaining errors can't actually be generated, yet. */ assert(reason == ZPOOL_STATUS_OK); } if (msgid != NULL) { printf(" "); printf_color(ANSI_BOLD, gettext("see:")); printf(gettext( " https://openzfs.github.io/openzfs-docs/msg/%s\n"), msgid); } if (config != NULL) { uint64_t nerr; nvlist_t **spares, **l2cache; uint_t nspares, nl2cache; pool_checkpoint_stat_t *pcs = NULL; pool_removal_stat_t *prs = NULL; print_scan_status(zhp, nvroot); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t **)&prs, &c); print_removal_status(zhp, prs); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c); print_checkpoint_status(pcs); cbp->cb_namewidth = max_width(zhp, nvroot, 0, 0, cbp->cb_name_flags | VDEV_NAME_TYPE_ID); if (cbp->cb_namewidth < 10) cbp->cb_namewidth = 10; color_start(ANSI_BOLD); (void) printf(gettext("config:\n\n")); (void) printf(gettext("\t%-*s %-8s %5s %5s %5s"), cbp->cb_namewidth, "NAME", "STATE", "READ", "WRITE", "CKSUM"); color_end(); if (cbp->cb_print_slow_ios) { printf_color(ANSI_BOLD, " %5s", gettext("SLOW")); } if (cbp->vcdl != NULL) print_cmd_columns(cbp->vcdl, 0); printf("\n"); print_status_config(zhp, cbp, zpool_get_name(zhp), nvroot, 0, B_FALSE, NULL); print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_BIAS_DEDUP); print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_BIAS_SPECIAL); print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_CLASS_LOGS); if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0) print_l2cache(zhp, cbp, l2cache, nl2cache); if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0) print_spares(zhp, cbp, spares, nspares); if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT, &nerr) == 0) { nvlist_t *nverrlist = NULL; /* * If the approximate error count is small, get a * precise count by fetching the entire log and * uniquifying the results. */ if (nerr > 0 && nerr < 100 && !cbp->cb_verbose && zpool_get_errlog(zhp, &nverrlist) == 0) { nvpair_t *elem; elem = NULL; nerr = 0; while ((elem = nvlist_next_nvpair(nverrlist, elem)) != NULL) { nerr++; } } nvlist_free(nverrlist); (void) printf("\n"); if (nerr == 0) (void) printf(gettext("errors: No known data " "errors\n")); else if (!cbp->cb_verbose) (void) printf(gettext("errors: %llu data " "errors, use '-v' for a list\n"), (u_longlong_t)nerr); else print_error_log(zhp); } if (cbp->cb_dedup_stats) print_dedup_stats(config); } else { (void) printf(gettext("config: The configuration cannot be " "determined.\n")); } return (0); } /* * zpool status [-c [script1,script2,...]] [-igLpPstvx] [-T d|u] [pool] ... * [interval [count]] * * -c CMD For each vdev, run command CMD * -i Display vdev initialization status. * -g Display guid for individual vdev name. * -L Follow links when resolving vdev path name. * -p Display values in parsable (exact) format. * -P Display full path for vdev name. * -s Display slow IOs column. * -v Display complete error logs * -x Display only pools with potential problems * -D Display dedup status (undocumented) * -t Display vdev TRIM status. * -T Display a timestamp in date(1) or Unix format * * Describes the health status of all pools or some subset. */ int zpool_do_status(int argc, char **argv) { int c; int ret; float interval = 0; unsigned long count = 0; status_cbdata_t cb = { 0 }; char *cmd = NULL; /* check options */ while ((c = getopt(argc, argv, "c:igLpPsvxDtT:")) != -1) { switch (c) { case 'c': if (cmd != NULL) { fprintf(stderr, gettext("Can't set -c flag twice\n")); exit(1); } if (getenv("ZPOOL_SCRIPTS_ENABLED") != NULL && !libzfs_envvar_is_set("ZPOOL_SCRIPTS_ENABLED")) { fprintf(stderr, gettext( "Can't run -c, disabled by " "ZPOOL_SCRIPTS_ENABLED.\n")); exit(1); } if ((getuid() <= 0 || geteuid() <= 0) && !libzfs_envvar_is_set("ZPOOL_SCRIPTS_AS_ROOT")) { fprintf(stderr, gettext( "Can't run -c with root privileges " "unless ZPOOL_SCRIPTS_AS_ROOT is set.\n")); exit(1); } cmd = optarg; break; case 'i': cb.cb_print_vdev_init = B_TRUE; break; case 'g': cb.cb_name_flags |= VDEV_NAME_GUID; break; case 'L': cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS; break; case 'p': cb.cb_literal = B_TRUE; break; case 'P': cb.cb_name_flags |= VDEV_NAME_PATH; break; case 's': cb.cb_print_slow_ios = B_TRUE; break; case 'v': cb.cb_verbose = B_TRUE; break; case 'x': cb.cb_explain = B_TRUE; break; case 'D': cb.cb_dedup_stats = B_TRUE; break; case 't': cb.cb_print_vdev_trim = B_TRUE; break; case 'T': get_timestamp_arg(*optarg); break; case '?': if (optopt == 'c') { print_zpool_script_list("status"); exit(0); } else { fprintf(stderr, gettext("invalid option '%c'\n"), optopt); } usage(B_FALSE); } } argc -= optind; argv += optind; get_interval_count(&argc, argv, &interval, &count); if (argc == 0) cb.cb_allpools = B_TRUE; cb.cb_first = B_TRUE; cb.cb_print_status = B_TRUE; for (;;) { if (timestamp_fmt != NODATE) print_timestamp(timestamp_fmt); if (cmd != NULL) cb.vcdl = all_pools_for_each_vdev_run(argc, argv, cmd, NULL, NULL, 0, 0); ret = for_each_pool(argc, argv, B_TRUE, NULL, cb.cb_literal, status_callback, &cb); if (cb.vcdl != NULL) free_vdev_cmd_data_list(cb.vcdl); if (argc == 0 && cb.cb_count == 0) (void) fprintf(stderr, gettext("no pools available\n")); else if (cb.cb_explain && cb.cb_first && cb.cb_allpools) (void) printf(gettext("all pools are healthy\n")); if (ret != 0) return (ret); if (interval == 0) break; if (count != 0 && --count == 0) break; (void) fsleep(interval); } return (0); } typedef struct upgrade_cbdata { int cb_first; int cb_argc; uint64_t cb_version; char **cb_argv; } upgrade_cbdata_t; static int check_unsupp_fs(zfs_handle_t *zhp, void *unsupp_fs) { int zfs_version = (int)zfs_prop_get_int(zhp, ZFS_PROP_VERSION); int *count = (int *)unsupp_fs; if (zfs_version > ZPL_VERSION) { (void) printf(gettext("%s (v%d) is not supported by this " "implementation of ZFS.\n"), zfs_get_name(zhp), zfs_version); (*count)++; } zfs_iter_filesystems(zhp, check_unsupp_fs, unsupp_fs); zfs_close(zhp); return (0); } static int upgrade_version(zpool_handle_t *zhp, uint64_t version) { int ret; nvlist_t *config; uint64_t oldversion; int unsupp_fs = 0; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &oldversion) == 0); char compat[ZFS_MAXPROPLEN]; if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat, ZFS_MAXPROPLEN, NULL, B_FALSE) != 0) compat[0] = '\0'; assert(SPA_VERSION_IS_SUPPORTED(oldversion)); assert(oldversion < version); ret = zfs_iter_root(zpool_get_handle(zhp), check_unsupp_fs, &unsupp_fs); if (ret != 0) return (ret); if (unsupp_fs) { (void) fprintf(stderr, gettext("Upgrade not performed due " "to %d unsupported filesystems (max v%d).\n"), unsupp_fs, (int)ZPL_VERSION); return (1); } if (strcmp(compat, ZPOOL_COMPAT_LEGACY) == 0) { (void) fprintf(stderr, gettext("Upgrade not performed because " "'compatibility' property set to '" ZPOOL_COMPAT_LEGACY "'.\n")); return (1); } ret = zpool_upgrade(zhp, version); if (ret != 0) return (ret); if (version >= SPA_VERSION_FEATURES) { (void) printf(gettext("Successfully upgraded " "'%s' from version %llu to feature flags.\n"), zpool_get_name(zhp), (u_longlong_t)oldversion); } else { (void) printf(gettext("Successfully upgraded " "'%s' from version %llu to version %llu.\n"), zpool_get_name(zhp), (u_longlong_t)oldversion, (u_longlong_t)version); } return (0); } static int upgrade_enable_all(zpool_handle_t *zhp, int *countp) { int i, ret, count; boolean_t firstff = B_TRUE; nvlist_t *enabled = zpool_get_features(zhp); char compat[ZFS_MAXPROPLEN]; if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat, ZFS_MAXPROPLEN, NULL, B_FALSE) != 0) compat[0] = '\0'; boolean_t requested_features[SPA_FEATURES]; if (zpool_do_load_compat(compat, requested_features) != ZPOOL_COMPATIBILITY_OK) return (-1); count = 0; for (i = 0; i < SPA_FEATURES; i++) { const char *fname = spa_feature_table[i].fi_uname; const char *fguid = spa_feature_table[i].fi_guid; if (!spa_feature_table[i].fi_zfs_mod_supported) continue; if (!nvlist_exists(enabled, fguid) && requested_features[i]) { char *propname; verify(-1 != asprintf(&propname, "feature@%s", fname)); ret = zpool_set_prop(zhp, propname, ZFS_FEATURE_ENABLED); if (ret != 0) { free(propname); return (ret); } count++; if (firstff) { (void) printf(gettext("Enabled the " "following features on '%s':\n"), zpool_get_name(zhp)); firstff = B_FALSE; } (void) printf(gettext(" %s\n"), fname); free(propname); } } if (countp != NULL) *countp = count; return (0); } static int upgrade_cb(zpool_handle_t *zhp, void *arg) { upgrade_cbdata_t *cbp = arg; nvlist_t *config; uint64_t version; boolean_t modified_pool = B_FALSE; int ret; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &version) == 0); assert(SPA_VERSION_IS_SUPPORTED(version)); if (version < cbp->cb_version) { cbp->cb_first = B_FALSE; ret = upgrade_version(zhp, cbp->cb_version); if (ret != 0) return (ret); modified_pool = B_TRUE; /* * If they did "zpool upgrade -a", then we could * be doing ioctls to different pools. We need * to log this history once to each pool, and bypass * the normal history logging that happens in main(). */ (void) zpool_log_history(g_zfs, history_str); log_history = B_FALSE; } if (cbp->cb_version >= SPA_VERSION_FEATURES) { int count; ret = upgrade_enable_all(zhp, &count); if (ret != 0) return (ret); if (count > 0) { cbp->cb_first = B_FALSE; modified_pool = B_TRUE; } } if (modified_pool) { (void) printf("\n"); (void) after_zpool_upgrade(zhp); } return (0); } static int upgrade_list_older_cb(zpool_handle_t *zhp, void *arg) { upgrade_cbdata_t *cbp = arg; nvlist_t *config; uint64_t version; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &version) == 0); assert(SPA_VERSION_IS_SUPPORTED(version)); if (version < SPA_VERSION_FEATURES) { if (cbp->cb_first) { (void) printf(gettext("The following pools are " "formatted with legacy version numbers and can\n" "be upgraded to use feature flags. After " "being upgraded, these pools\nwill no " "longer be accessible by software that does not " "support feature\nflags.\n\n" "Note that setting a pool's 'compatibility' " "feature to '" ZPOOL_COMPAT_LEGACY "' will\n" "inhibit upgrades.\n\n")); (void) printf(gettext("VER POOL\n")); (void) printf(gettext("--- ------------\n")); cbp->cb_first = B_FALSE; } (void) printf("%2llu %s\n", (u_longlong_t)version, zpool_get_name(zhp)); } return (0); } static int upgrade_list_disabled_cb(zpool_handle_t *zhp, void *arg) { upgrade_cbdata_t *cbp = arg; nvlist_t *config; uint64_t version; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &version) == 0); if (version >= SPA_VERSION_FEATURES) { int i; boolean_t poolfirst = B_TRUE; nvlist_t *enabled = zpool_get_features(zhp); for (i = 0; i < SPA_FEATURES; i++) { const char *fguid = spa_feature_table[i].fi_guid; const char *fname = spa_feature_table[i].fi_uname; if (!spa_feature_table[i].fi_zfs_mod_supported) continue; if (!nvlist_exists(enabled, fguid)) { if (cbp->cb_first) { (void) printf(gettext("\nSome " "supported features are not " "enabled on the following pools. " "Once a\nfeature is enabled the " "pool may become incompatible with " "software\nthat does not support " "the feature. See " "zpool-features(7) for " "details.\n\n" "Note that the pool " "'compatibility' feature can be " "used to inhibit\nfeature " "upgrades.\n\n")); (void) printf(gettext("POOL " "FEATURE\n")); (void) printf(gettext("------" "---------\n")); cbp->cb_first = B_FALSE; } if (poolfirst) { (void) printf(gettext("%s\n"), zpool_get_name(zhp)); poolfirst = B_FALSE; } (void) printf(gettext(" %s\n"), fname); } /* * If they did "zpool upgrade -a", then we could * be doing ioctls to different pools. We need * to log this history once to each pool, and bypass * the normal history logging that happens in main(). */ (void) zpool_log_history(g_zfs, history_str); log_history = B_FALSE; } } return (0); } /* ARGSUSED */ static int upgrade_one(zpool_handle_t *zhp, void *data) { boolean_t modified_pool = B_FALSE; upgrade_cbdata_t *cbp = data; uint64_t cur_version; int ret; if (strcmp("log", zpool_get_name(zhp)) == 0) { (void) fprintf(stderr, gettext("'log' is now a reserved word\n" "Pool 'log' must be renamed using export and import" " to upgrade.\n")); return (1); } cur_version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL); if (cur_version > cbp->cb_version) { (void) printf(gettext("Pool '%s' is already formatted " "using more current version '%llu'.\n\n"), zpool_get_name(zhp), (u_longlong_t)cur_version); return (0); } if (cbp->cb_version != SPA_VERSION && cur_version == cbp->cb_version) { (void) printf(gettext("Pool '%s' is already formatted " "using version %llu.\n\n"), zpool_get_name(zhp), (u_longlong_t)cbp->cb_version); return (0); } if (cur_version != cbp->cb_version) { modified_pool = B_TRUE; ret = upgrade_version(zhp, cbp->cb_version); if (ret != 0) return (ret); } if (cbp->cb_version >= SPA_VERSION_FEATURES) { int count = 0; ret = upgrade_enable_all(zhp, &count); if (ret != 0) return (ret); if (count != 0) { modified_pool = B_TRUE; } else if (cur_version == SPA_VERSION) { (void) printf(gettext("Pool '%s' already has all " "supported and requested features enabled.\n"), zpool_get_name(zhp)); } } if (modified_pool) { (void) printf("\n"); (void) after_zpool_upgrade(zhp); } return (0); } /* * zpool upgrade * zpool upgrade -v * zpool upgrade [-V version] <-a | pool ...> * * With no arguments, display downrev'd ZFS pool available for upgrade. * Individual pools can be upgraded by specifying the pool, and '-a' will * upgrade all pools. */ int zpool_do_upgrade(int argc, char **argv) { int c; upgrade_cbdata_t cb = { 0 }; int ret = 0; boolean_t showversions = B_FALSE; boolean_t upgradeall = B_FALSE; char *end; /* check options */ while ((c = getopt(argc, argv, ":avV:")) != -1) { switch (c) { case 'a': upgradeall = B_TRUE; break; case 'v': showversions = B_TRUE; break; case 'V': cb.cb_version = strtoll(optarg, &end, 10); if (*end != '\0' || !SPA_VERSION_IS_SUPPORTED(cb.cb_version)) { (void) fprintf(stderr, gettext("invalid version '%s'\n"), optarg); usage(B_FALSE); } break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); usage(B_FALSE); break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } cb.cb_argc = argc; cb.cb_argv = argv; argc -= optind; argv += optind; if (cb.cb_version == 0) { cb.cb_version = SPA_VERSION; } else if (!upgradeall && argc == 0) { (void) fprintf(stderr, gettext("-V option is " "incompatible with other arguments\n")); usage(B_FALSE); } if (showversions) { if (upgradeall || argc != 0) { (void) fprintf(stderr, gettext("-v option is " "incompatible with other arguments\n")); usage(B_FALSE); } } else if (upgradeall) { if (argc != 0) { (void) fprintf(stderr, gettext("-a option should not " "be used along with a pool name\n")); usage(B_FALSE); } } (void) printf(gettext("This system supports ZFS pool feature " "flags.\n\n")); if (showversions) { int i; (void) printf(gettext("The following features are " "supported:\n\n")); (void) printf(gettext("FEAT DESCRIPTION\n")); (void) printf("----------------------------------------------" "---------------\n"); for (i = 0; i < SPA_FEATURES; i++) { zfeature_info_t *fi = &spa_feature_table[i]; if (!fi->fi_zfs_mod_supported) continue; const char *ro = (fi->fi_flags & ZFEATURE_FLAG_READONLY_COMPAT) ? " (read-only compatible)" : ""; (void) printf("%-37s%s\n", fi->fi_uname, ro); (void) printf(" %s\n", fi->fi_desc); } (void) printf("\n"); (void) printf(gettext("The following legacy versions are also " "supported:\n\n")); (void) printf(gettext("VER DESCRIPTION\n")); (void) printf("--- -----------------------------------------" "---------------\n"); (void) printf(gettext(" 1 Initial ZFS version\n")); (void) printf(gettext(" 2 Ditto blocks " "(replicated metadata)\n")); (void) printf(gettext(" 3 Hot spares and double parity " "RAID-Z\n")); (void) printf(gettext(" 4 zpool history\n")); (void) printf(gettext(" 5 Compression using the gzip " "algorithm\n")); (void) printf(gettext(" 6 bootfs pool property\n")); (void) printf(gettext(" 7 Separate intent log devices\n")); (void) printf(gettext(" 8 Delegated administration\n")); (void) printf(gettext(" 9 refquota and refreservation " "properties\n")); (void) printf(gettext(" 10 Cache devices\n")); (void) printf(gettext(" 11 Improved scrub performance\n")); (void) printf(gettext(" 12 Snapshot properties\n")); (void) printf(gettext(" 13 snapused property\n")); (void) printf(gettext(" 14 passthrough-x aclinherit\n")); (void) printf(gettext(" 15 user/group space accounting\n")); (void) printf(gettext(" 16 stmf property support\n")); (void) printf(gettext(" 17 Triple-parity RAID-Z\n")); (void) printf(gettext(" 18 Snapshot user holds\n")); (void) printf(gettext(" 19 Log device removal\n")); (void) printf(gettext(" 20 Compression using zle " "(zero-length encoding)\n")); (void) printf(gettext(" 21 Deduplication\n")); (void) printf(gettext(" 22 Received properties\n")); (void) printf(gettext(" 23 Slim ZIL\n")); (void) printf(gettext(" 24 System attributes\n")); (void) printf(gettext(" 25 Improved scrub stats\n")); (void) printf(gettext(" 26 Improved snapshot deletion " "performance\n")); (void) printf(gettext(" 27 Improved snapshot creation " "performance\n")); (void) printf(gettext(" 28 Multiple vdev replacements\n")); (void) printf(gettext("\nFor more information on a particular " "version, including supported releases,\n")); (void) printf(gettext("see the ZFS Administration Guide.\n\n")); } else if (argc == 0 && upgradeall) { cb.cb_first = B_TRUE; ret = zpool_iter(g_zfs, upgrade_cb, &cb); if (ret == 0 && cb.cb_first) { if (cb.cb_version == SPA_VERSION) { (void) printf(gettext("All pools are already " "formatted using feature flags.\n\n")); (void) printf(gettext("Every feature flags " "pool already has all supported and " "requested features enabled.\n")); } else { (void) printf(gettext("All pools are already " "formatted with version %llu or higher.\n"), (u_longlong_t)cb.cb_version); } } } else if (argc == 0) { cb.cb_first = B_TRUE; ret = zpool_iter(g_zfs, upgrade_list_older_cb, &cb); assert(ret == 0); if (cb.cb_first) { (void) printf(gettext("All pools are formatted " "using feature flags.\n\n")); } else { (void) printf(gettext("\nUse 'zpool upgrade -v' " "for a list of available legacy versions.\n")); } cb.cb_first = B_TRUE; ret = zpool_iter(g_zfs, upgrade_list_disabled_cb, &cb); assert(ret == 0); if (cb.cb_first) { (void) printf(gettext("Every feature flags pool has " "all supported and requested features enabled.\n")); } else { (void) printf(gettext("\n")); } } else { ret = for_each_pool(argc, argv, B_FALSE, NULL, B_FALSE, upgrade_one, &cb); } return (ret); } typedef struct hist_cbdata { boolean_t first; boolean_t longfmt; boolean_t internal; } hist_cbdata_t; static void print_history_records(nvlist_t *nvhis, hist_cbdata_t *cb) { nvlist_t **records; uint_t numrecords; int i; verify(nvlist_lookup_nvlist_array(nvhis, ZPOOL_HIST_RECORD, &records, &numrecords) == 0); for (i = 0; i < numrecords; i++) { nvlist_t *rec = records[i]; char tbuf[64] = ""; if (nvlist_exists(rec, ZPOOL_HIST_TIME)) { time_t tsec; struct tm t; tsec = fnvlist_lookup_uint64(records[i], ZPOOL_HIST_TIME); (void) localtime_r(&tsec, &t); (void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t); } if (nvlist_exists(rec, ZPOOL_HIST_ELAPSED_NS)) { uint64_t elapsed_ns = fnvlist_lookup_int64(records[i], ZPOOL_HIST_ELAPSED_NS); (void) snprintf(tbuf + strlen(tbuf), sizeof (tbuf) - strlen(tbuf), " (%lldms)", (long long)elapsed_ns / 1000 / 1000); } if (nvlist_exists(rec, ZPOOL_HIST_CMD)) { (void) printf("%s %s", tbuf, fnvlist_lookup_string(rec, ZPOOL_HIST_CMD)); } else if (nvlist_exists(rec, ZPOOL_HIST_INT_EVENT)) { int ievent = fnvlist_lookup_uint64(rec, ZPOOL_HIST_INT_EVENT); if (!cb->internal) continue; if (ievent >= ZFS_NUM_LEGACY_HISTORY_EVENTS) { (void) printf("%s unrecognized record:\n", tbuf); dump_nvlist(rec, 4); continue; } (void) printf("%s [internal %s txg:%lld] %s", tbuf, zfs_history_event_names[ievent], (longlong_t)fnvlist_lookup_uint64( rec, ZPOOL_HIST_TXG), fnvlist_lookup_string(rec, ZPOOL_HIST_INT_STR)); } else if (nvlist_exists(rec, ZPOOL_HIST_INT_NAME)) { if (!cb->internal) continue; (void) printf("%s [txg:%lld] %s", tbuf, (longlong_t)fnvlist_lookup_uint64( rec, ZPOOL_HIST_TXG), fnvlist_lookup_string(rec, ZPOOL_HIST_INT_NAME)); if (nvlist_exists(rec, ZPOOL_HIST_DSNAME)) { (void) printf(" %s (%llu)", fnvlist_lookup_string(rec, ZPOOL_HIST_DSNAME), (u_longlong_t)fnvlist_lookup_uint64(rec, ZPOOL_HIST_DSID)); } (void) printf(" %s", fnvlist_lookup_string(rec, ZPOOL_HIST_INT_STR)); } else if (nvlist_exists(rec, ZPOOL_HIST_IOCTL)) { if (!cb->internal) continue; (void) printf("%s ioctl %s\n", tbuf, fnvlist_lookup_string(rec, ZPOOL_HIST_IOCTL)); if (nvlist_exists(rec, ZPOOL_HIST_INPUT_NVL)) { (void) printf(" input:\n"); dump_nvlist(fnvlist_lookup_nvlist(rec, ZPOOL_HIST_INPUT_NVL), 8); } if (nvlist_exists(rec, ZPOOL_HIST_OUTPUT_NVL)) { (void) printf(" output:\n"); dump_nvlist(fnvlist_lookup_nvlist(rec, ZPOOL_HIST_OUTPUT_NVL), 8); } if (nvlist_exists(rec, ZPOOL_HIST_OUTPUT_SIZE)) { (void) printf(" output nvlist omitted; " "original size: %lldKB\n", (longlong_t)fnvlist_lookup_int64(rec, ZPOOL_HIST_OUTPUT_SIZE) / 1024); } if (nvlist_exists(rec, ZPOOL_HIST_ERRNO)) { (void) printf(" errno: %lld\n", (longlong_t)fnvlist_lookup_int64(rec, ZPOOL_HIST_ERRNO)); } } else { if (!cb->internal) continue; (void) printf("%s unrecognized record:\n", tbuf); dump_nvlist(rec, 4); } if (!cb->longfmt) { (void) printf("\n"); continue; } (void) printf(" ["); if (nvlist_exists(rec, ZPOOL_HIST_WHO)) { uid_t who = fnvlist_lookup_uint64(rec, ZPOOL_HIST_WHO); struct passwd *pwd = getpwuid(who); (void) printf("user %d ", (int)who); if (pwd != NULL) (void) printf("(%s) ", pwd->pw_name); } if (nvlist_exists(rec, ZPOOL_HIST_HOST)) { (void) printf("on %s", fnvlist_lookup_string(rec, ZPOOL_HIST_HOST)); } if (nvlist_exists(rec, ZPOOL_HIST_ZONE)) { (void) printf(":%s", fnvlist_lookup_string(rec, ZPOOL_HIST_ZONE)); } (void) printf("]"); (void) printf("\n"); } } /* * Print out the command history for a specific pool. */ static int get_history_one(zpool_handle_t *zhp, void *data) { nvlist_t *nvhis; int ret; hist_cbdata_t *cb = (hist_cbdata_t *)data; uint64_t off = 0; boolean_t eof = B_FALSE; cb->first = B_FALSE; (void) printf(gettext("History for '%s':\n"), zpool_get_name(zhp)); while (!eof) { if ((ret = zpool_get_history(zhp, &nvhis, &off, &eof)) != 0) return (ret); print_history_records(nvhis, cb); nvlist_free(nvhis); } (void) printf("\n"); return (ret); } /* * zpool history * * Displays the history of commands that modified pools. */ int zpool_do_history(int argc, char **argv) { hist_cbdata_t cbdata = { 0 }; int ret; int c; cbdata.first = B_TRUE; /* check options */ while ((c = getopt(argc, argv, "li")) != -1) { switch (c) { case 'l': cbdata.longfmt = B_TRUE; break; case 'i': cbdata.internal = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; ret = for_each_pool(argc, argv, B_FALSE, NULL, B_FALSE, get_history_one, &cbdata); if (argc == 0 && cbdata.first == B_TRUE) { (void) fprintf(stderr, gettext("no pools available\n")); return (0); } return (ret); } typedef struct ev_opts { int verbose; int scripted; int follow; int clear; char poolname[ZFS_MAX_DATASET_NAME_LEN]; } ev_opts_t; static void zpool_do_events_short(nvlist_t *nvl, ev_opts_t *opts) { char ctime_str[26], str[32], *ptr; int64_t *tv; uint_t n; verify(nvlist_lookup_int64_array(nvl, FM_EREPORT_TIME, &tv, &n) == 0); memset(str, ' ', 32); (void) ctime_r((const time_t *)&tv[0], ctime_str); (void) memcpy(str, ctime_str+4, 6); /* 'Jun 30' */ (void) memcpy(str+7, ctime_str+20, 4); /* '1993' */ (void) memcpy(str+12, ctime_str+11, 8); /* '21:49:08' */ (void) sprintf(str+20, ".%09lld", (longlong_t)tv[1]); /* '.123456789' */ if (opts->scripted) (void) printf(gettext("%s\t"), str); else (void) printf(gettext("%s "), str); verify(nvlist_lookup_string(nvl, FM_CLASS, &ptr) == 0); (void) printf(gettext("%s\n"), ptr); } static void zpool_do_events_nvprint(nvlist_t *nvl, int depth) { nvpair_t *nvp; for (nvp = nvlist_next_nvpair(nvl, NULL); nvp != NULL; nvp = nvlist_next_nvpair(nvl, nvp)) { data_type_t type = nvpair_type(nvp); const char *name = nvpair_name(nvp); boolean_t b; uint8_t i8; uint16_t i16; uint32_t i32; uint64_t i64; char *str; nvlist_t *cnv; printf(gettext("%*s%s = "), depth, "", name); switch (type) { case DATA_TYPE_BOOLEAN: printf(gettext("%s"), "1"); break; case DATA_TYPE_BOOLEAN_VALUE: (void) nvpair_value_boolean_value(nvp, &b); printf(gettext("%s"), b ? "1" : "0"); break; case DATA_TYPE_BYTE: (void) nvpair_value_byte(nvp, &i8); printf(gettext("0x%x"), i8); break; case DATA_TYPE_INT8: (void) nvpair_value_int8(nvp, (void *)&i8); printf(gettext("0x%x"), i8); break; case DATA_TYPE_UINT8: (void) nvpair_value_uint8(nvp, &i8); printf(gettext("0x%x"), i8); break; case DATA_TYPE_INT16: (void) nvpair_value_int16(nvp, (void *)&i16); printf(gettext("0x%x"), i16); break; case DATA_TYPE_UINT16: (void) nvpair_value_uint16(nvp, &i16); printf(gettext("0x%x"), i16); break; case DATA_TYPE_INT32: (void) nvpair_value_int32(nvp, (void *)&i32); printf(gettext("0x%x"), i32); break; case DATA_TYPE_UINT32: (void) nvpair_value_uint32(nvp, &i32); printf(gettext("0x%x"), i32); break; case DATA_TYPE_INT64: (void) nvpair_value_int64(nvp, (void *)&i64); printf(gettext("0x%llx"), (u_longlong_t)i64); break; case DATA_TYPE_UINT64: (void) nvpair_value_uint64(nvp, &i64); /* * translate vdev state values to readable * strings to aide zpool events consumers */ if (strcmp(name, FM_EREPORT_PAYLOAD_ZFS_VDEV_STATE) == 0 || strcmp(name, FM_EREPORT_PAYLOAD_ZFS_VDEV_LASTSTATE) == 0) { printf(gettext("\"%s\" (0x%llx)"), zpool_state_to_name(i64, VDEV_AUX_NONE), (u_longlong_t)i64); } else { printf(gettext("0x%llx"), (u_longlong_t)i64); } break; case DATA_TYPE_HRTIME: (void) nvpair_value_hrtime(nvp, (void *)&i64); printf(gettext("0x%llx"), (u_longlong_t)i64); break; case DATA_TYPE_STRING: (void) nvpair_value_string(nvp, &str); printf(gettext("\"%s\""), str ? str : ""); break; case DATA_TYPE_NVLIST: printf(gettext("(embedded nvlist)\n")); (void) nvpair_value_nvlist(nvp, &cnv); zpool_do_events_nvprint(cnv, depth + 8); printf(gettext("%*s(end %s)"), depth, "", name); break; case DATA_TYPE_NVLIST_ARRAY: { nvlist_t **val; uint_t i, nelem; (void) nvpair_value_nvlist_array(nvp, &val, &nelem); printf(gettext("(%d embedded nvlists)\n"), nelem); for (i = 0; i < nelem; i++) { printf(gettext("%*s%s[%d] = %s\n"), depth, "", name, i, "(embedded nvlist)"); zpool_do_events_nvprint(val[i], depth + 8); printf(gettext("%*s(end %s[%i])\n"), depth, "", name, i); } printf(gettext("%*s(end %s)\n"), depth, "", name); } break; case DATA_TYPE_INT8_ARRAY: { int8_t *val; uint_t i, nelem; (void) nvpair_value_int8_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_UINT8_ARRAY: { uint8_t *val; uint_t i, nelem; (void) nvpair_value_uint8_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_INT16_ARRAY: { int16_t *val; uint_t i, nelem; (void) nvpair_value_int16_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_UINT16_ARRAY: { uint16_t *val; uint_t i, nelem; (void) nvpair_value_uint16_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_INT32_ARRAY: { int32_t *val; uint_t i, nelem; (void) nvpair_value_int32_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_UINT32_ARRAY: { uint32_t *val; uint_t i, nelem; (void) nvpair_value_uint32_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_INT64_ARRAY: { int64_t *val; uint_t i, nelem; (void) nvpair_value_int64_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%llx "), (u_longlong_t)val[i]); break; } case DATA_TYPE_UINT64_ARRAY: { uint64_t *val; uint_t i, nelem; (void) nvpair_value_uint64_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%llx "), (u_longlong_t)val[i]); break; } case DATA_TYPE_STRING_ARRAY: { char **str; uint_t i, nelem; (void) nvpair_value_string_array(nvp, &str, &nelem); for (i = 0; i < nelem; i++) printf(gettext("\"%s\" "), str[i] ? str[i] : ""); break; } case DATA_TYPE_BOOLEAN_ARRAY: case DATA_TYPE_BYTE_ARRAY: case DATA_TYPE_DOUBLE: case DATA_TYPE_DONTCARE: case DATA_TYPE_UNKNOWN: printf(gettext("")); break; } printf(gettext("\n")); } } static int zpool_do_events_next(ev_opts_t *opts) { nvlist_t *nvl; int zevent_fd, ret, dropped; char *pool; zevent_fd = open(ZFS_DEV, O_RDWR); VERIFY(zevent_fd >= 0); if (!opts->scripted) (void) printf(gettext("%-30s %s\n"), "TIME", "CLASS"); while (1) { ret = zpool_events_next(g_zfs, &nvl, &dropped, (opts->follow ? ZEVENT_NONE : ZEVENT_NONBLOCK), zevent_fd); if (ret || nvl == NULL) break; if (dropped > 0) (void) printf(gettext("dropped %d events\n"), dropped); if (strlen(opts->poolname) > 0 && nvlist_lookup_string(nvl, FM_FMRI_ZFS_POOL, &pool) == 0 && strcmp(opts->poolname, pool) != 0) continue; zpool_do_events_short(nvl, opts); if (opts->verbose) { zpool_do_events_nvprint(nvl, 8); printf(gettext("\n")); } (void) fflush(stdout); nvlist_free(nvl); } VERIFY(0 == close(zevent_fd)); return (ret); } static int zpool_do_events_clear(ev_opts_t *opts) { int count, ret; ret = zpool_events_clear(g_zfs, &count); if (!ret) (void) printf(gettext("cleared %d events\n"), count); return (ret); } /* * zpool events [-vHf [pool] | -c] * * Displays events logs by ZFS. */ int zpool_do_events(int argc, char **argv) { ev_opts_t opts = { 0 }; int ret; int c; /* check options */ while ((c = getopt(argc, argv, "vHfc")) != -1) { switch (c) { case 'v': opts.verbose = 1; break; case 'H': opts.scripted = 1; break; case 'f': opts.follow = 1; break; case 'c': opts.clear = 1; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } else if (argc == 1) { (void) strlcpy(opts.poolname, argv[0], sizeof (opts.poolname)); if (!zfs_name_valid(opts.poolname, ZFS_TYPE_POOL)) { (void) fprintf(stderr, gettext("invalid pool name '%s'\n"), opts.poolname); usage(B_FALSE); } } if ((argc == 1 || opts.verbose || opts.scripted || opts.follow) && opts.clear) { (void) fprintf(stderr, gettext("invalid options combined with -c\n")); usage(B_FALSE); } if (opts.clear) ret = zpool_do_events_clear(&opts); else ret = zpool_do_events_next(&opts); return (ret); } static int get_callback(zpool_handle_t *zhp, void *data) { zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data; char value[MAXNAMELEN]; zprop_source_t srctype; zprop_list_t *pl; for (pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) { /* * Skip the special fake placeholder. This will also skip * over the name property when 'all' is specified. */ if (pl->pl_prop == ZPOOL_PROP_NAME && pl == cbp->cb_proplist) continue; if (pl->pl_prop == ZPROP_INVAL && (zpool_prop_feature(pl->pl_user_prop) || zpool_prop_unsupported(pl->pl_user_prop))) { srctype = ZPROP_SRC_LOCAL; if (zpool_prop_get_feature(zhp, pl->pl_user_prop, value, sizeof (value)) == 0) { zprop_print_one_property(zpool_get_name(zhp), cbp, pl->pl_user_prop, value, srctype, NULL, NULL); } } else { if (zpool_get_prop(zhp, pl->pl_prop, value, sizeof (value), &srctype, cbp->cb_literal) != 0) continue; zprop_print_one_property(zpool_get_name(zhp), cbp, zpool_prop_to_name(pl->pl_prop), value, srctype, NULL, NULL); } } return (0); } /* * zpool get [-Hp] [-o "all" | field[,...]] <"all" | property[,...]> ... * * -H Scripted mode. Don't display headers, and separate properties * by a single tab. * -o List of columns to display. Defaults to * "name,property,value,source". * -p Display values in parsable (exact) format. * * Get properties of pools in the system. Output space statistics * for each one as well as other attributes. */ int zpool_do_get(int argc, char **argv) { zprop_get_cbdata_t cb = { 0 }; zprop_list_t fake_name = { 0 }; int ret; int c, i; char *value; cb.cb_first = B_TRUE; /* * Set up default columns and sources. */ cb.cb_sources = ZPROP_SRC_ALL; cb.cb_columns[0] = GET_COL_NAME; cb.cb_columns[1] = GET_COL_PROPERTY; cb.cb_columns[2] = GET_COL_VALUE; cb.cb_columns[3] = GET_COL_SOURCE; cb.cb_type = ZFS_TYPE_POOL; /* check options */ while ((c = getopt(argc, argv, ":Hpo:")) != -1) { switch (c) { case 'p': cb.cb_literal = B_TRUE; break; case 'H': cb.cb_scripted = B_TRUE; break; case 'o': bzero(&cb.cb_columns, sizeof (cb.cb_columns)); i = 0; while (*optarg != '\0') { static char *col_subopts[] = { "name", "property", "value", "source", "all", NULL }; if (i == ZFS_GET_NCOLS) { (void) fprintf(stderr, gettext("too " "many fields given to -o " "option\n")); usage(B_FALSE); } switch (getsubopt(&optarg, col_subopts, &value)) { case 0: cb.cb_columns[i++] = GET_COL_NAME; break; case 1: cb.cb_columns[i++] = GET_COL_PROPERTY; break; case 2: cb.cb_columns[i++] = GET_COL_VALUE; break; case 3: cb.cb_columns[i++] = GET_COL_SOURCE; break; case 4: if (i > 0) { (void) fprintf(stderr, gettext("\"all\" conflicts " "with specific fields " "given to -o option\n")); usage(B_FALSE); } cb.cb_columns[0] = GET_COL_NAME; cb.cb_columns[1] = GET_COL_PROPERTY; cb.cb_columns[2] = GET_COL_VALUE; cb.cb_columns[3] = GET_COL_SOURCE; i = ZFS_GET_NCOLS; break; default: (void) fprintf(stderr, gettext("invalid column name " "'%s'\n"), value); usage(B_FALSE); } } break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing property " "argument\n")); usage(B_FALSE); } if (zprop_get_list(g_zfs, argv[0], &cb.cb_proplist, ZFS_TYPE_POOL) != 0) usage(B_FALSE); argc--; argv++; if (cb.cb_proplist != NULL) { fake_name.pl_prop = ZPOOL_PROP_NAME; fake_name.pl_width = strlen(gettext("NAME")); fake_name.pl_next = cb.cb_proplist; cb.cb_proplist = &fake_name; } ret = for_each_pool(argc, argv, B_TRUE, &cb.cb_proplist, cb.cb_literal, get_callback, &cb); if (cb.cb_proplist == &fake_name) zprop_free_list(fake_name.pl_next); else zprop_free_list(cb.cb_proplist); return (ret); } typedef struct set_cbdata { char *cb_propname; char *cb_value; boolean_t cb_any_successful; } set_cbdata_t; static int set_callback(zpool_handle_t *zhp, void *data) { int error; set_cbdata_t *cb = (set_cbdata_t *)data; /* Check if we have out-of-bounds features */ if (strcmp(cb->cb_propname, ZPOOL_CONFIG_COMPATIBILITY) == 0) { boolean_t features[SPA_FEATURES]; if (zpool_do_load_compat(cb->cb_value, features) != ZPOOL_COMPATIBILITY_OK) return (-1); nvlist_t *enabled = zpool_get_features(zhp); spa_feature_t i; for (i = 0; i < SPA_FEATURES; i++) { const char *fguid = spa_feature_table[i].fi_guid; if (nvlist_exists(enabled, fguid) && !features[i]) break; } if (i < SPA_FEATURES) (void) fprintf(stderr, gettext("Warning: one or " "more features already enabled on pool '%s'\n" "are not present in this compatibility set.\n"), zpool_get_name(zhp)); } /* if we're setting a feature, check it's in compatibility set */ if (zpool_prop_feature(cb->cb_propname) && strcmp(cb->cb_value, ZFS_FEATURE_ENABLED) == 0) { char *fname = strchr(cb->cb_propname, '@') + 1; spa_feature_t f; if (zfeature_lookup_name(fname, &f) == 0) { char compat[ZFS_MAXPROPLEN]; if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat, ZFS_MAXPROPLEN, NULL, B_FALSE) != 0) compat[0] = '\0'; boolean_t features[SPA_FEATURES]; if (zpool_do_load_compat(compat, features) != ZPOOL_COMPATIBILITY_OK) { (void) fprintf(stderr, gettext("Error: " "cannot enable feature '%s' on pool '%s'\n" "because the pool's 'compatibility' " "property cannot be parsed.\n"), fname, zpool_get_name(zhp)); return (-1); } if (!features[f]) { (void) fprintf(stderr, gettext("Error: " "cannot enable feature '%s' on pool '%s'\n" "as it is not specified in this pool's " "current compatibility set.\n" "Consider setting 'compatibility' to a " "less restrictive set, or to 'off'.\n"), fname, zpool_get_name(zhp)); return (-1); } } } error = zpool_set_prop(zhp, cb->cb_propname, cb->cb_value); if (!error) cb->cb_any_successful = B_TRUE; return (error); } int zpool_do_set(int argc, char **argv) { set_cbdata_t cb = { 0 }; int error; if (argc > 1 && argv[1][0] == '-') { (void) fprintf(stderr, gettext("invalid option '%c'\n"), argv[1][1]); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing property=value " "argument\n")); usage(B_FALSE); } if (argc < 3) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc > 3) { (void) fprintf(stderr, gettext("too many pool names\n")); usage(B_FALSE); } cb.cb_propname = argv[1]; cb.cb_value = strchr(cb.cb_propname, '='); if (cb.cb_value == NULL) { (void) fprintf(stderr, gettext("missing value in " "property=value argument\n")); usage(B_FALSE); } *(cb.cb_value) = '\0'; cb.cb_value++; error = for_each_pool(argc - 2, argv + 2, B_TRUE, NULL, B_FALSE, set_callback, &cb); return (error); } /* Add up the total number of bytes left to initialize/trim across all vdevs */ static uint64_t vdev_activity_remaining(nvlist_t *nv, zpool_wait_activity_t activity) { uint64_t bytes_remaining; nvlist_t **child; uint_t c, children; vdev_stat_t *vs; assert(activity == ZPOOL_WAIT_INITIALIZE || activity == ZPOOL_WAIT_TRIM); verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c) == 0); if (activity == ZPOOL_WAIT_INITIALIZE && vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE) bytes_remaining = vs->vs_initialize_bytes_est - vs->vs_initialize_bytes_done; else if (activity == ZPOOL_WAIT_TRIM && vs->vs_trim_state == VDEV_TRIM_ACTIVE) bytes_remaining = vs->vs_trim_bytes_est - vs->vs_trim_bytes_done; else bytes_remaining = 0; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (c = 0; c < children; c++) bytes_remaining += vdev_activity_remaining(child[c], activity); return (bytes_remaining); } /* Add up the total number of bytes left to rebuild across top-level vdevs */ static uint64_t vdev_activity_top_remaining(nvlist_t *nv) { uint64_t bytes_remaining = 0; nvlist_t **child; uint_t children; int error; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (uint_t c = 0; c < children; c++) { vdev_rebuild_stat_t *vrs; uint_t i; error = nvlist_lookup_uint64_array(child[c], ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i); if (error == 0) { if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) { bytes_remaining += (vrs->vrs_bytes_est - vrs->vrs_bytes_rebuilt); } } } return (bytes_remaining); } /* Whether any vdevs are 'spare' or 'replacing' vdevs */ static boolean_t vdev_any_spare_replacing(nvlist_t *nv) { nvlist_t **child; uint_t c, children; char *vdev_type; (void) nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &vdev_type); if (strcmp(vdev_type, VDEV_TYPE_REPLACING) == 0 || strcmp(vdev_type, VDEV_TYPE_SPARE) == 0 || strcmp(vdev_type, VDEV_TYPE_DRAID_SPARE) == 0) { return (B_TRUE); } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (c = 0; c < children; c++) { if (vdev_any_spare_replacing(child[c])) return (B_TRUE); } return (B_FALSE); } typedef struct wait_data { char *wd_poolname; boolean_t wd_scripted; boolean_t wd_exact; boolean_t wd_headers_once; boolean_t wd_should_exit; /* Which activities to wait for */ boolean_t wd_enabled[ZPOOL_WAIT_NUM_ACTIVITIES]; float wd_interval; pthread_cond_t wd_cv; pthread_mutex_t wd_mutex; } wait_data_t; /* * Print to stdout a single line, containing one column for each activity that * we are waiting for specifying how many bytes of work are left for that * activity. */ static void print_wait_status_row(wait_data_t *wd, zpool_handle_t *zhp, int row) { nvlist_t *config, *nvroot; uint_t c; int i; pool_checkpoint_stat_t *pcs = NULL; pool_scan_stat_t *pss = NULL; pool_removal_stat_t *prs = NULL; char *headers[] = {"DISCARD", "FREE", "INITIALIZE", "REPLACE", "REMOVE", "RESILVER", "SCRUB", "TRIM"}; int col_widths[ZPOOL_WAIT_NUM_ACTIVITIES]; /* Calculate the width of each column */ for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) { /* * Make sure we have enough space in the col for pretty-printed * numbers and for the column header, and then leave a couple * spaces between cols for readability. */ col_widths[i] = MAX(strlen(headers[i]), 6) + 2; } /* Print header if appropriate */ int term_height = terminal_height(); boolean_t reprint_header = (!wd->wd_headers_once && term_height > 0 && row % (term_height-1) == 0); if (!wd->wd_scripted && (row == 0 || reprint_header)) { for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) { if (wd->wd_enabled[i]) (void) printf("%*s", col_widths[i], headers[i]); } (void) printf("\n"); } /* Bytes of work remaining in each activity */ int64_t bytes_rem[ZPOOL_WAIT_NUM_ACTIVITIES] = {0}; bytes_rem[ZPOOL_WAIT_FREE] = zpool_get_prop_int(zhp, ZPOOL_PROP_FREEING, NULL); config = zpool_get_config(zhp, NULL); nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c); if (pcs != NULL && pcs->pcs_state == CS_CHECKPOINT_DISCARDING) bytes_rem[ZPOOL_WAIT_CKPT_DISCARD] = pcs->pcs_space; (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t **)&prs, &c); if (prs != NULL && prs->prs_state == DSS_SCANNING) bytes_rem[ZPOOL_WAIT_REMOVE] = prs->prs_to_copy - prs->prs_copied; (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&pss, &c); if (pss != NULL && pss->pss_state == DSS_SCANNING && pss->pss_pass_scrub_pause == 0) { int64_t rem = pss->pss_to_examine - pss->pss_issued; if (pss->pss_func == POOL_SCAN_SCRUB) bytes_rem[ZPOOL_WAIT_SCRUB] = rem; else bytes_rem[ZPOOL_WAIT_RESILVER] = rem; } else if (check_rebuilding(nvroot, NULL)) { bytes_rem[ZPOOL_WAIT_RESILVER] = vdev_activity_top_remaining(nvroot); } bytes_rem[ZPOOL_WAIT_INITIALIZE] = vdev_activity_remaining(nvroot, ZPOOL_WAIT_INITIALIZE); bytes_rem[ZPOOL_WAIT_TRIM] = vdev_activity_remaining(nvroot, ZPOOL_WAIT_TRIM); /* * A replace finishes after resilvering finishes, so the amount of work * left for a replace is the same as for resilvering. * * It isn't quite correct to say that if we have any 'spare' or * 'replacing' vdevs and a resilver is happening, then a replace is in * progress, like we do here. When a hot spare is used, the faulted vdev * is not removed after the hot spare is resilvered, so parent 'spare' * vdev is not removed either. So we could have a 'spare' vdev, but be * resilvering for a different reason. However, we use it as a heuristic * because we don't have access to the DTLs, which could tell us whether * or not we have really finished resilvering a hot spare. */ if (vdev_any_spare_replacing(nvroot)) bytes_rem[ZPOOL_WAIT_REPLACE] = bytes_rem[ZPOOL_WAIT_RESILVER]; if (timestamp_fmt != NODATE) print_timestamp(timestamp_fmt); for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) { char buf[64]; if (!wd->wd_enabled[i]) continue; if (wd->wd_exact) (void) snprintf(buf, sizeof (buf), "%" PRIi64, bytes_rem[i]); else zfs_nicenum(bytes_rem[i], buf, sizeof (buf)); if (wd->wd_scripted) (void) printf(i == 0 ? "%s" : "\t%s", buf); else (void) printf(" %*s", col_widths[i] - 1, buf); } (void) printf("\n"); (void) fflush(stdout); } static void * wait_status_thread(void *arg) { wait_data_t *wd = (wait_data_t *)arg; zpool_handle_t *zhp; if ((zhp = zpool_open(g_zfs, wd->wd_poolname)) == NULL) return (void *)(1); for (int row = 0; ; row++) { boolean_t missing; struct timespec timeout; int ret = 0; (void) clock_gettime(CLOCK_REALTIME, &timeout); if (zpool_refresh_stats(zhp, &missing) != 0 || missing || zpool_props_refresh(zhp) != 0) { zpool_close(zhp); return (void *)(uintptr_t)(missing ? 0 : 1); } print_wait_status_row(wd, zhp, row); timeout.tv_sec += floor(wd->wd_interval); long nanos = timeout.tv_nsec + (wd->wd_interval - floor(wd->wd_interval)) * NANOSEC; if (nanos >= NANOSEC) { timeout.tv_sec++; timeout.tv_nsec = nanos - NANOSEC; } else { timeout.tv_nsec = nanos; } pthread_mutex_lock(&wd->wd_mutex); if (!wd->wd_should_exit) ret = pthread_cond_timedwait(&wd->wd_cv, &wd->wd_mutex, &timeout); pthread_mutex_unlock(&wd->wd_mutex); if (ret == 0) { break; /* signaled by main thread */ } else if (ret != ETIMEDOUT) { (void) fprintf(stderr, gettext("pthread_cond_timedwait " "failed: %s\n"), strerror(ret)); zpool_close(zhp); return (void *)(uintptr_t)(1); } } zpool_close(zhp); return (void *)(0); } int zpool_do_wait(int argc, char **argv) { boolean_t verbose = B_FALSE; int c; char *value; int i; unsigned long count; pthread_t status_thr; int error = 0; zpool_handle_t *zhp; wait_data_t wd; wd.wd_scripted = B_FALSE; wd.wd_exact = B_FALSE; wd.wd_headers_once = B_FALSE; wd.wd_should_exit = B_FALSE; pthread_mutex_init(&wd.wd_mutex, NULL); pthread_cond_init(&wd.wd_cv, NULL); /* By default, wait for all types of activity. */ for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) wd.wd_enabled[i] = B_TRUE; while ((c = getopt(argc, argv, "HpT:t:")) != -1) { switch (c) { case 'H': wd.wd_scripted = B_TRUE; break; case 'n': wd.wd_headers_once = B_TRUE; break; case 'p': wd.wd_exact = B_TRUE; break; case 'T': get_timestamp_arg(*optarg); break; case 't': { static char *col_subopts[] = { "discard", "free", "initialize", "replace", "remove", "resilver", "scrub", "trim", NULL }; /* Reset activities array */ bzero(&wd.wd_enabled, sizeof (wd.wd_enabled)); while (*optarg != '\0') { int activity = getsubopt(&optarg, col_subopts, &value); if (activity < 0) { (void) fprintf(stderr, gettext("invalid activity '%s'\n"), value); usage(B_FALSE); } wd.wd_enabled[activity] = B_TRUE; } break; } case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; get_interval_count(&argc, argv, &wd.wd_interval, &count); if (count != 0) { /* This subcmd only accepts an interval, not a count */ (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } if (wd.wd_interval != 0) verbose = B_TRUE; if (argc < 1) { (void) fprintf(stderr, gettext("missing 'pool' argument\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } wd.wd_poolname = argv[0]; if ((zhp = zpool_open(g_zfs, wd.wd_poolname)) == NULL) return (1); if (verbose) { /* * We use a separate thread for printing status updates because * the main thread will call lzc_wait(), which blocks as long * as an activity is in progress, which can be a long time. */ if (pthread_create(&status_thr, NULL, wait_status_thread, &wd) != 0) { (void) fprintf(stderr, gettext("failed to create status" "thread: %s\n"), strerror(errno)); zpool_close(zhp); return (1); } } /* * Loop over all activities that we are supposed to wait for until none * of them are in progress. Note that this means we can end up waiting * for more activities to complete than just those that were in progress * when we began waiting; if an activity we are interested in begins * while we are waiting for another activity, we will wait for both to * complete before exiting. */ for (;;) { boolean_t missing = B_FALSE; boolean_t any_waited = B_FALSE; for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) { boolean_t waited; if (!wd.wd_enabled[i]) continue; error = zpool_wait_status(zhp, i, &missing, &waited); if (error != 0 || missing) break; any_waited = (any_waited || waited); } if (error != 0 || missing || !any_waited) break; } zpool_close(zhp); if (verbose) { uintptr_t status; pthread_mutex_lock(&wd.wd_mutex); wd.wd_should_exit = B_TRUE; pthread_cond_signal(&wd.wd_cv); pthread_mutex_unlock(&wd.wd_mutex); (void) pthread_join(status_thr, (void *)&status); if (status != 0) error = status; } pthread_mutex_destroy(&wd.wd_mutex); pthread_cond_destroy(&wd.wd_cv); return (error); } static int find_command_idx(char *command, int *idx) { int i; for (i = 0; i < NCOMMAND; i++) { if (command_table[i].name == NULL) continue; if (strcmp(command, command_table[i].name) == 0) { *idx = i; return (0); } } return (1); } /* * Display version message */ static int zpool_do_version(int argc, char **argv) { if (zfs_version_print() == -1) return (1); return (0); } /* * Do zpool_load_compat() and print error message on failure */ static zpool_compat_status_t zpool_do_load_compat(const char *compat, boolean_t *list) { char report[1024]; zpool_compat_status_t ret; ret = zpool_load_compat(compat, list, report, 1024); switch (ret) { case ZPOOL_COMPATIBILITY_OK: break; case ZPOOL_COMPATIBILITY_NOFILES: case ZPOOL_COMPATIBILITY_BADFILE: case ZPOOL_COMPATIBILITY_BADTOKEN: (void) fprintf(stderr, "Error: %s\n", report); break; case ZPOOL_COMPATIBILITY_WARNTOKEN: (void) fprintf(stderr, "Warning: %s\n", report); ret = ZPOOL_COMPATIBILITY_OK; break; } return (ret); } int main(int argc, char **argv) { int ret = 0; int i = 0; char *cmdname; char **newargv; (void) setlocale(LC_ALL, ""); (void) setlocale(LC_NUMERIC, "C"); (void) textdomain(TEXT_DOMAIN); srand(time(NULL)); opterr = 0; /* * Make sure the user has specified some command. */ if (argc < 2) { (void) fprintf(stderr, gettext("missing command\n")); usage(B_FALSE); } cmdname = argv[1]; /* * Special case '-?' */ if ((strcmp(cmdname, "-?") == 0) || strcmp(cmdname, "--help") == 0) usage(B_TRUE); /* * Special case '-V|--version' */ if ((strcmp(cmdname, "-V") == 0) || (strcmp(cmdname, "--version") == 0)) return (zpool_do_version(argc, argv)); if ((g_zfs = libzfs_init()) == NULL) { (void) fprintf(stderr, "%s\n", libzfs_error_init(errno)); return (1); } libzfs_print_on_error(g_zfs, B_TRUE); zfs_save_arguments(argc, argv, history_str, sizeof (history_str)); /* * Many commands modify input strings for string parsing reasons. * We create a copy to protect the original argv. */ newargv = malloc((argc + 1) * sizeof (newargv[0])); for (i = 0; i < argc; i++) newargv[i] = strdup(argv[i]); newargv[argc] = NULL; /* * Run the appropriate command. */ if (find_command_idx(cmdname, &i) == 0) { current_command = &command_table[i]; ret = command_table[i].func(argc - 1, newargv + 1); } else if (strchr(cmdname, '=')) { verify(find_command_idx("set", &i) == 0); current_command = &command_table[i]; ret = command_table[i].func(argc, newargv); } else if (strcmp(cmdname, "freeze") == 0 && argc == 3) { /* * 'freeze' is a vile debugging abomination, so we treat * it as such. */ zfs_cmd_t zc = {"\0"}; (void) strlcpy(zc.zc_name, argv[2], sizeof (zc.zc_name)); ret = zfs_ioctl(g_zfs, ZFS_IOC_POOL_FREEZE, &zc); if (ret != 0) { (void) fprintf(stderr, gettext("failed to freeze pool: %d\n"), errno); ret = 1; } log_history = 0; } else { (void) fprintf(stderr, gettext("unrecognized " "command '%s'\n"), cmdname); usage(B_FALSE); ret = 1; } for (i = 0; i < argc; i++) free(newargv[i]); free(newargv); if (ret == 0 && log_history) (void) zpool_log_history(g_zfs, history_str); libzfs_fini(g_zfs); /* * The 'ZFS_ABORT' environment variable causes us to dump core on exit * for the purposes of running ::findleaks. */ if (getenv("ZFS_ABORT") != NULL) { (void) printf("dumping core by request\n"); abort(); } return (ret); } diff --git a/man/man4/zfs.4 b/man/man4/zfs.4 index 71a95c3bd812..d58523f831df 100644 --- a/man/man4/zfs.4 +++ b/man/man4/zfs.4 @@ -1,2461 +1,2468 @@ .\" .\" Copyright (c) 2013 by Turbo Fredriksson . All rights reserved. .\" Copyright (c) 2019, 2021 by Delphix. All rights reserved. .\" Copyright (c) 2019 Datto Inc. .\" The contents of this file are subject to the terms of the Common Development .\" and Distribution License (the "License"). You may not use this file except .\" in compliance with the License. You can obtain a copy of the license at .\" usr/src/OPENSOLARIS.LICENSE or http://www.opensolaris.org/os/licensing. .\" .\" See the License for the specific language governing permissions and .\" limitations under the License. When distributing Covered Code, include this .\" CDDL HEADER in each file and include the License file at .\" usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this .\" CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your .\" own identifying information: .\" Portions Copyright [yyyy] [name of copyright owner] .\" .Dd January 10, 2023 .Dt ZFS 4 .Os . .Sh NAME .Nm zfs .Nd tuning of the ZFS kernel module . .Sh DESCRIPTION The ZFS module supports these parameters: .Bl -tag -width Ds .It Sy dbuf_cache_max_bytes Ns = Ns Sy ULONG_MAX Ns B Pq ulong Maximum size in bytes of the dbuf cache. The target size is determined by the MIN versus .No 1/2^ Ns Sy dbuf_cache_shift Pq 1/32nd of the target ARC size. The behavior of the dbuf cache and its associated settings can be observed via the .Pa /proc/spl/kstat/zfs/dbufstats kstat. . .It Sy dbuf_metadata_cache_max_bytes Ns = Ns Sy ULONG_MAX Ns B Pq ulong Maximum size in bytes of the metadata dbuf cache. The target size is determined by the MIN versus .No 1/2^ Ns Sy dbuf_metadata_cache_shift Pq 1/64th of the target ARC size. The behavior of the metadata dbuf cache and its associated settings can be observed via the .Pa /proc/spl/kstat/zfs/dbufstats kstat. . .It Sy dbuf_cache_hiwater_pct Ns = Ns Sy 10 Ns % Pq uint The percentage over .Sy dbuf_cache_max_bytes when dbufs must be evicted directly. . .It Sy dbuf_cache_lowater_pct Ns = Ns Sy 10 Ns % Pq uint The percentage below .Sy dbuf_cache_max_bytes when the evict thread stops evicting dbufs. . .It Sy dbuf_cache_shift Ns = Ns Sy 5 Pq int Set the size of the dbuf cache .Pq Sy dbuf_cache_max_bytes to a log2 fraction of the target ARC size. . .It Sy dbuf_metadata_cache_shift Ns = Ns Sy 6 Pq int Set the size of the dbuf metadata cache .Pq Sy dbuf_metadata_cache_max_bytes to a log2 fraction of the target ARC size. . .It Sy dmu_object_alloc_chunk_shift Ns = Ns Sy 7 Po 128 Pc Pq int dnode slots allocated in a single operation as a power of 2. The default value minimizes lock contention for the bulk operation performed. . .It Sy dmu_prefetch_max Ns = Ns Sy 134217728 Ns B Po 128MB Pc Pq int Limit the amount we can prefetch with one call to this amount in bytes. This helps to limit the amount of memory that can be used by prefetching. . .It Sy ignore_hole_birth Pq int Alias for .Sy send_holes_without_birth_time . . .It Sy l2arc_feed_again Ns = Ns Sy 1 Ns | Ns 0 Pq int Turbo L2ARC warm-up. When the L2ARC is cold the fill interval will be set as fast as possible. . .It Sy l2arc_feed_min_ms Ns = Ns Sy 200 Pq ulong Min feed interval in milliseconds. Requires .Sy l2arc_feed_again Ns = Ns Ar 1 and only applicable in related situations. . .It Sy l2arc_feed_secs Ns = Ns Sy 1 Pq ulong Seconds between L2ARC writing. . .It Sy l2arc_headroom Ns = Ns Sy 2 Pq ulong How far through the ARC lists to search for L2ARC cacheable content, expressed as a multiplier of .Sy l2arc_write_max . ARC persistence across reboots can be achieved with persistent L2ARC by setting this parameter to .Sy 0 , allowing the full length of ARC lists to be searched for cacheable content. . .It Sy l2arc_headroom_boost Ns = Ns Sy 200 Ns % Pq ulong Scales .Sy l2arc_headroom by this percentage when L2ARC contents are being successfully compressed before writing. A value of .Sy 100 disables this feature. . .It Sy l2arc_exclude_special Ns = Ns Sy 0 Ns | Ns 1 Pq int Controls whether buffers present on special vdevs are eligibile for caching into L2ARC. If set to 1, exclude dbufs on special vdevs from being cached to L2ARC. . .It Sy l2arc_mfuonly Ns = Ns Sy 0 Ns | Ns 1 Pq int Controls whether only MFU metadata and data are cached from ARC into L2ARC. This may be desired to avoid wasting space on L2ARC when reading/writing large amounts of data that are not expected to be accessed more than once. .Pp The default is off, meaning both MRU and MFU data and metadata are cached. When turning off this feature, some MRU buffers will still be present in ARC and eventually cached on L2ARC. .No If Sy l2arc_noprefetch Ns = Ns Sy 0 , some prefetched buffers will be cached to L2ARC, and those might later transition to MRU, in which case the .Sy l2arc_mru_asize No arcstat will not be Sy 0 . .Pp Regardless of .Sy l2arc_noprefetch , some MFU buffers might be evicted from ARC, accessed later on as prefetches and transition to MRU as prefetches. If accessed again they are counted as MRU and the .Sy l2arc_mru_asize No arcstat will not be Sy 0 . .Pp The ARC status of L2ARC buffers when they were first cached in L2ARC can be seen in the .Sy l2arc_mru_asize , Sy l2arc_mfu_asize , No and Sy l2arc_prefetch_asize arcstats when importing the pool or onlining a cache device if persistent L2ARC is enabled. .Pp The .Sy evict_l2_eligible_mru arcstat does not take into account if this option is enabled as the information provided by the .Sy evict_l2_eligible_m[rf]u arcstats can be used to decide if toggling this option is appropriate for the current workload. . .It Sy l2arc_meta_percent Ns = Ns Sy 33 Ns % Pq int Percent of ARC size allowed for L2ARC-only headers. Since L2ARC buffers are not evicted on memory pressure, too many headers on a system with an irrationally large L2ARC can render it slow or unusable. This parameter limits L2ARC writes and rebuilds to achieve the target. . .It Sy l2arc_trim_ahead Ns = Ns Sy 0 Ns % Pq ulong Trims ahead of the current write size .Pq Sy l2arc_write_max on L2ARC devices by this percentage of write size if we have filled the device. If set to .Sy 100 we TRIM twice the space required to accommodate upcoming writes. A minimum of .Sy 64MB will be trimmed. It also enables TRIM of the whole L2ARC device upon creation or addition to an existing pool or if the header of the device is invalid upon importing a pool or onlining a cache device. A value of .Sy 0 disables TRIM on L2ARC altogether and is the default as it can put significant stress on the underlying storage devices. This will vary depending of how well the specific device handles these commands. . .It Sy l2arc_noprefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int Do not write buffers to L2ARC if they were prefetched but not used by applications. In case there are prefetched buffers in L2ARC and this option is later set, we do not read the prefetched buffers from L2ARC. Unsetting this option is useful for caching sequential reads from the disks to L2ARC and serve those reads from L2ARC later on. This may be beneficial in case the L2ARC device is significantly faster in sequential reads than the disks of the pool. .Pp Use .Sy 1 to disable and .Sy 0 to enable caching/reading prefetches to/from L2ARC. . .It Sy l2arc_norw Ns = Ns Sy 0 Ns | Ns 1 Pq int No reads during writes. . .It Sy l2arc_write_boost Ns = Ns Sy 8388608 Ns B Po 8MB Pc Pq ulong Cold L2ARC devices will have .Sy l2arc_write_max increased by this amount while they remain cold. . .It Sy l2arc_write_max Ns = Ns Sy 8388608 Ns B Po 8MB Pc Pq ulong Max write bytes per interval. . .It Sy l2arc_rebuild_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Rebuild the L2ARC when importing a pool (persistent L2ARC). This can be disabled if there are problems importing a pool or attaching an L2ARC device (e.g. the L2ARC device is slow in reading stored log metadata, or the metadata has become somehow fragmented/unusable). . .It Sy l2arc_rebuild_blocks_min_l2size Ns = Ns Sy 1073741824 Ns B Po 1GB Pc Pq ulong Mininum size of an L2ARC device required in order to write log blocks in it. The log blocks are used upon importing the pool to rebuild the persistent L2ARC. .Pp For L2ARC devices less than 1GB, the amount of data .Fn l2arc_evict evicts is significant compared to the amount of restored L2ARC data. In this case, do not write log blocks in L2ARC in order not to waste space. . .It Sy metaslab_aliquot Ns = Ns Sy 1048576 Ns B Po 1MB Pc Pq ulong Metaslab granularity, in bytes. This is roughly similar to what would be referred to as the "stripe size" in traditional RAID arrays. In normal operation, ZFS will try to write this amount of data to each disk before moving on to the next top-level vdev. . .It Sy metaslab_bias_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable metaslab group biasing based on their vdevs' over- or under-utilization relative to the pool. . .It Sy metaslab_force_ganging Ns = Ns Sy 16777217 Ns B Ns B Po 16MB + 1B Pc Pq ulong Make some blocks above a certain size be gang blocks. This option is used by the test suite to facilitate testing. . .It Sy zfs_default_bs Ns = Ns Sy 9 Po 512 B Pc Pq int Default dnode block size as a power of 2. . .It Sy zfs_default_ibs Ns = Ns Sy 17 Po 128 KiB Pc Pq int Default dnode indirect block size as a power of 2. . .It Sy zfs_history_output_max Ns = Ns Sy 1048576 Ns B Ns B Po 1MB Pc Pq int When attempting to log an output nvlist of an ioctl in the on-disk history, the output will not be stored if it is larger than this size (in bytes). This must be less than .Sy DMU_MAX_ACCESS Pq 64MB . This applies primarily to .Fn zfs_ioc_channel_program Pq cf. Xr zfs-program 8 . . .It Sy zfs_keep_log_spacemaps_at_export Ns = Ns Sy 0 Ns | Ns 1 Pq int Prevent log spacemaps from being destroyed during pool exports and destroys. . .It Sy zfs_metaslab_segment_weight_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable/disable segment-based metaslab selection. . .It Sy zfs_metaslab_switch_threshold Ns = Ns Sy 2 Pq int When using segment-based metaslab selection, continue allocating from the active metaslab until this option's worth of buckets have been exhausted. . .It Sy metaslab_debug_load Ns = Ns Sy 0 Ns | Ns 1 Pq int Load all metaslabs during pool import. . .It Sy metaslab_debug_unload Ns = Ns Sy 0 Ns | Ns 1 Pq int Prevent metaslabs from being unloaded. . .It Sy metaslab_fragmentation_factor_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable use of the fragmentation metric in computing metaslab weights. . .It Sy metaslab_df_max_search Ns = Ns Sy 16777216 Ns B Po 16MB Pc Pq int Maximum distance to search forward from the last offset. Without this limit, fragmented pools can see .Em >100`000 iterations and .Fn metaslab_block_picker becomes the performance limiting factor on high-performance storage. .Pp With the default setting of .Sy 16MB , we typically see less than .Em 500 iterations, even with very fragmented .Sy ashift Ns = Ns Sy 9 pools. The maximum number of iterations possible is .Sy metaslab_df_max_search / 2^(ashift+1) . With the default setting of .Sy 16MB this is .Em 16*1024 Pq with Sy ashift Ns = Ns Sy 9 or .Em 2*1024 Pq with Sy ashift Ns = Ns Sy 12 . . .It Sy metaslab_df_use_largest_segment Ns = Ns Sy 0 Ns | Ns 1 Pq int If not searching forward (due to .Sy metaslab_df_max_search , metaslab_df_free_pct , .No or Sy metaslab_df_alloc_threshold ) , this tunable controls which segment is used. If set, we will use the largest free segment. If unset, we will use a segment of at least the requested size. . .It Sy zfs_metaslab_max_size_cache_sec Ns = Ns Sy 3600 Ns s Po 1h Pc Pq ulong When we unload a metaslab, we cache the size of the largest free chunk. We use that cached size to determine whether or not to load a metaslab for a given allocation. As more frees accumulate in that metaslab while it's unloaded, the cached max size becomes less and less accurate. After a number of seconds controlled by this tunable, we stop considering the cached max size and start considering only the histogram instead. . .It Sy zfs_metaslab_mem_limit Ns = Ns Sy 25 Ns % Pq int When we are loading a new metaslab, we check the amount of memory being used to store metaslab range trees. If it is over a threshold, we attempt to unload the least recently used metaslab to prevent the system from clogging all of its memory with range trees. This tunable sets the percentage of total system memory that is the threshold. . .It Sy zfs_metaslab_try_hard_before_gang Ns = Ns Sy 0 Ns | Ns 1 Pq int .Bl -item -compact .It If unset, we will first try normal allocation. .It If that fails then we will do a gang allocation. .It If that fails then we will do a "try hard" gang allocation. .It If that fails then we will have a multi-layer gang block. .El .Pp .Bl -item -compact .It If set, we will first try normal allocation. .It If that fails then we will do a "try hard" allocation. .It If that fails we will do a gang allocation. .It If that fails we will do a "try hard" gang allocation. .It If that fails then we will have a multi-layer gang block. .El . .It Sy zfs_metaslab_find_max_tries Ns = Ns Sy 100 Pq int When not trying hard, we only consider this number of the best metaslabs. This improves performance, especially when there are many metaslabs per vdev and the allocation can't actually be satisfied (so we would otherwise iterate all metaslabs). . .It Sy zfs_vdev_default_ms_count Ns = Ns Sy 200 Pq int When a vdev is added, target this number of metaslabs per top-level vdev. . .It Sy zfs_vdev_default_ms_shift Ns = Ns Sy 29 Po 512MB Pc Pq int Default limit for metaslab size. . .It Sy zfs_vdev_max_auto_ashift Ns = Ns Sy 14 Pq ulong Maximum ashift used when optimizing for logical -> physical sector size on new top-level vdevs. May be increased up to .Sy ASHIFT_MAX Po 16 Pc , but this may negatively impact pool space efficiency. . .It Sy zfs_vdev_min_auto_ashift Ns = Ns Sy ASHIFT_MIN Po 9 Pc Pq ulong Minimum ashift used when creating new top-level vdevs. . .It Sy zfs_vdev_min_ms_count Ns = Ns Sy 16 Pq int Minimum number of metaslabs to create in a top-level vdev. . .It Sy vdev_validate_skip Ns = Ns Sy 0 Ns | Ns 1 Pq int Skip label validation steps during pool import. Changing is not recommended unless you know what you're doing and are recovering a damaged label. . .It Sy zfs_vdev_ms_count_limit Ns = Ns Sy 131072 Po 128k Pc Pq int Practical upper limit of total metaslabs per top-level vdev. . .It Sy metaslab_preload_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable metaslab group preloading. . .It Sy metaslab_lba_weighting_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Give more weight to metaslabs with lower LBAs, assuming they have greater bandwidth, as is typically the case on a modern constant angular velocity disk drive. . .It Sy metaslab_unload_delay Ns = Ns Sy 32 Pq int After a metaslab is used, we keep it loaded for this many TXGs, to attempt to reduce unnecessary reloading. Note that both this many TXGs and .Sy metaslab_unload_delay_ms milliseconds must pass before unloading will occur. . .It Sy metaslab_unload_delay_ms Ns = Ns Sy 600000 Ns ms Po 10min Pc Pq int After a metaslab is used, we keep it loaded for this many milliseconds, to attempt to reduce unnecessary reloading. Note, that both this many milliseconds and .Sy metaslab_unload_delay TXGs must pass before unloading will occur. . .It Sy reference_history Ns = Ns Sy 3 Pq int Maximum reference holders being tracked when reference_tracking_enable is active. . .It Sy reference_tracking_enable Ns = Ns Sy 0 Ns | Ns 1 Pq int Track reference holders to .Sy refcount_t objects (debug builds only). . .It Sy send_holes_without_birth_time Ns = Ns Sy 1 Ns | Ns 0 Pq int When set, the .Sy hole_birth optimization will not be used, and all holes will always be sent during a .Nm zfs Cm send . This is useful if you suspect your datasets are affected by a bug in .Sy hole_birth . . .It Sy spa_config_path Ns = Ns Pa /etc/zfs/zpool.cache Pq charp SPA config file. . .It Sy spa_asize_inflation Ns = Ns Sy 24 Pq int Multiplication factor used to estimate actual disk consumption from the size of data being written. The default value is a worst case estimate, but lower values may be valid for a given pool depending on its configuration. Pool administrators who understand the factors involved may wish to specify a more realistic inflation factor, particularly if they operate close to quota or capacity limits. . .It Sy spa_load_print_vdev_tree Ns = Ns Sy 0 Ns | Ns 1 Pq int Whether to print the vdev tree in the debugging message buffer during pool import. . .It Sy spa_load_verify_data Ns = Ns Sy 1 Ns | Ns 0 Pq int Whether to traverse data blocks during an "extreme rewind" .Pq Fl X import. .Pp An extreme rewind import normally performs a full traversal of all blocks in the pool for verification. If this parameter is unset, the traversal skips non-metadata blocks. It can be toggled once the import has started to stop or start the traversal of non-metadata blocks. . .It Sy spa_load_verify_metadata Ns = Ns Sy 1 Ns | Ns 0 Pq int Whether to traverse blocks during an "extreme rewind" .Pq Fl X pool import. .Pp An extreme rewind import normally performs a full traversal of all blocks in the pool for verification. If this parameter is unset, the traversal is not performed. It can be toggled once the import has started to stop or start the traversal. . .It Sy spa_load_verify_shift Ns = Ns Sy 4 Po 1/16th Pc Pq int Sets the maximum number of bytes to consume during pool import to the log2 fraction of the target ARC size. . .It Sy spa_slop_shift Ns = Ns Sy 5 Po 1/32nd Pc Pq int Normally, we don't allow the last .Sy 3.2% Pq Sy 1/2^spa_slop_shift of space in the pool to be consumed. This ensures that we don't run the pool completely out of space, due to unaccounted changes (e.g. to the MOS). It also limits the worst-case time to allocate space. If we have less than this amount of free space, most ZPL operations (e.g. write, create) will return .Sy ENOSPC . . .It Sy vdev_removal_max_span Ns = Ns Sy 32768 Ns B Po 32kB Pc Pq int During top-level vdev removal, chunks of data are copied from the vdev which may include free space in order to trade bandwidth for IOPS. This parameter determines the maximum span of free space, in bytes, which will be included as "unnecessary" data in a chunk of copied data. .Pp The default value here was chosen to align with .Sy zfs_vdev_read_gap_limit , which is a similar concept when doing regular reads (but there's no reason it has to be the same). . .It Sy vdev_file_logical_ashift Ns = Ns Sy 9 Po 512B Pc Pq ulong Logical ashift for file-based devices. . .It Sy vdev_file_physical_ashift Ns = Ns Sy 9 Po 512B Pc Pq ulong Physical ashift for file-based devices. . .It Sy zap_iterate_prefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int If set, when we start iterating over a ZAP object, prefetch the entire object (all leaf blocks). However, this is limited by .Sy dmu_prefetch_max . . .It Sy zfetch_array_rd_sz Ns = Ns Sy 1048576 Ns B Po 1MB Pc Pq ulong If prefetching is enabled, disable prefetching for reads larger than this size. . .It Sy zfetch_min_distance Ns = Ns Sy 4194304 Ns B Po 4 MiB Pc Pq uint Min bytes to prefetch per stream. Prefetch distance starts from the demand access size and quickly grows to this value, doubling on each hit. After that it may grow further by 1/8 per hit, but only if some prefetch since last time haven't completed in time to satisfy demand request, i.e. prefetch depth didn't cover the read latency or the pool got saturated. . .It Sy zfetch_max_distance Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq uint Max bytes to prefetch per stream. . .It Sy zfetch_max_idistance Ns = Ns Sy 67108864 Ns B Po 64MB Pc Pq uint Max bytes to prefetch indirects for per stream. . .It Sy zfetch_max_streams Ns = Ns Sy 8 Pq uint Max number of streams per zfetch (prefetch streams per file). . .It Sy zfetch_min_sec_reap Ns = Ns Sy 1 Pq uint Min time before inactive prefetch stream can be reclaimed . .It Sy zfetch_max_sec_reap Ns = Ns Sy 2 Pq uint Max time before inactive prefetch stream can be deleted . .It Sy zfs_abd_scatter_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enables ARC from using scatter/gather lists and forces all allocations to be linear in kernel memory. Disabling can improve performance in some code paths at the expense of fragmented kernel memory. . .It Sy zfs_abd_scatter_max_order Ns = Ns Sy MAX_ORDER-1 Pq uint Maximum number of consecutive memory pages allocated in a single block for scatter/gather lists. .Pp The value of .Sy MAX_ORDER depends on kernel configuration. . .It Sy zfs_abd_scatter_min_size Ns = Ns Sy 1536 Ns B Po 1.5kB Pc Pq uint This is the minimum allocation size that will use scatter (page-based) ABDs. Smaller allocations will use linear ABDs. . .It Sy zfs_arc_dnode_limit Ns = Ns Sy 0 Ns B Pq ulong When the number of bytes consumed by dnodes in the ARC exceeds this number of bytes, try to unpin some of it in response to demand for non-metadata. This value acts as a ceiling to the amount of dnode metadata, and defaults to .Sy 0 , which indicates that a percent which is based on .Sy zfs_arc_dnode_limit_percent of the ARC meta buffers that may be used for dnodes. .Pp Also see .Sy zfs_arc_meta_prune which serves a similar purpose but is used when the amount of metadata in the ARC exceeds .Sy zfs_arc_meta_limit rather than in response to overall demand for non-metadata. . .It Sy zfs_arc_dnode_limit_percent Ns = Ns Sy 10 Ns % Pq ulong Percentage that can be consumed by dnodes of ARC meta buffers. .Pp See also .Sy zfs_arc_dnode_limit , which serves a similar purpose but has a higher priority if nonzero. . .It Sy zfs_arc_dnode_reduce_percent Ns = Ns Sy 10 Ns % Pq ulong Percentage of ARC dnodes to try to scan in response to demand for non-metadata when the number of bytes consumed by dnodes exceeds .Sy zfs_arc_dnode_limit . . .It Sy zfs_arc_average_blocksize Ns = Ns Sy 8192 Ns B Po 8kB Pc Pq int The ARC's buffer hash table is sized based on the assumption of an average block size of this value. This works out to roughly 1MB of hash table per 1GB of physical memory with 8-byte pointers. For configurations with a known larger average block size, this value can be increased to reduce the memory footprint. . .It Sy zfs_arc_eviction_pct Ns = Ns Sy 200 Ns % Pq int When .Fn arc_is_overflowing , .Fn arc_get_data_impl waits for this percent of the requested amount of data to be evicted. For example, by default, for every .Em 2kB that's evicted, .Em 1kB of it may be "reused" by a new allocation. Since this is above .Sy 100 Ns % , it ensures that progress is made towards getting .Sy arc_size No under Sy arc_c . Since this is finite, it ensures that allocations can still happen, even during the potentially long time that .Sy arc_size No is more than Sy arc_c . . .It Sy zfs_arc_evict_batch_limit Ns = Ns Sy 10 Pq int Number ARC headers to evict per sub-list before proceeding to another sub-list. This batch-style operation prevents entire sub-lists from being evicted at once but comes at a cost of additional unlocking and locking. . .It Sy zfs_arc_grow_retry Ns = Ns Sy 0 Ns s Pq int If set to a non zero value, it will replace the .Sy arc_grow_retry value with this value. The .Sy arc_grow_retry .No value Pq default Sy 5 Ns s is the number of seconds the ARC will wait before trying to resume growth after a memory pressure event. . .It Sy zfs_arc_lotsfree_percent Ns = Ns Sy 10 Ns % Pq int Throttle I/O when free system memory drops below this percentage of total system memory. Setting this value to .Sy 0 will disable the throttle. . .It Sy zfs_arc_max Ns = Ns Sy 0 Ns B Pq ulong Max size of ARC in bytes. If .Sy 0 , then the max size of ARC is determined by the amount of system memory installed. Under Linux, half of system memory will be used as the limit. Under .Fx , the larger of .Sy all_system_memory - 1GB No and Sy 5/8 * all_system_memory will be used as the limit. This value must be at least .Sy 67108864 Ns B Pq 64MB . .Pp This value can be changed dynamically, with some caveats. It cannot be set back to .Sy 0 while running, and reducing it below the current ARC size will not cause the ARC to shrink without memory pressure to induce shrinking. . .It Sy zfs_arc_meta_adjust_restarts Ns = Ns Sy 4096 Pq ulong The number of restart passes to make while scanning the ARC attempting the free buffers in order to stay below the .Sy fs_arc_meta_limit . This value should not need to be tuned but is available to facilitate performance analysis. . .It Sy zfs_arc_meta_limit Ns = Ns Sy 0 Ns B Pq ulong The maximum allowed size in bytes that metadata buffers are allowed to consume in the ARC. When this limit is reached, metadata buffers will be reclaimed, even if the overall .Sy arc_c_max has not been reached. It defaults to .Sy 0 , which indicates that a percentage based on .Sy zfs_arc_meta_limit_percent of the ARC may be used for metadata. .Pp This value my be changed dynamically, except that must be set to an explicit value .Pq cannot be set back to Sy 0 . . .It Sy zfs_arc_meta_limit_percent Ns = Ns Sy 75 Ns % Pq ulong Percentage of ARC buffers that can be used for metadata. .Pp See also .Sy zfs_arc_meta_limit , which serves a similar purpose but has a higher priority if nonzero. . .It Sy zfs_arc_meta_min Ns = Ns Sy 0 Ns B Pq ulong The minimum allowed size in bytes that metadata buffers may consume in the ARC. . .It Sy zfs_arc_meta_prune Ns = Ns Sy 10000 Pq int The number of dentries and inodes to be scanned looking for entries which can be dropped. This may be required when the ARC reaches the .Sy zfs_arc_meta_limit because dentries and inodes can pin buffers in the ARC. Increasing this value will cause to dentry and inode caches to be pruned more aggressively. Setting this value to .Sy 0 will disable pruning the inode and dentry caches. . .It Sy zfs_arc_meta_strategy Ns = Ns Sy 1 Ns | Ns 0 Pq int Define the strategy for ARC metadata buffer eviction (meta reclaim strategy): .Bl -tag -compact -offset 4n -width "0 (META_ONLY)" .It Sy 0 Pq META_ONLY evict only the ARC metadata buffers .It Sy 1 Pq BALANCED additional data buffers may be evicted if required to evict the required number of metadata buffers. .El . .It Sy zfs_arc_min Ns = Ns Sy 0 Ns B Pq ulong Min size of ARC in bytes. .No If set to Sy 0 , arc_c_min will default to consuming the larger of .Sy 32MB No or Sy all_system_memory/32 . . .It Sy zfs_arc_min_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 1s Pc Pq int Minimum time prefetched blocks are locked in the ARC. . .It Sy zfs_arc_min_prescient_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 6s Pc Pq int Minimum time "prescient prefetched" blocks are locked in the ARC. These blocks are meant to be prefetched fairly aggressively ahead of the code that may use them. . .It Sy zfs_arc_prune_task_threads Ns = Ns Sy 1 Pq int Number of arc_prune threads. .Fx does not need more than one. Linux may theoretically use one per mount point up to number of CPUs, but that was not proven to be useful. . .It Sy zfs_max_missing_tvds Ns = Ns Sy 0 Pq int Number of missing top-level vdevs which will be allowed during pool import (only in read-only mode). . .It Sy zfs_max_nvlist_src_size Ns = Sy 0 Pq ulong Maximum size in bytes allowed to be passed as .Sy zc_nvlist_src_size for ioctls on .Pa /dev/zfs . This prevents a user from causing the kernel to allocate an excessive amount of memory. When the limit is exceeded, the ioctl fails with .Sy EINVAL and a description of the error is sent to the .Pa zfs-dbgmsg log. This parameter should not need to be touched under normal circumstances. If .Sy 0 , equivalent to a quarter of the user-wired memory limit under .Fx and to .Sy 134217728 Ns B Pq 128MB under Linux. . .It Sy zfs_multilist_num_sublists Ns = Ns Sy 0 Pq int To allow more fine-grained locking, each ARC state contains a series of lists for both data and metadata objects. Locking is performed at the level of these "sub-lists". This parameters controls the number of sub-lists per ARC state, and also applies to other uses of the multilist data structure. .Pp If .Sy 0 , equivalent to the greater of the number of online CPUs and .Sy 4 . . .It Sy zfs_arc_overflow_shift Ns = Ns Sy 8 Pq int The ARC size is considered to be overflowing if it exceeds the current ARC target size .Pq Sy arc_c by thresholds determined by this parameter. Exceeding by .Sy ( arc_c >> zfs_arc_overflow_shift ) * 0.5 starts ARC reclamation process. If that appears insufficient, exceeding by .Sy ( arc_c >> zfs_arc_overflow_shift ) * 1.5 blocks new buffer allocation until the reclaim thread catches up. Started reclamation process continues till ARC size returns below the target size. .Pp The default value of .Sy 8 causes the ARC to start reclamation if it exceeds the target size by .Em 0.2% of the target size, and block allocations by .Em 0.6% . . .It Sy zfs_arc_p_min_shift Ns = Ns Sy 0 Pq int If nonzero, this will update .Sy arc_p_min_shift Pq default Sy 4 with the new value. .Sy arc_p_min_shift No is used as a shift of Sy arc_c when calculating the minumum .Sy arc_p No size. . .It Sy zfs_arc_p_dampener_disable Ns = Ns Sy 1 Ns | Ns 0 Pq int Disable .Sy arc_p adapt dampener, which reduces the maximum single adjustment to .Sy arc_p . . .It Sy zfs_arc_shrink_shift Ns = Ns Sy 0 Pq int If nonzero, this will update .Sy arc_shrink_shift Pq default Sy 7 with the new value. . .It Sy zfs_arc_pc_percent Ns = Ns Sy 0 Ns % Po off Pc Pq uint Percent of pagecache to reclaim ARC to. .Pp This tunable allows the ZFS ARC to play more nicely with the kernel's LRU pagecache. It can guarantee that the ARC size won't collapse under scanning pressure on the pagecache, yet still allows the ARC to be reclaimed down to .Sy zfs_arc_min if necessary. This value is specified as percent of pagecache size (as measured by .Sy NR_FILE_PAGES ) , where that percent may exceed .Sy 100 . This only operates during memory pressure/reclaim. . .It Sy zfs_arc_shrinker_limit Ns = Ns Sy 10000 Pq int This is a limit on how many pages the ARC shrinker makes available for eviction in response to one page allocation attempt. Note that in practice, the kernel's shrinker can ask us to evict up to about four times this for one allocation attempt. .Pp The default limit of .Sy 10000 Pq in practice, Em 160MB No per allocation attempt with 4kB pages limits the amount of time spent attempting to reclaim ARC memory to less than 100ms per allocation attempt, even with a small average compressed block size of ~8kB. .Pp The parameter can be set to 0 (zero) to disable the limit, and only applies on Linux. . .It Sy zfs_arc_sys_free Ns = Ns Sy 0 Ns B Pq ulong The target number of bytes the ARC should leave as free memory on the system. If zero, equivalent to the bigger of .Sy 512kB No and Sy all_system_memory/64 . . .It Sy zfs_autoimport_disable Ns = Ns Sy 1 Ns | Ns 0 Pq int Disable pool import at module load by ignoring the cache file .Pq Sy spa_config_path . . .It Sy zfs_checksum_events_per_second Ns = Ns Sy 20 Ns /s Pq uint Rate limit checksum events to this many per second. Note that this should not be set below the ZED thresholds (currently 10 checksums over 10 seconds) or else the daemon may not trigger any action. . .It Sy zfs_commit_timeout_pct Ns = Ns Sy 5 Ns % Pq int This controls the amount of time that a ZIL block (lwb) will remain "open" when it isn't "full", and it has a thread waiting for it to be committed to stable storage. The timeout is scaled based on a percentage of the last lwb latency to avoid significantly impacting the latency of each individual transaction record (itx). . .It Sy zfs_condense_indirect_commit_entry_delay_ms Ns = Ns Sy 0 Ns ms Pq int Vdev indirection layer (used for device removal) sleeps for this many milliseconds during mapping generation. Intended for use with the test suite to throttle vdev removal speed. . .It Sy zfs_condense_indirect_obsolete_pct Ns = Ns Sy 25 Ns % Pq int Minimum percent of obsolete bytes in vdev mapping required to attempt to condense .Pq see Sy zfs_condense_indirect_vdevs_enable . Intended for use with the test suite to facilitate triggering condensing as needed. . .It Sy zfs_condense_indirect_vdevs_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable condensing indirect vdev mappings. When set, attempt to condense indirect vdev mappings if the mapping uses more than .Sy zfs_condense_min_mapping_bytes bytes of memory and if the obsolete space map object uses more than .Sy zfs_condense_max_obsolete_bytes bytes on-disk. The condensing process is an attempt to save memory by removing obsolete mappings. . .It Sy zfs_condense_max_obsolete_bytes Ns = Ns Sy 1073741824 Ns B Po 1GB Pc Pq ulong Only attempt to condense indirect vdev mappings if the on-disk size of the obsolete space map object is greater than this number of bytes .Pq see Sy zfs_condense_indirect_vdevs_enable . . .It Sy zfs_condense_min_mapping_bytes Ns = Ns Sy 131072 Ns B Po 128kB Pc Pq ulong Minimum size vdev mapping to attempt to condense .Pq see Sy zfs_condense_indirect_vdevs_enable . . .It Sy zfs_dbgmsg_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int Internally ZFS keeps a small log to facilitate debugging. The log is enabled by default, and can be disabled by unsetting this option. The contents of the log can be accessed by reading .Pa /proc/spl/kstat/zfs/dbgmsg . Writing .Sy 0 to the file clears the log. .Pp This setting does not influence debug prints due to .Sy zfs_flags . . .It Sy zfs_dbgmsg_maxsize Ns = Ns Sy 4194304 Ns B Po 4MB Pc Pq int Maximum size of the internal ZFS debug log. . .It Sy zfs_dbuf_state_index Ns = Ns Sy 0 Pq int Historically used for controlling what reporting was available under .Pa /proc/spl/kstat/zfs . No effect. . .It Sy zfs_deadman_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int When a pool sync operation takes longer than .Sy zfs_deadman_synctime_ms , or when an individual I/O operation takes longer than .Sy zfs_deadman_ziotime_ms , then the operation is considered to be "hung". If .Sy zfs_deadman_enabled is set, then the deadman behavior is invoked as described by .Sy zfs_deadman_failmode . By default, the deadman is enabled and set to .Sy wait which results in "hung" I/Os only being logged. The deadman is automatically disabled when a pool gets suspended. . .It Sy zfs_deadman_failmode Ns = Ns Sy wait Pq charp Controls the failure behavior when the deadman detects a "hung" I/O operation. Valid values are: .Bl -tag -compact -offset 4n -width "continue" .It Sy wait Wait for a "hung" operation to complete. For each "hung" operation a "deadman" event will be posted describing that operation. .It Sy continue Attempt to recover from a "hung" operation by re-dispatching it to the I/O pipeline if possible. .It Sy panic Panic the system. This can be used to facilitate automatic fail-over to a properly configured fail-over partner. .El . .It Sy zfs_deadman_checktime_ms Ns = Ns Sy 60000 Ns ms Po 1min Pc Pq int Check time in milliseconds. This defines the frequency at which we check for hung I/O requests and potentially invoke the .Sy zfs_deadman_failmode behavior. . .It Sy zfs_deadman_synctime_ms Ns = Ns Sy 600000 Ns ms Po 10min Pc Pq ulong Interval in milliseconds after which the deadman is triggered and also the interval after which a pool sync operation is considered to be "hung". Once this limit is exceeded the deadman will be invoked every .Sy zfs_deadman_checktime_ms milliseconds until the pool sync completes. . .It Sy zfs_deadman_ziotime_ms Ns = Ns Sy 300000 Ns ms Po 5min Pc Pq ulong Interval in milliseconds after which the deadman is triggered and an individual I/O operation is considered to be "hung". As long as the operation remains "hung", the deadman will be invoked every .Sy zfs_deadman_checktime_ms milliseconds until the operation completes. . .It Sy zfs_dedup_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int Enable prefetching dedup-ed blocks which are going to be freed. . .It Sy zfs_delay_min_dirty_percent Ns = Ns Sy 60 Ns % Pq int Start to delay each transaction once there is this amount of dirty data, expressed as a percentage of .Sy zfs_dirty_data_max . This value should be at least .Sy zfs_vdev_async_write_active_max_dirty_percent . .No See Sx ZFS TRANSACTION DELAY . . .It Sy zfs_delay_scale Ns = Ns Sy 500000 Pq int This controls how quickly the transaction delay approaches infinity. Larger values cause longer delays for a given amount of dirty data. .Pp For the smoothest delay, this value should be about 1 billion divided by the maximum number of operations per second. This will smoothly handle between ten times and a tenth of this number. .No See Sx ZFS TRANSACTION DELAY . .Pp .Sy zfs_delay_scale * zfs_dirty_data_max Em must be smaller than Sy 2^64 . . .It Sy zfs_disable_ivset_guid_check Ns = Ns Sy 0 Ns | Ns 1 Pq int Disables requirement for IVset GUIDs to be present and match when doing a raw receive of encrypted datasets. Intended for users whose pools were created with OpenZFS pre-release versions and now have compatibility issues. . .It Sy zfs_key_max_salt_uses Ns = Ns Sy 400000000 Po 4*10^8 Pc Pq ulong Maximum number of uses of a single salt value before generating a new one for encrypted datasets. The default value is also the maximum. . .It Sy zfs_object_mutex_size Ns = Ns Sy 64 Pq uint Size of the znode hashtable used for holds. .Pp Due to the need to hold locks on objects that may not exist yet, kernel mutexes are not created per-object and instead a hashtable is used where collisions will result in objects waiting when there is not actually contention on the same object. . .It Sy zfs_slow_io_events_per_second Ns = Ns Sy 20 Ns /s Pq int Rate limit delay and deadman zevents (which report slow I/Os) to this many per second. . .It Sy zfs_unflushed_max_mem_amt Ns = Ns Sy 1073741824 Ns B Po 1GB Pc Pq ulong Upper-bound limit for unflushed metadata changes to be held by the log spacemap in memory, in bytes. . .It Sy zfs_unflushed_max_mem_ppm Ns = Ns Sy 1000 Ns ppm Po 0.1% Pc Pq ulong Part of overall system memory that ZFS allows to be used for unflushed metadata changes by the log spacemap, in millionths. . .It Sy zfs_unflushed_log_block_max Ns = Ns Sy 131072 Po 128k Pc Pq ulong Describes the maximum number of log spacemap blocks allowed for each pool. The default value means that the space in all the log spacemaps can add up to no more than .Sy 131072 blocks (which means .Em 16GB of logical space before compression and ditto blocks, assuming that blocksize is .Em 128kB ) . .Pp This tunable is important because it involves a trade-off between import time after an unclean export and the frequency of flushing metaslabs. The higher this number is, the more log blocks we allow when the pool is active which means that we flush metaslabs less often and thus decrease the number of I/Os for spacemap updates per TXG. At the same time though, that means that in the event of an unclean export, there will be more log spacemap blocks for us to read, inducing overhead in the import time of the pool. The lower the number, the amount of flushing increases, destroying log blocks quicker as they become obsolete faster, which leaves less blocks to be read during import time after a crash. .Pp Each log spacemap block existing during pool import leads to approximately one extra logical I/O issued. This is the reason why this tunable is exposed in terms of blocks rather than space used. . .It Sy zfs_unflushed_log_block_min Ns = Ns Sy 1000 Pq ulong If the number of metaslabs is small and our incoming rate is high, we could get into a situation that we are flushing all our metaslabs every TXG. Thus we always allow at least this many log blocks. . .It Sy zfs_unflushed_log_block_pct Ns = Ns Sy 400 Ns % Pq ulong Tunable used to determine the number of blocks that can be used for the spacemap log, expressed as a percentage of the total number of unflushed metaslabs in the pool. . .It Sy zfs_unflushed_log_txg_max Ns = Ns Sy 1000 Pq ulong Tunable limiting maximum time in TXGs any metaslab may remain unflushed. It effectively limits maximum number of unflushed per-TXG spacemap logs that need to be read after unclean pool export. . .It Sy zfs_unlink_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq uint When enabled, files will not be asynchronously removed from the list of pending unlinks and the space they consume will be leaked. Once this option has been disabled and the dataset is remounted, the pending unlinks will be processed and the freed space returned to the pool. This option is used by the test suite. . .It Sy zfs_delete_blocks Ns = Ns Sy 20480 Pq ulong This is the used to define a large file for the purposes of deletion. Files containing more than .Sy zfs_delete_blocks will be deleted asynchronously, while smaller files are deleted synchronously. Decreasing this value will reduce the time spent in an .Xr unlink 2 system call, at the expense of a longer delay before the freed space is available. . .It Sy zfs_dirty_data_max Ns = Pq int Determines the dirty space limit in bytes. Once this limit is exceeded, new writes are halted until space frees up. This parameter takes precedence over .Sy zfs_dirty_data_max_percent . .No See Sx ZFS TRANSACTION DELAY . .Pp Defaults to .Sy physical_ram/10 , capped at .Sy zfs_dirty_data_max_max . . .It Sy zfs_dirty_data_max_max Ns = Pq int Maximum allowable value of .Sy zfs_dirty_data_max , expressed in bytes. This limit is only enforced at module load time, and will be ignored if .Sy zfs_dirty_data_max is later changed. This parameter takes precedence over .Sy zfs_dirty_data_max_max_percent . .No See Sx ZFS TRANSACTION DELAY . .Pp Defaults to .Sy physical_ram/4 , . .It Sy zfs_dirty_data_max_max_percent Ns = Ns Sy 25 Ns % Pq int Maximum allowable value of .Sy zfs_dirty_data_max , expressed as a percentage of physical RAM. This limit is only enforced at module load time, and will be ignored if .Sy zfs_dirty_data_max is later changed. The parameter .Sy zfs_dirty_data_max_max takes precedence over this one. .No See Sx ZFS TRANSACTION DELAY . . .It Sy zfs_dirty_data_max_percent Ns = Ns Sy 10 Ns % Pq int Determines the dirty space limit, expressed as a percentage of all memory. Once this limit is exceeded, new writes are halted until space frees up. The parameter .Sy zfs_dirty_data_max takes precedence over this one. .No See Sx ZFS TRANSACTION DELAY . .Pp Subject to .Sy zfs_dirty_data_max_max . . .It Sy zfs_dirty_data_sync_percent Ns = Ns Sy 20 Ns % Pq int Start syncing out a transaction group if there's at least this much dirty data .Pq as a percentage of Sy zfs_dirty_data_max . This should be less than .Sy zfs_vdev_async_write_active_min_dirty_percent . . .It Sy zfs_wrlog_data_max Ns = Pq int The upper limit of write-transaction zil log data size in bytes. Write operations are throttled when approaching the limit until log data is cleared out after transaction group sync. Because of some overhead, it should be set at least 2 times the size of .Sy zfs_dirty_data_max .No to prevent harming normal write throughput. It also should be smaller than the size of the slog device if slog is present. .Pp Defaults to .Sy zfs_dirty_data_max*2 . .It Sy zfs_fallocate_reserve_percent Ns = Ns Sy 110 Ns % Pq uint Since ZFS is a copy-on-write filesystem with snapshots, blocks cannot be preallocated for a file in order to guarantee that later writes will not run out of space. Instead, .Xr fallocate 2 space preallocation only checks that sufficient space is currently available in the pool or the user's project quota allocation, and then creates a sparse file of the requested size. The requested space is multiplied by .Sy zfs_fallocate_reserve_percent to allow additional space for indirect blocks and other internal metadata. Setting this to .Sy 0 disables support for .Xr fallocate 2 and causes it to return .Sy EOPNOTSUPP . . .It Sy zfs_fletcher_4_impl Ns = Ns Sy fastest Pq string Select a fletcher 4 implementation. .Pp Supported selectors are: .Sy fastest , scalar , sse2 , ssse3 , avx2 , avx512f , avx512bw , .No and Sy aarch64_neon . All except .Sy fastest No and Sy scalar require instruction set extensions to be available, and will only appear if ZFS detects that they are present at runtime. If multiple implementations of fletcher 4 are available, the .Sy fastest will be chosen using a micro benchmark. Selecting .Sy scalar results in the original CPU-based calculation being used. Selecting any option other than .Sy fastest No or Sy scalar results in vector instructions from the respective CPU instruction set being used. . .It Sy zfs_free_bpobj_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable/disable the processing of the free_bpobj object. . .It Sy zfs_async_block_max_blocks Ns = Ns Sy ULONG_MAX Po unlimited Pc Pq ulong Maximum number of blocks freed in a single TXG. . .It Sy zfs_max_async_dedup_frees Ns = Ns Sy 100000 Po 10^5 Pc Pq ulong Maximum number of dedup blocks freed in a single TXG. . .It Sy zfs_override_estimate_recordsize Ns = Ns Sy 0 Pq ulong If nonzer, override record size calculation for .Nm zfs Cm send estimates. . .It Sy zfs_vdev_async_read_max_active Ns = Ns Sy 3 Pq int Maximum asynchronous read I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_read_min_active Ns = Ns Sy 1 Pq int Minimum asynchronous read I/O operation active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_active_max_dirty_percent Ns = Ns Sy 60 Ns % Pq int When the pool has more than this much dirty data, use .Sy zfs_vdev_async_write_max_active to limit active async writes. If the dirty data is between the minimum and maximum, the active I/O limit is linearly interpolated. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_active_min_dirty_percent Ns = Ns Sy 30 Ns % Pq int When the pool has less than this much dirty data, use .Sy zfs_vdev_async_write_min_active to limit active async writes. If the dirty data is between the minimum and maximum, the active I/O limit is linearly interpolated. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_max_active Ns = Ns Sy 30 Pq int Maximum asynchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_min_active Ns = Ns Sy 2 Pq int Minimum asynchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . .Pp Lower values are associated with better latency on rotational media but poorer resilver performance. The default value of .Sy 2 was chosen as a compromise. A value of .Sy 3 has been shown to improve resilver performance further at a cost of further increasing latency. . .It Sy zfs_vdev_initializing_max_active Ns = Ns Sy 1 Pq int Maximum initializing I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_initializing_min_active Ns = Ns Sy 1 Pq int Minimum initializing I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_max_active Ns = Ns Sy 1000 Pq int The maximum number of I/O operations active to each device. Ideally, this will be at least the sum of each queue's .Sy max_active . .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_open_timeout_ms Ns = Ns Sy 1000 Pq uint Timeout value to wait before determining a device is missing during import. This is helpful for transient missing paths due to links being briefly removed and recreated in response to udev events. . .It Sy zfs_vdev_rebuild_max_active Ns = Ns Sy 3 Pq int Maximum sequential resilver I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_rebuild_min_active Ns = Ns Sy 1 Pq int Minimum sequential resilver I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_removal_max_active Ns = Ns Sy 2 Pq int Maximum removal I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_removal_min_active Ns = Ns Sy 1 Pq int Minimum removal I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_scrub_max_active Ns = Ns Sy 2 Pq int Maximum scrub I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_scrub_min_active Ns = Ns Sy 1 Pq int Minimum scrub I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_read_max_active Ns = Ns Sy 10 Pq int Maximum synchronous read I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_read_min_active Ns = Ns Sy 10 Pq int Minimum synchronous read I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_write_max_active Ns = Ns Sy 10 Pq int Maximum synchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_write_min_active Ns = Ns Sy 10 Pq int Minimum synchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_trim_max_active Ns = Ns Sy 2 Pq int Maximum trim/discard I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_trim_min_active Ns = Ns Sy 1 Pq int Minimum trim/discard I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_nia_delay Ns = Ns Sy 5 Pq int For non-interactive I/O (scrub, resilver, removal, initialize and rebuild), the number of concurrently-active I/O operations is limited to .Sy zfs_*_min_active , unless the vdev is "idle". When there are no interactive I/O operatinons active (synchronous or otherwise), and .Sy zfs_vdev_nia_delay operations have completed since the last interactive operation, then the vdev is considered to be "idle", and the number of concurrently-active non-interactive operations is increased to .Sy zfs_*_max_active . .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_nia_credit Ns = Ns Sy 5 Pq int Some HDDs tend to prioritize sequential I/O so strongly, that concurrent random I/O latency reaches several seconds. On some HDDs this happens even if sequential I/O operations are submitted one at a time, and so setting .Sy zfs_*_max_active Ns = Sy 1 does not help. To prevent non-interactive I/O, like scrub, from monopolizing the device, no more than .Sy zfs_vdev_nia_credit operations can be sent while there are outstanding incomplete interactive operations. This enforced wait ensures the HDD services the interactive I/O within a reasonable amount of time. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_queue_depth_pct Ns = Ns Sy 1000 Ns % Pq int Maximum number of queued allocations per top-level vdev expressed as a percentage of .Sy zfs_vdev_async_write_max_active , which allows the system to detect devices that are more capable of handling allocations and to allocate more blocks to those devices. This allows for dynamic allocation distribution when devices are imbalanced, as fuller devices will tend to be slower than empty devices. .Pp Also see .Sy zio_dva_throttle_enabled . . .It Sy zfs_expire_snapshot Ns = Ns Sy 300 Ns s Pq int Time before expiring .Pa .zfs/snapshot . . .It Sy zfs_admin_snapshot Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow the creation, removal, or renaming of entries in the .Sy .zfs/snapshot directory to cause the creation, destruction, or renaming of snapshots. When enabled, this functionality works both locally and over NFS exports which have the .Em no_root_squash option set. . .It Sy zfs_flags Ns = Ns Sy 0 Pq int Set additional debugging flags. The following flags may be bitwise-ored together: .TS box; lbz r l l . Value Symbolic Name Description _ 1 ZFS_DEBUG_DPRINTF Enable dprintf entries in the debug log. * 2 ZFS_DEBUG_DBUF_VERIFY Enable extra dbuf verifications. * 4 ZFS_DEBUG_DNODE_VERIFY Enable extra dnode verifications. 8 ZFS_DEBUG_SNAPNAMES Enable snapshot name verification. 16 ZFS_DEBUG_MODIFY Check for illegally modified ARC buffers. 64 ZFS_DEBUG_ZIO_FREE Enable verification of block frees. 128 ZFS_DEBUG_HISTOGRAM_VERIFY Enable extra spacemap histogram verifications. 256 ZFS_DEBUG_METASLAB_VERIFY Verify space accounting on disk matches in-memory \fBrange_trees\fP. 512 ZFS_DEBUG_SET_ERROR Enable \fBSET_ERROR\fP and dprintf entries in the debug log. 1024 ZFS_DEBUG_INDIRECT_REMAP Verify split blocks created by device removal. 2048 ZFS_DEBUG_TRIM Verify TRIM ranges are always within the allocatable range tree. 4096 ZFS_DEBUG_LOG_SPACEMAP Verify that the log summary is consistent with the spacemap log and enable \fBzfs_dbgmsgs\fP for metaslab loading and flushing. .TE .Sy \& * No Requires debug build. . .It Sy zfs_btree_verify_intensity Ns = Ns Sy 0 Pq uint Enables btree verification. The following settings are culminative: .TS box; lbz r l l . Value Description 1 Verify height. 2 Verify pointers from children to parent. 3 Verify element counts. 4 Verify element order. (expensive) * 5 Verify unused memory is poisoned. (expensive) .TE .Sy \& * No Requires debug build. . .It Sy zfs_free_leak_on_eio Ns = Ns Sy 0 Ns | Ns 1 Pq int If destroy encounters an .Sy EIO while reading metadata (e.g. indirect blocks), space referenced by the missing metadata can not be freed. Normally this causes the background destroy to become "stalled", as it is unable to make forward progress. While in this stalled state, all remaining space to free from the error-encountering filesystem is "temporarily leaked". Set this flag to cause it to ignore the .Sy EIO , permanently leak the space from indirect blocks that can not be read, and continue to free everything else that it can. .Pp The default "stalling" behavior is useful if the storage partially fails (i.e. some but not all I/O operations fail), and then later recovers. In this case, we will be able to continue pool operations while it is partially failed, and when it recovers, we can continue to free the space, with no leaks. Note, however, that this case is actually fairly rare. .Pp Typically pools either .Bl -enum -compact -offset 4n -width "1." .It fail completely (but perhaps temporarily, e.g. due to a top-level vdev going offline), or .It have localized, permanent errors (e.g. disk returns the wrong data due to bit flip or firmware bug). .El In the former case, this setting does not matter because the pool will be suspended and the sync thread will not be able to make forward progress regardless. In the latter, because the error is permanent, the best we can do is leak the minimum amount of space, which is what setting this flag will do. It is therefore reasonable for this flag to normally be set, but we chose the more conservative approach of not setting it, so that there is no possibility of leaking space in the "partial temporary" failure case. . .It Sy zfs_free_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1s Pc Pq int During a .Nm zfs Cm destroy operation using the .Sy async_destroy feature, a minimum of this much time will be spent working on freeing blocks per TXG. . .It Sy zfs_obsolete_min_time_ms Ns = Ns Sy 500 Ns ms Pq int Similar to .Sy zfs_free_min_time_ms , but for cleanup of old indirection records for removed vdevs. . .It Sy zfs_immediate_write_sz Ns = Ns Sy 32768 Ns B Po 32kB Pc Pq long Largest data block to write to the ZIL. Larger blocks will be treated as if the dataset being written to had the .Sy logbias Ns = Ns Sy throughput property set. . .It Sy zfs_initialize_value Ns = Ns Sy 16045690984833335022 Po 0xDEADBEEFDEADBEEE Pc Pq ulong Pattern written to vdev free space by .Xr zpool-initialize 8 . . .It Sy zfs_initialize_chunk_size Ns = Ns Sy 1048576 Ns B Po 1MB Pc Pq ulong Size of writes used by .Xr zpool-initialize 8 . This option is used by the test suite. . .It Sy zfs_livelist_max_entries Ns = Ns Sy 500000 Po 5*10^5 Pc Pq ulong The threshold size (in block pointers) at which we create a new sub-livelist. Larger sublists are more costly from a memory perspective but the fewer sublists there are, the lower the cost of insertion. . .It Sy zfs_livelist_min_percent_shared Ns = Ns Sy 75 Ns % Pq int If the amount of shared space between a snapshot and its clone drops below this threshold, the clone turns off the livelist and reverts to the old deletion method. This is in place because livelists no long give us a benefit once a clone has been overwritten enough. . .It Sy zfs_livelist_condense_new_alloc Ns = Ns Sy 0 Pq int Incremented each time an extra ALLOC blkptr is added to a livelist entry while it is being condensed. This option is used by the test suite to track race conditions. . .It Sy zfs_livelist_condense_sync_cancel Ns = Ns Sy 0 Pq int Incremented each time livelist condensing is canceled while in .Fn spa_livelist_condense_sync . This option is used by the test suite to track race conditions. . .It Sy zfs_livelist_condense_sync_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int When set, the livelist condense process pauses indefinitely before executing the synctask - .Fn spa_livelist_condense_sync . This option is used by the test suite to trigger race conditions. . .It Sy zfs_livelist_condense_zthr_cancel Ns = Ns Sy 0 Pq int Incremented each time livelist condensing is canceled while in .Fn spa_livelist_condense_cb . This option is used by the test suite to track race conditions. . .It Sy zfs_livelist_condense_zthr_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int When set, the livelist condense process pauses indefinitely before executing the open context condensing work in .Fn spa_livelist_condense_cb . This option is used by the test suite to trigger race conditions. . .It Sy zfs_lua_max_instrlimit Ns = Ns Sy 100000000 Po 10^8 Pc Pq ulong The maximum execution time limit that can be set for a ZFS channel program, specified as a number of Lua instructions. . .It Sy zfs_lua_max_memlimit Ns = Ns Sy 104857600 Po 100MB Pc Pq ulong The maximum memory limit that can be set for a ZFS channel program, specified in bytes. . .It Sy zfs_max_dataset_nesting Ns = Ns Sy 50 Pq int The maximum depth of nested datasets. This value can be tuned temporarily to fix existing datasets that exceed the predefined limit. . .It Sy zfs_max_log_walking Ns = Ns Sy 5 Pq ulong The number of past TXGs that the flushing algorithm of the log spacemap feature uses to estimate incoming log blocks. . .It Sy zfs_max_logsm_summary_length Ns = Ns Sy 10 Pq ulong Maximum number of rows allowed in the summary of the spacemap log. . .It Sy zfs_max_recordsize Ns = Ns Sy 1048576 Po 1MB Pc Pq int We currently support block sizes from .Em 512B No to Em 16MB . The benefits of larger blocks, and thus larger I/O, need to be weighed against the cost of COWing a giant block to modify one byte. Additionally, very large blocks can have an impact on I/O latency, and also potentially on the memory allocator. Therefore, we do not allow the recordsize to be set larger than this tunable. Larger blocks can be created by changing it, and pools with larger blocks can always be imported and used, regardless of this setting. . .It Sy zfs_allow_redacted_dataset_mount Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow datasets received with redacted send/receive to be mounted. Normally disabled because these datasets may be missing key data. . .It Sy zfs_min_metaslabs_to_flush Ns = Ns Sy 1 Pq ulong Minimum number of metaslabs to flush per dirty TXG. . .It Sy zfs_metaslab_fragmentation_threshold Ns = Ns Sy 70 Ns % Pq int Allow metaslabs to keep their active state as long as their fragmentation percentage is no more than this value. An active metaslab that exceeds this threshold will no longer keep its active status allowing better metaslabs to be selected. . .It Sy zfs_mg_fragmentation_threshold Ns = Ns Sy 95 Ns % Pq int Metaslab groups are considered eligible for allocations if their fragmentation metric (measured as a percentage) is less than or equal to this value. If a metaslab group exceeds this threshold then it will be skipped unless all metaslab groups within the metaslab class have also crossed this threshold. . .It Sy zfs_mg_noalloc_threshold Ns = Ns Sy 0 Ns % Pq int Defines a threshold at which metaslab groups should be eligible for allocations. The value is expressed as a percentage of free space beyond which a metaslab group is always eligible for allocations. If a metaslab group's free space is less than or equal to the threshold, the allocator will avoid allocating to that group unless all groups in the pool have reached the threshold. Once all groups have reached the threshold, all groups are allowed to accept allocations. The default value of .Sy 0 disables the feature and causes all metaslab groups to be eligible for allocations. .Pp This parameter allows one to deal with pools having heavily imbalanced vdevs such as would be the case when a new vdev has been added. Setting the threshold to a non-zero percentage will stop allocations from being made to vdevs that aren't filled to the specified percentage and allow lesser filled vdevs to acquire more allocations than they otherwise would under the old .Sy zfs_mg_alloc_failures facility. . .It Sy zfs_ddt_data_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int If enabled, ZFS will place DDT data into the special allocation class. . .It Sy zfs_user_indirect_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int If enabled, ZFS will place user data indirect blocks into the special allocation class. . .It Sy zfs_multihost_history Ns = Ns Sy 0 Pq int Historical statistics for this many latest multihost updates will be available in .Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /multihost . . .It Sy zfs_multihost_interval Ns = Ns Sy 1000 Ns ms Po 1s Pc Pq ulong Used to control the frequency of multihost writes which are performed when the .Sy multihost pool property is on. This is one of the factors used to determine the length of the activity check during import. .Pp The multihost write period is .Sy zfs_multihost_interval / leaf-vdevs . On average a multihost write will be issued for each leaf vdev every .Sy zfs_multihost_interval milliseconds. In practice, the observed period can vary with the I/O load and this observed value is the delay which is stored in the uberblock. . .It Sy zfs_multihost_import_intervals Ns = Ns Sy 20 Pq uint Used to control the duration of the activity test on import. Smaller values of .Sy zfs_multihost_import_intervals will reduce the import time but increase the risk of failing to detect an active pool. The total activity check time is never allowed to drop below one second. .Pp On import the activity check waits a minimum amount of time determined by .Sy zfs_multihost_interval * zfs_multihost_import_intervals , or the same product computed on the host which last had the pool imported, whichever is greater. The activity check time may be further extended if the value of MMP delay found in the best uberblock indicates actual multihost updates happened at longer intervals than .Sy zfs_multihost_interval . A minimum of .Em 100ms is enforced. .Pp .Sy 0 No is equivalent to Sy 1 . . .It Sy zfs_multihost_fail_intervals Ns = Ns Sy 10 Pq uint Controls the behavior of the pool when multihost write failures or delays are detected. .Pp When .Sy 0 , multihost write failures or delays are ignored. The failures will still be reported to the ZED which depending on its configuration may take action such as suspending the pool or offlining a device. .Pp Otherwise, the pool will be suspended if .Sy zfs_multihost_fail_intervals * zfs_multihost_interval milliseconds pass without a successful MMP write. This guarantees the activity test will see MMP writes if the pool is imported. .Sy 1 No is equivalent to Sy 2 ; this is necessary to prevent the pool from being suspended due to normal, small I/O latency variations. . .It Sy zfs_no_scrub_io Ns = Ns Sy 0 Ns | Ns 1 Pq int Set to disable scrub I/O. This results in scrubs not actually scrubbing data and simply doing a metadata crawl of the pool instead. . .It Sy zfs_no_scrub_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int Set to disable block prefetching for scrubs. . .It Sy zfs_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable cache flush operations on disks when writing. Setting this will cause pool corruption on power loss if a volatile out-of-order write cache is enabled. . .It Sy zfs_nopwrite_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Allow no-operation writes. The occurrence of nopwrites will further depend on other pool properties .Pq i.a. the checksumming and compression algorithms . . .It Sy zfs_dmu_offset_next_sync Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable forcing TXG sync to find holes. When enabled forces ZFS to sync data when .Sy SEEK_HOLE No or Sy SEEK_DATA flags are used allowing holes in a file to be accurately reported. When disabled holes will not be reported in recently dirtied files. . .It Sy zfs_pd_bytes_max Ns = Ns Sy 52428800 Ns B Po 50MB Pc Pq int The number of bytes which should be prefetched during a pool traversal, like .Nm zfs Cm send or other data crawling operations. . .It Sy zfs_traverse_indirect_prefetch_limit Ns = Ns Sy 32 Pq int The number of blocks pointed by indirect (non-L0) block which should be prefetched during a pool traversal, like .Nm zfs Cm send or other data crawling operations. . .It Sy zfs_per_txg_dirty_frees_percent Ns = Ns Sy 30 Ns % Pq ulong Control percentage of dirtied indirect blocks from frees allowed into one TXG. After this threshold is crossed, additional frees will wait until the next TXG. .Sy 0 No disables this throttle. . .It Sy zfs_prefetch_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable predictive prefetch. Note that it leaves "prescient" prefetch (for. e.g.\& .Nm zfs Cm send ) intact. Unlike predictive prefetch, prescient prefetch never issues I/O that ends up not being needed, so it can't hurt performance. . .It Sy zfs_qat_checksum_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable QAT hardware acceleration for SHA256 checksums. May be unset after the ZFS modules have been loaded to initialize the QAT hardware as long as support is compiled in and the QAT driver is present. . .It Sy zfs_qat_compress_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable QAT hardware acceleration for gzip compression. May be unset after the ZFS modules have been loaded to initialize the QAT hardware as long as support is compiled in and the QAT driver is present. . .It Sy zfs_qat_encrypt_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable QAT hardware acceleration for AES-GCM encryption. May be unset after the ZFS modules have been loaded to initialize the QAT hardware as long as support is compiled in and the QAT driver is present. . .It Sy zfs_vnops_read_chunk_size Ns = Ns Sy 1048576 Ns B Po 1MB Pc Pq long Bytes to read per chunk. . .It Sy zfs_read_history Ns = Ns Sy 0 Pq int Historical statistics for this many latest reads will be available in .Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /reads . . .It Sy zfs_read_history_hits Ns = Ns Sy 0 Ns | Ns 1 Pq int Include cache hits in read history . .It Sy zfs_rebuild_max_segment Ns = Ns Sy 1048576 Ns B Po 1MB Pc Pq ulong Maximum read segment size to issue when sequentially resilvering a top-level vdev. . .It Sy zfs_rebuild_scrub_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Automatically start a pool scrub when the last active sequential resilver completes in order to verify the checksums of all blocks which have been resilvered. This is enabled by default and strongly recommended. . .It Sy zfs_rebuild_vdev_limit Ns = Ns Sy 33554432 Ns B Po 32MB Pc Pq ulong Maximum amount of I/O that can be concurrently issued for a sequential resilver per leaf device, given in bytes. . .It Sy zfs_reconstruct_indirect_combinations_max Ns = Ns Sy 4096 Pq int If an indirect split block contains more than this many possible unique combinations when being reconstructed, consider it too computationally expensive to check them all. Instead, try at most this many randomly selected combinations each time the block is accessed. This allows all segment copies to participate fairly in the reconstruction when all combinations cannot be checked and prevents repeated use of one bad copy. . .It Sy zfs_recover Ns = Ns Sy 0 Ns | Ns 1 Pq int Set to attempt to recover from fatal errors. This should only be used as a last resort, as it typically results in leaked space, or worse. . .It Sy zfs_removal_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int Ignore hard IO errors during device removal. When set, if a device encounters a hard IO error during the removal process the removal will not be cancelled. This can result in a normally recoverable block becoming permanently damaged and is hence not recommended. This should only be used as a last resort when the pool cannot be returned to a healthy state prior to removing the device. . .It Sy zfs_removal_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq int This is used by the test suite so that it can ensure that certain actions happen while in the middle of a removal. . .It Sy zfs_remove_max_segment Ns = Ns Sy 16777216 Ns B Po 16MB Pc Pq int The largest contiguous segment that we will attempt to allocate when removing a device. If there is a performance problem with attempting to allocate large blocks, consider decreasing this. The default value is also the maximum. . .It Sy zfs_resilver_disable_defer Ns = Ns Sy 0 Ns | Ns 1 Pq int Ignore the .Sy resilver_defer feature, causing an operation that would start a resilver to immediately restart the one in progress. . .It Sy zfs_resilver_min_time_ms Ns = Ns Sy 3000 Ns ms Po 3s Pc Pq int Resilvers are processed by the sync thread. While resilvering, it will spend at least this much time working on a resilver between TXG flushes. . .It Sy zfs_scan_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int If set, remove the DTL (dirty time list) upon completion of a pool scan (scrub), even if there were unrepairable errors. Intended to be used during pool repair or recovery to stop resilvering when the pool is next imported. . .It Sy zfs_scrub_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1s Pc Pq int Scrubs are processed by the sync thread. While scrubbing, it will spend at least this much time working on a scrub between TXG flushes. . .It Sy zfs_scan_checkpoint_intval Ns = Ns Sy 7200 Ns s Po 2h Pc Pq int To preserve progress across reboots, the sequential scan algorithm periodically needs to stop metadata scanning and issue all the verification I/O to disk. The frequency of this flushing is determined by this tunable. . .It Sy zfs_scan_fill_weight Ns = Ns Sy 3 Pq int This tunable affects how scrub and resilver I/O segments are ordered. A higher number indicates that we care more about how filled in a segment is, while a lower number indicates we care more about the size of the extent without considering the gaps within a segment. This value is only tunable upon module insertion. Changing the value afterwards will have no affect on scrub or resilver performance. . .It Sy zfs_scan_issue_strategy Ns = Ns Sy 0 Pq int Determines the order that data will be verified while scrubbing or resilvering: .Bl -tag -compact -offset 4n -width "a" .It Sy 1 Data will be verified as sequentially as possible, given the amount of memory reserved for scrubbing .Pq see Sy zfs_scan_mem_lim_fact . This may improve scrub performance if the pool's data is very fragmented. .It Sy 2 The largest mostly-contiguous chunk of found data will be verified first. By deferring scrubbing of small segments, we may later find adjacent data to coalesce and increase the segment size. .It Sy 0 .No Use strategy Sy 1 No during normal verification .No and strategy Sy 2 No while taking a checkpoint. .El . .It Sy zfs_scan_legacy Ns = Ns Sy 0 Ns | Ns 1 Pq int If unset, indicates that scrubs and resilvers will gather metadata in memory before issuing sequential I/O. Otherwise indicates that the legacy algorithm will be used, where I/O is initiated as soon as it is discovered. Unsetting will not affect scrubs or resilvers that are already in progress. . .It Sy zfs_scan_max_ext_gap Ns = Ns Sy 2097152 Ns B Po 2MB Pc Pq int Sets the largest gap in bytes between scrub/resilver I/O operations that will still be considered sequential for sorting purposes. Changing this value will not affect scrubs or resilvers that are already in progress. . .It Sy zfs_scan_mem_lim_fact Ns = Ns Sy 20 Ns ^-1 Pq int Maximum fraction of RAM used for I/O sorting by sequential scan algorithm. This tunable determines the hard limit for I/O sorting memory usage. When the hard limit is reached we stop scanning metadata and start issuing data verification I/O. This is done until we get below the soft limit. . .It Sy zfs_scan_mem_lim_soft_fact Ns = Ns Sy 20 Ns ^-1 Pq int The fraction of the hard limit used to determined the soft limit for I/O sorting by the sequential scan algorithm. When we cross this limit from below no action is taken. When we cross this limit from above it is because we are issuing verification I/O. In this case (unless the metadata scan is done) we stop issuing verification I/O and start scanning metadata again until we get to the hard limit. . +.It Sy zfs_scan_report_txgs Ns = Ns Sy 0 Ns | Ns 1 Pq uint +When reporting resilver throughput and estimated completion time use the +performance observed over roughly the last +.Sy zfs_scan_report_txgs +TXGs. +When set to zero performance is calculated over the time between checkpoints. +. .It Sy zfs_scan_strict_mem_lim Ns = Ns Sy 0 Ns | Ns 1 Pq int Enforce tight memory limits on pool scans when a sequential scan is in progress. When disabled, the memory limit may be exceeded by fast disks. . .It Sy zfs_scan_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq int Freezes a scrub/resilver in progress without actually pausing it. Intended for testing/debugging. . .It Sy zfs_scan_vdev_limit Ns = Ns Sy 4194304 Ns B Po 4MB Pc Pq int Maximum amount of data that can be concurrently issued at once for scrubs and resilvers per leaf device, given in bytes. . .It Sy zfs_send_corrupt_data Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow sending of corrupt data (ignore read/checksum errors when sending). . .It Sy zfs_send_unmodified_spill_blocks Ns = Ns Sy 1 Ns | Ns 0 Pq int Include unmodified spill blocks in the send stream. Under certain circumstances, previous versions of ZFS could incorrectly remove the spill block from an existing object. Including unmodified copies of the spill blocks creates a backwards-compatible stream which will recreate a spill block if it was incorrectly removed. . .It Sy zfs_send_no_prefetch_queue_ff Ns = Ns Sy 20 Ns ^-1 Pq int The fill fraction of the .Nm zfs Cm send internal queues. The fill fraction controls the timing with which internal threads are woken up. . .It Sy zfs_send_no_prefetch_queue_length Ns = Ns Sy 1048576 Ns B Po 1MB Pc Pq int The maximum number of bytes allowed in .Nm zfs Cm send Ns 's internal queues. . .It Sy zfs_send_queue_ff Ns = Ns Sy 20 Ns ^-1 Pq int The fill fraction of the .Nm zfs Cm send prefetch queue. The fill fraction controls the timing with which internal threads are woken up. . .It Sy zfs_send_queue_length Ns = Ns Sy 16777216 Ns B Po 16MB Pc Pq int The maximum number of bytes allowed that will be prefetched by .Nm zfs Cm send . This value must be at least twice the maximum block size in use. . .It Sy zfs_recv_queue_ff Ns = Ns Sy 20 Ns ^-1 Pq int The fill fraction of the .Nm zfs Cm receive queue. The fill fraction controls the timing with which internal threads are woken up. . .It Sy zfs_recv_queue_length Ns = Ns Sy 16777216 Ns B Po 16MB Pc Pq int The maximum number of bytes allowed in the .Nm zfs Cm receive queue. This value must be at least twice the maximum block size in use. . .It Sy zfs_recv_write_batch_size Ns = Ns Sy 1048576 Ns B Po 1MB Pc Pq int The maximum amount of data, in bytes, that .Nm zfs Cm receive will write in one DMU transaction. This is the uncompressed size, even when receiving a compressed send stream. This setting will not reduce the write size below a single block. Capped at a maximum of .Sy 32MB . . .It Sy zfs_override_estimate_recordsize Ns = Ns Sy 0 Ns | Ns 1 Pq ulong Setting this variable overrides the default logic for estimating block sizes when doing a .Nm zfs Cm send . The default heuristic is that the average block size will be the current recordsize. Override this value if most data in your dataset is not of that size and you require accurate zfs send size estimates. . .It Sy zfs_sync_pass_deferred_free Ns = Ns Sy 2 Pq int Flushing of data to disk is done in passes. Defer frees starting in this pass. . .It Sy zfs_spa_discard_memory_limit Ns = Ns Sy 16777216 Ns B Po 16MB Pc Pq int Maximum memory used for prefetching a checkpoint's space map on each vdev while discarding the checkpoint. . .It Sy zfs_special_class_metadata_reserve_pct Ns = Ns Sy 25 Ns % Pq int Only allow small data blocks to be allocated on the special and dedup vdev types when the available free space percentage on these vdevs exceeds this value. This ensures reserved space is available for pool metadata as the special vdevs approach capacity. . .It Sy zfs_sync_pass_dont_compress Ns = Ns Sy 8 Pq int Starting in this sync pass, disable compression (including of metadata). With the default setting, in practice, we don't have this many sync passes, so this has no effect. .Pp The original intent was that disabling compression would help the sync passes to converge. However, in practice, disabling compression increases the average number of sync passes; because when we turn compression off, many blocks' size will change, and thus we have to re-allocate (not overwrite) them. It also increases the number of .Em 128kB allocations (e.g. for indirect blocks and spacemaps) because these will not be compressed. The .Em 128kB allocations are especially detrimental to performance on highly fragmented systems, which may have very few free segments of this size, and may need to load new metaslabs to satisfy these allocations. . .It Sy zfs_sync_pass_rewrite Ns = Ns Sy 2 Pq int Rewrite new block pointers starting in this pass. . .It Sy zfs_sync_taskq_batch_pct Ns = Ns Sy 75 Ns % Pq int This controls the number of threads used by .Sy dp_sync_taskq . The default value of .Sy 75% will create a maximum of one thread per CPU. . .It Sy zfs_trim_extent_bytes_max Ns = Ns Sy 134217728 Ns B Po 128MB Pc Pq uint Maximum size of TRIM command. Larger ranges will be split into chunks no larger than this value before issuing. . .It Sy zfs_trim_extent_bytes_min Ns = Ns Sy 32768 Ns B Po 32kB Pc Pq uint Minimum size of TRIM commands. TRIM ranges smaller than this will be skipped, unless they're part of a larger range which was chunked. This is done because it's common for these small TRIMs to negatively impact overall performance. . .It Sy zfs_trim_metaslab_skip Ns = Ns Sy 0 Ns | Ns 1 Pq uint Skip uninitialized metaslabs during the TRIM process. This option is useful for pools constructed from large thinly-provisioned devices where TRIM operations are slow. As a pool ages, an increasing fraction of the pool's metaslabs will be initialized, progressively degrading the usefulness of this option. This setting is stored when starting a manual TRIM and will persist for the duration of the requested TRIM. . .It Sy zfs_trim_queue_limit Ns = Ns Sy 10 Pq uint Maximum number of queued TRIMs outstanding per leaf vdev. The number of concurrent TRIM commands issued to the device is controlled by .Sy zfs_vdev_trim_min_active No and Sy zfs_vdev_trim_max_active . . .It Sy zfs_trim_txg_batch Ns = Ns Sy 32 Pq uint The number of transaction groups' worth of frees which should be aggregated before TRIM operations are issued to the device. This setting represents a trade-off between issuing larger, more efficient TRIM operations and the delay before the recently trimmed space is available for use by the device. .Pp Increasing this value will allow frees to be aggregated for a longer time. This will result is larger TRIM operations and potentially increased memory usage. Decreasing this value will have the opposite effect. The default of .Sy 32 was determined to be a reasonable compromise. . .It Sy zfs_txg_history Ns = Ns Sy 0 Pq int Historical statistics for this many latest TXGs will be available in .Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /TXGs . . .It Sy zfs_txg_timeout Ns = Ns Sy 5 Ns s Pq int Flush dirty data to disk at least every this many seconds (maximum TXG duration). . .It Sy zfs_vdev_aggregate_trim Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow TRIM I/Os to be aggregated. This is normally not helpful because the extents to be trimmed will have been already been aggregated by the metaslab. This option is provided for debugging and performance analysis. . .It Sy zfs_vdev_aggregation_limit Ns = Ns Sy 1048576 Ns B Po 1MB Pc Pq int Max vdev I/O aggregation size. . .It Sy zfs_vdev_aggregation_limit_non_rotating Ns = Ns Sy 131072 Ns B Po 128kB Pc Pq int Max vdev I/O aggregation size for non-rotating media. . .It Sy zfs_vdev_cache_bshift Ns = Ns Sy 16 Po 64kB Pc Pq int Shift size to inflate reads to. . .It Sy zfs_vdev_cache_max Ns = Ns Sy 16384 Ns B Po 16kB Pc Pq int Inflate reads smaller than this value to meet the .Sy zfs_vdev_cache_bshift size .Pq default Sy 64kB . . .It Sy zfs_vdev_cache_size Ns = Ns Sy 0 Pq int Total size of the per-disk cache in bytes. .Pp Currently this feature is disabled, as it has been found to not be helpful for performance and in some cases harmful. . .It Sy zfs_vdev_mirror_rotating_inc Ns = Ns Sy 0 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member when an I/O operation immediately follows its predecessor on rotational vdevs for the purpose of making decisions based on load. . .It Sy zfs_vdev_mirror_rotating_seek_inc Ns = Ns Sy 5 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member when an I/O operation lacks locality as defined by .Sy zfs_vdev_mirror_rotating_seek_offset . Operations within this that are not immediately following the previous operation are incremented by half. . .It Sy zfs_vdev_mirror_rotating_seek_offset Ns = Ns Sy 1048576 Ns B Po 1MB Pc Pq int The maximum distance for the last queued I/O operation in which the balancing algorithm considers an operation to have locality. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_mirror_non_rotating_inc Ns = Ns Sy 0 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member on non-rotational vdevs when I/O operations do not immediately follow one another. . .It Sy zfs_vdev_mirror_non_rotating_seek_inc Ns = Ns Sy 1 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member when an I/O operation lacks locality as defined by the .Sy zfs_vdev_mirror_rotating_seek_offset . Operations within this that are not immediately following the previous operation are incremented by half. . .It Sy zfs_vdev_read_gap_limit Ns = Ns Sy 32768 Ns B Po 32kB Pc Pq int Aggregate read I/O operations if the on-disk gap between them is within this threshold. . .It Sy zfs_vdev_write_gap_limit Ns = Ns Sy 4096 Ns B Po 4kB Pc Pq int Aggregate write I/O operations if the on-disk gap between them is within this threshold. . .It Sy zfs_vdev_raidz_impl Ns = Ns Sy fastest Pq string Select the raidz parity implementation to use. .Pp Variants that don't depend on CPU-specific features may be selected on module load, as they are supported on all systems. The remaining options may only be set after the module is loaded, as they are available only if the implementations are compiled in and supported on the running system. .Pp Once the module is loaded, .Pa /sys/module/zfs/parameters/zfs_vdev_raidz_impl will show the available options, with the currently selected one enclosed in square brackets. .Pp .TS lb l l . fastest selected by built-in benchmark original original implementation scalar scalar implementation sse2 SSE2 instruction set 64-bit x86 ssse3 SSSE3 instruction set 64-bit x86 avx2 AVX2 instruction set 64-bit x86 avx512f AVX512F instruction set 64-bit x86 avx512bw AVX512F & AVX512BW instruction sets 64-bit x86 aarch64_neon NEON Aarch64/64-bit ARMv8 aarch64_neonx2 NEON with more unrolling Aarch64/64-bit ARMv8 powerpc_altivec Altivec PowerPC .TE . .It Sy zfs_vdev_scheduler Pq charp .Sy DEPRECATED . Prints warning to kernel log for compatibility. . .It Sy zfs_zevent_len_max Ns = Ns Sy 512 Pq int Max event queue length. Events in the queue can be viewed with .Xr zpool-events 8 . . .It Sy zfs_zevent_retain_max Ns = Ns Sy 2000 Pq int Maximum recent zevent records to retain for duplicate checking. Setting this to .Sy 0 disables duplicate detection. . .It Sy zfs_zevent_retain_expire_secs Ns = Ns Sy 900 Ns s Po 15min Pc Pq int Lifespan for a recent ereport that was retained for duplicate checking. . .It Sy zfs_zil_clean_taskq_maxalloc Ns = Ns Sy 1048576 Pq int The maximum number of taskq entries that are allowed to be cached. When this limit is exceeded transaction records (itxs) will be cleaned synchronously. . .It Sy zfs_zil_clean_taskq_minalloc Ns = Ns Sy 1024 Pq int The number of taskq entries that are pre-populated when the taskq is first created and are immediately available for use. . .It Sy zfs_zil_clean_taskq_nthr_pct Ns = Ns Sy 100 Ns % Pq int This controls the number of threads used by .Sy dp_zil_clean_taskq . The default value of .Sy 100% will create a maximum of one thread per cpu. . .It Sy zil_maxblocksize Ns = Ns Sy 131072 Ns B Po 128kB Pc Pq int This sets the maximum block size used by the ZIL. On very fragmented pools, lowering this .Pq typically to Sy 36kB can improve performance. . .It Sy zil_min_commit_timeout Ns = Ns Sy 5000 Pq u64 This sets the minimum delay in nanoseconds ZIL care to delay block commit, waiting for more records. If ZIL writes are too fast, kernel may not be able sleep for so short interval, increasing log latency above allowed by .Sy zfs_commit_timeout_pct . . .It Sy zil_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable the cache flush commands that are normally sent to disk by the ZIL after an LWB write has completed. Setting this will cause ZIL corruption on power loss if a volatile out-of-order write cache is enabled. . .It Sy zil_replay_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable intent logging replay. Can be disabled for recovery from corrupted ZIL. . .It Sy zil_slog_bulk Ns = Ns Sy 786432 Ns B Po 768kB Pc Pq ulong Limit SLOG write size per commit executed with synchronous priority. Any writes above that will be executed with lower (asynchronous) priority to limit potential SLOG device abuse by single active ZIL writer. . .It Sy zfs_embedded_slog_min_ms Ns = Ns Sy 64 Pq int Usually, one metaslab from each normal-class vdev is dedicated for use by the ZIL to log synchronous writes. However, if there are fewer than .Sy zfs_embedded_slog_min_ms metaslabs in the vdev, this functionality is disabled. This ensures that we don't set aside an unreasonable amount of space for the ZIL. . .It Sy zio_deadman_log_all Ns = Ns Sy 0 Ns | Ns 1 Pq int If non-zero, the zio deadman will produce debugging messages .Pq see Sy zfs_dbgmsg_enable for all zios, rather than only for leaf zios possessing a vdev. This is meant to be used by developers to gain diagnostic information for hang conditions which don't involve a mutex or other locking primitive: typically conditions in which a thread in the zio pipeline is looping indefinitely. . .It Sy zio_slow_io_ms Ns = Ns Sy 30000 Ns ms Po 30s Pc Pq int When an I/O operation takes more than this much time to complete, it's marked as slow. Each slow operation causes a delay zevent. Slow I/O counters can be seen with .Nm zpool Cm status Fl s . . .It Sy zio_dva_throttle_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Throttle block allocations in the I/O pipeline. This allows for dynamic allocation distribution when devices are imbalanced. When enabled, the maximum number of pending allocations per top-level vdev is limited by .Sy zfs_vdev_queue_depth_pct . . .It Sy zio_requeue_io_start_cut_in_line Ns = Ns Sy 0 Ns | Ns 1 Pq int Prioritize requeued I/O. . .It Sy zio_taskq_batch_pct Ns = Ns Sy 80 Ns % Pq uint Percentage of online CPUs which will run a worker thread for I/O. These workers are responsible for I/O work such as compression and checksum calculations. Fractional number of CPUs will be rounded down. .Pp The default value of .Sy 80% was chosen to avoid using all CPUs which can result in latency issues and inconsistent application performance, especially when slower compression and/or checksumming is enabled. . .It Sy zio_taskq_batch_tpq Ns = Ns Sy 0 Pq uint Number of worker threads per taskq. Lower values improve I/O ordering and CPU utilization, while higher reduces lock contention. .Pp If .Sy 0 , generate a system-dependent value close to 6 threads per taskq. . .It Sy zvol_inhibit_dev Ns = Ns Sy 0 Ns | Ns 1 Pq uint Do not create zvol device nodes. This may slightly improve startup time on systems with a very large number of zvols. . .It Sy zvol_major Ns = Ns Sy 230 Pq uint Major number for zvol block devices. . .It Sy zvol_max_discard_blocks Ns = Ns Sy 16384 Pq ulong Discard (TRIM) operations done on zvols will be done in batches of this many blocks, where block size is determined by the .Sy volblocksize property of a zvol. . .It Sy zvol_prefetch_bytes Ns = Ns Sy 131072 Ns B Po 128kB Pc Pq uint When adding a zvol to the system, prefetch this many bytes from the start and end of the volume. Prefetching these regions of the volume is desirable, because they are likely to be accessed immediately by .Xr blkid 8 or the kernel partitioner. . .It Sy zvol_request_sync Ns = Ns Sy 0 Ns | Ns 1 Pq uint When processing I/O requests for a zvol, submit them synchronously. This effectively limits the queue depth to .Em 1 for each I/O submitter. When unset, requests are handled asynchronously by a thread pool. The number of requests which can be handled concurrently is controlled by .Sy zvol_threads . . .It Sy zvol_threads Ns = Ns Sy 32 Pq uint Max number of threads which can handle zvol I/O requests concurrently. . .It Sy zvol_volmode Ns = Ns Sy 1 Pq uint Defines zvol block devices behaviour when .Sy volmode Ns = Ns Sy default : .Bl -tag -compact -offset 4n -width "a" .It Sy 1 .No equivalent to Sy full .It Sy 2 .No equivalent to Sy dev .It Sy 3 .No equivalent to Sy none .El .El . .Sh ZFS I/O SCHEDULER ZFS issues I/O operations to leaf vdevs to satisfy and complete I/O operations. The scheduler determines when and in what order those operations are issued. The scheduler divides operations into five I/O classes, prioritized in the following order: sync read, sync write, async read, async write, and scrub/resilver. Each queue defines the minimum and maximum number of concurrent operations that may be issued to the device. In addition, the device has an aggregate maximum, .Sy zfs_vdev_max_active . Note that the sum of the per-queue minima must not exceed the aggregate maximum. If the sum of the per-queue maxima exceeds the aggregate maximum, then the number of active operations may reach .Sy zfs_vdev_max_active , in which case no further operations will be issued, regardless of whether all per-queue minima have been met. .Pp For many physical devices, throughput increases with the number of concurrent operations, but latency typically suffers. Furthermore, physical devices typically have a limit at which more concurrent operations have no effect on throughput or can actually cause it to decrease. .Pp The scheduler selects the next operation to issue by first looking for an I/O class whose minimum has not been satisfied. Once all are satisfied and the aggregate maximum has not been hit, the scheduler looks for classes whose maximum has not been satisfied. Iteration through the I/O classes is done in the order specified above. No further operations are issued if the aggregate maximum number of concurrent operations has been hit, or if there are no operations queued for an I/O class that has not hit its maximum. Every time an I/O operation is queued or an operation completes, the scheduler looks for new operations to issue. .Pp In general, smaller .Sy max_active Ns s will lead to lower latency of synchronous operations. Larger .Sy max_active Ns s may lead to higher overall throughput, depending on underlying storage. .Pp The ratio of the queues' .Sy max_active Ns s determines the balance of performance between reads, writes, and scrubs. For example, increasing .Sy zfs_vdev_scrub_max_active will cause the scrub or resilver to complete more quickly, but reads and writes to have higher latency and lower throughput. .Pp All I/O classes have a fixed maximum number of outstanding operations, except for the async write class. Asynchronous writes represent the data that is committed to stable storage during the syncing stage for transaction groups. Transaction groups enter the syncing state periodically, so the number of queued async writes will quickly burst up and then bleed down to zero. Rather than servicing them as quickly as possible, the I/O scheduler changes the maximum number of active async write operations according to the amount of dirty data in the pool. Since both throughput and latency typically increase with the number of concurrent operations issued to physical devices, reducing the burstiness in the number of concurrent operations also stabilizes the response time of operations from other – and in particular synchronous – queues. In broad strokes, the I/O scheduler will issue more concurrent operations from the async write queue as there's more dirty data in the pool. . .Ss Async Writes The number of concurrent operations issued for the async write I/O class follows a piece-wise linear function defined by a few adjustable points: .Bd -literal | o---------| <-- \fBzfs_vdev_async_write_max_active\fP ^ | /^ | | | / | | active | / | | I/O | / | | count | / | | | / | | |-------o | | <-- \fBzfs_vdev_async_write_min_active\fP 0|_______^______|_________| 0% | | 100% of \fBzfs_dirty_data_max\fP | | | `-- \fBzfs_vdev_async_write_active_max_dirty_percent\fP `--------- \fBzfs_vdev_async_write_active_min_dirty_percent\fP .Ed .Pp Until the amount of dirty data exceeds a minimum percentage of the dirty data allowed in the pool, the I/O scheduler will limit the number of concurrent operations to the minimum. As that threshold is crossed, the number of concurrent operations issued increases linearly to the maximum at the specified maximum percentage of the dirty data allowed in the pool. .Pp Ideally, the amount of dirty data on a busy pool will stay in the sloped part of the function between .Sy zfs_vdev_async_write_active_min_dirty_percent and .Sy zfs_vdev_async_write_active_max_dirty_percent . If it exceeds the maximum percentage, this indicates that the rate of incoming data is greater than the rate that the backend storage can handle. In this case, we must further throttle incoming writes, as described in the next section. . .Sh ZFS TRANSACTION DELAY We delay transactions when we've determined that the backend storage isn't able to accommodate the rate of incoming writes. .Pp If there is already a transaction waiting, we delay relative to when that transaction will finish waiting. This way the calculated delay time is independent of the number of threads concurrently executing transactions. .Pp If we are the only waiter, wait relative to when the transaction started, rather than the current time. This credits the transaction for "time already served", e.g. reading indirect blocks. .Pp The minimum time for a transaction to take is calculated as .Dl min_time = min( Ns Sy zfs_delay_scale No * (dirty - min) / (max - dirty), 100ms) .Pp The delay has two degrees of freedom that can be adjusted via tunables. The percentage of dirty data at which we start to delay is defined by .Sy zfs_delay_min_dirty_percent . This should typically be at or above .Sy zfs_vdev_async_write_active_max_dirty_percent , so that we only start to delay after writing at full speed has failed to keep up with the incoming write rate. The scale of the curve is defined by .Sy zfs_delay_scale . Roughly speaking, this variable determines the amount of delay at the midpoint of the curve. .Bd -literal delay 10ms +-------------------------------------------------------------*+ | *| 9ms + *+ | *| 8ms + *+ | * | 7ms + * + | * | 6ms + * + | * | 5ms + * + | * | 4ms + * + | * | 3ms + * + | * | 2ms + (midpoint) * + | | ** | 1ms + v *** + | \fBzfs_delay_scale\fP ----------> ******** | 0 +-------------------------------------*********----------------+ 0% <- \fBzfs_dirty_data_max\fP -> 100% .Ed .Pp Note, that since the delay is added to the outstanding time remaining on the most recent transaction it's effectively the inverse of IOPS. Here, the midpoint of .Em 500us translates to .Em 2000 IOPS . The shape of the curve was chosen such that small changes in the amount of accumulated dirty data in the first three quarters of the curve yield relatively small differences in the amount of delay. .Pp The effects can be easier to understand when the amount of delay is represented on a logarithmic scale: .Bd -literal delay 100ms +-------------------------------------------------------------++ + + | | + *+ 10ms + *+ + ** + | (midpoint) ** | + | ** + 1ms + v **** + + \fBzfs_delay_scale\fP ----------> ***** + | **** | + **** + 100us + ** + + * + | * | + * + 10us + * + + + | | + + +--------------------------------------------------------------+ 0% <- \fBzfs_dirty_data_max\fP -> 100% .Ed .Pp Note here that only as the amount of dirty data approaches its limit does the delay start to increase rapidly. The goal of a properly tuned system should be to keep the amount of dirty data out of that range by first ensuring that the appropriate limits are set for the I/O scheduler to reach optimal throughput on the back-end storage, and then by changing the value of .Sy zfs_delay_scale to increase the steepness of the curve. diff --git a/module/zfs/dsl_scan.c b/module/zfs/dsl_scan.c index f3c639b0d04e..49ea1d47cf71 100644 --- a/module/zfs/dsl_scan.c +++ b/module/zfs/dsl_scan.c @@ -1,4490 +1,4518 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2021 by Delphix. All rights reserved. * Copyright 2016 Gary Mills * Copyright (c) 2017, 2019, Datto Inc. All rights reserved. * Copyright (c) 2015, Nexenta Systems, Inc. All rights reserved. * Copyright 2019 Joyent, Inc. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef _KERNEL #include #endif /* * Grand theory statement on scan queue sorting * * Scanning is implemented by recursively traversing all indirection levels * in an object and reading all blocks referenced from said objects. This * results in us approximately traversing the object from lowest logical * offset to the highest. For best performance, we would want the logical * blocks to be physically contiguous. However, this is frequently not the * case with pools given the allocation patterns of copy-on-write filesystems. * So instead, we put the I/Os into a reordering queue and issue them in a * way that will most benefit physical disks (LBA-order). * * Queue management: * * Ideally, we would want to scan all metadata and queue up all block I/O * prior to starting to issue it, because that allows us to do an optimal * sorting job. This can however consume large amounts of memory. Therefore * we continuously monitor the size of the queues and constrain them to 5% * (zfs_scan_mem_lim_fact) of physmem. If the queues grow larger than this * limit, we clear out a few of the largest extents at the head of the queues * to make room for more scanning. Hopefully, these extents will be fairly * large and contiguous, allowing us to approach sequential I/O throughput * even without a fully sorted tree. * * Metadata scanning takes place in dsl_scan_visit(), which is called from * dsl_scan_sync() every spa_sync(). If we have either fully scanned all * metadata on the pool, or we need to make room in memory because our * queues are too large, dsl_scan_visit() is postponed and * scan_io_queues_run() is called from dsl_scan_sync() instead. This implies * that metadata scanning and queued I/O issuing are mutually exclusive. This * allows us to provide maximum sequential I/O throughput for the majority of * I/O's issued since sequential I/O performance is significantly negatively * impacted if it is interleaved with random I/O. * * Implementation Notes * * One side effect of the queued scanning algorithm is that the scanning code * needs to be notified whenever a block is freed. This is needed to allow * the scanning code to remove these I/Os from the issuing queue. Additionally, * we do not attempt to queue gang blocks to be issued sequentially since this * is very hard to do and would have an extremely limited performance benefit. * Instead, we simply issue gang I/Os as soon as we find them using the legacy * algorithm. * * Backwards compatibility * * This new algorithm is backwards compatible with the legacy on-disk data * structures (and therefore does not require a new feature flag). * Periodically during scanning (see zfs_scan_checkpoint_intval), the scan * will stop scanning metadata (in logical order) and wait for all outstanding * sorted I/O to complete. Once this is done, we write out a checkpoint * bookmark, indicating that we have scanned everything logically before it. * If the pool is imported on a machine without the new sorting algorithm, * the scan simply resumes from the last checkpoint using the legacy algorithm. */ typedef int (scan_cb_t)(dsl_pool_t *, const blkptr_t *, const zbookmark_phys_t *); static scan_cb_t dsl_scan_scrub_cb; static int scan_ds_queue_compare(const void *a, const void *b); static int scan_prefetch_queue_compare(const void *a, const void *b); static void scan_ds_queue_clear(dsl_scan_t *scn); static void scan_ds_prefetch_queue_clear(dsl_scan_t *scn); static boolean_t scan_ds_queue_contains(dsl_scan_t *scn, uint64_t dsobj, uint64_t *txg); static void scan_ds_queue_insert(dsl_scan_t *scn, uint64_t dsobj, uint64_t txg); static void scan_ds_queue_remove(dsl_scan_t *scn, uint64_t dsobj); static void scan_ds_queue_sync(dsl_scan_t *scn, dmu_tx_t *tx); static uint64_t dsl_scan_count_data_disks(vdev_t *vd); extern int zfs_vdev_async_write_active_min_dirty_percent; static int zfs_scan_blkstats = 0; +/* + * 'zpool status' uses bytes processed per pass to report throughput and + * estimate time remaining. We define a pass to start when the scanning + * phase completes for a sequential resilver. Optionally, this value + * may be used to reset the pass statistics every N txgs to provide an + * estimated completion time based on currently observed performance. + */ +static uint_t zfs_scan_report_txgs = 0; + /* * By default zfs will check to ensure it is not over the hard memory * limit before each txg. If finer-grained control of this is needed * this value can be set to 1 to enable checking before scanning each * block. */ int zfs_scan_strict_mem_lim = B_FALSE; /* * Maximum number of parallelly executed bytes per leaf vdev. We attempt * to strike a balance here between keeping the vdev queues full of I/Os * at all times and not overflowing the queues to cause long latency, * which would cause long txg sync times. No matter what, we will not * overload the drives with I/O, since that is protected by * zfs_vdev_scrub_max_active. */ unsigned long zfs_scan_vdev_limit = 4 << 20; int zfs_scan_issue_strategy = 0; int zfs_scan_legacy = B_FALSE; /* don't queue & sort zios, go direct */ unsigned long zfs_scan_max_ext_gap = 2 << 20; /* in bytes */ /* * fill_weight is non-tunable at runtime, so we copy it at module init from * zfs_scan_fill_weight. Runtime adjustments to zfs_scan_fill_weight would * break queue sorting. */ int zfs_scan_fill_weight = 3; static uint64_t fill_weight; /* See dsl_scan_should_clear() for details on the memory limit tunables */ uint64_t zfs_scan_mem_lim_min = 16 << 20; /* bytes */ uint64_t zfs_scan_mem_lim_soft_max = 128 << 20; /* bytes */ int zfs_scan_mem_lim_fact = 20; /* fraction of physmem */ int zfs_scan_mem_lim_soft_fact = 20; /* fraction of mem lim above */ int zfs_scrub_min_time_ms = 1000; /* min millisecs to scrub per txg */ int zfs_obsolete_min_time_ms = 500; /* min millisecs to obsolete per txg */ int zfs_free_min_time_ms = 1000; /* min millisecs to free per txg */ int zfs_resilver_min_time_ms = 3000; /* min millisecs to resilver per txg */ int zfs_scan_checkpoint_intval = 7200; /* in seconds */ int zfs_scan_suspend_progress = 0; /* set to prevent scans from progressing */ int zfs_no_scrub_io = B_FALSE; /* set to disable scrub i/o */ int zfs_no_scrub_prefetch = B_FALSE; /* set to disable scrub prefetch */ enum ddt_class zfs_scrub_ddt_class_max = DDT_CLASS_DUPLICATE; /* max number of blocks to free in a single TXG */ unsigned long zfs_async_block_max_blocks = ULONG_MAX; /* max number of dedup blocks to free in a single TXG */ unsigned long zfs_max_async_dedup_frees = 100000; int zfs_resilver_disable_defer = 0; /* set to disable resilver deferring */ /* * We wait a few txgs after importing a pool to begin scanning so that * the import / mounting code isn't held up by scrub / resilver IO. * Unfortunately, it is a bit difficult to determine exactly how long * this will take since userspace will trigger fs mounts asynchronously * and the kernel will create zvol minors asynchronously. As a result, * the value provided here is a bit arbitrary, but represents a * reasonable estimate of how many txgs it will take to finish fully * importing a pool */ #define SCAN_IMPORT_WAIT_TXGS 5 #define DSL_SCAN_IS_SCRUB_RESILVER(scn) \ ((scn)->scn_phys.scn_func == POOL_SCAN_SCRUB || \ (scn)->scn_phys.scn_func == POOL_SCAN_RESILVER) /* * Enable/disable the processing of the free_bpobj object. */ int zfs_free_bpobj_enabled = 1; /* the order has to match pool_scan_type */ static scan_cb_t *scan_funcs[POOL_SCAN_FUNCS] = { NULL, dsl_scan_scrub_cb, /* POOL_SCAN_SCRUB */ dsl_scan_scrub_cb, /* POOL_SCAN_RESILVER */ }; /* In core node for the scn->scn_queue. Represents a dataset to be scanned */ typedef struct { uint64_t sds_dsobj; uint64_t sds_txg; avl_node_t sds_node; } scan_ds_t; /* * This controls what conditions are placed on dsl_scan_sync_state(): * SYNC_OPTIONAL) write out scn_phys iff scn_queues_pending == 0 * SYNC_MANDATORY) write out scn_phys always. scn_queues_pending must be 0. * SYNC_CACHED) if scn_queues_pending == 0, write out scn_phys. Otherwise * write out the scn_phys_cached version. * See dsl_scan_sync_state for details. */ typedef enum { SYNC_OPTIONAL, SYNC_MANDATORY, SYNC_CACHED } state_sync_type_t; /* * This struct represents the minimum information needed to reconstruct a * zio for sequential scanning. This is useful because many of these will * accumulate in the sequential IO queues before being issued, so saving * memory matters here. */ typedef struct scan_io { /* fields from blkptr_t */ uint64_t sio_blk_prop; uint64_t sio_phys_birth; uint64_t sio_birth; zio_cksum_t sio_cksum; uint32_t sio_nr_dvas; /* fields from zio_t */ uint32_t sio_flags; zbookmark_phys_t sio_zb; /* members for queue sorting */ union { avl_node_t sio_addr_node; /* link into issuing queue */ list_node_t sio_list_node; /* link for issuing to disk */ } sio_nodes; /* * There may be up to SPA_DVAS_PER_BP DVAs here from the bp, * depending on how many were in the original bp. Only the * first DVA is really used for sorting and issuing purposes. * The other DVAs (if provided) simply exist so that the zio * layer can find additional copies to repair from in the * event of an error. This array must go at the end of the * struct to allow this for the variable number of elements. */ dva_t sio_dva[0]; } scan_io_t; #define SIO_SET_OFFSET(sio, x) DVA_SET_OFFSET(&(sio)->sio_dva[0], x) #define SIO_SET_ASIZE(sio, x) DVA_SET_ASIZE(&(sio)->sio_dva[0], x) #define SIO_GET_OFFSET(sio) DVA_GET_OFFSET(&(sio)->sio_dva[0]) #define SIO_GET_ASIZE(sio) DVA_GET_ASIZE(&(sio)->sio_dva[0]) #define SIO_GET_END_OFFSET(sio) \ (SIO_GET_OFFSET(sio) + SIO_GET_ASIZE(sio)) #define SIO_GET_MUSED(sio) \ (sizeof (scan_io_t) + ((sio)->sio_nr_dvas * sizeof (dva_t))) struct dsl_scan_io_queue { dsl_scan_t *q_scn; /* associated dsl_scan_t */ vdev_t *q_vd; /* top-level vdev that this queue represents */ zio_t *q_zio; /* scn_zio_root child for waiting on IO */ /* trees used for sorting I/Os and extents of I/Os */ range_tree_t *q_exts_by_addr; zfs_btree_t q_exts_by_size; avl_tree_t q_sios_by_addr; uint64_t q_sio_memused; uint64_t q_last_ext_addr; /* members for zio rate limiting */ uint64_t q_maxinflight_bytes; uint64_t q_inflight_bytes; kcondvar_t q_zio_cv; /* used under vd->vdev_scan_io_queue_lock */ /* per txg statistics */ uint64_t q_total_seg_size_this_txg; uint64_t q_segs_this_txg; uint64_t q_total_zio_size_this_txg; uint64_t q_zios_this_txg; }; /* private data for dsl_scan_prefetch_cb() */ typedef struct scan_prefetch_ctx { zfs_refcount_t spc_refcnt; /* refcount for memory management */ dsl_scan_t *spc_scn; /* dsl_scan_t for the pool */ boolean_t spc_root; /* is this prefetch for an objset? */ uint8_t spc_indblkshift; /* dn_indblkshift of current dnode */ uint16_t spc_datablkszsec; /* dn_idatablkszsec of current dnode */ } scan_prefetch_ctx_t; /* private data for dsl_scan_prefetch() */ typedef struct scan_prefetch_issue_ctx { avl_node_t spic_avl_node; /* link into scn->scn_prefetch_queue */ scan_prefetch_ctx_t *spic_spc; /* spc for the callback */ blkptr_t spic_bp; /* bp to prefetch */ zbookmark_phys_t spic_zb; /* bookmark to prefetch */ } scan_prefetch_issue_ctx_t; static void scan_exec_io(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags, const zbookmark_phys_t *zb, dsl_scan_io_queue_t *queue); static void scan_io_queue_insert_impl(dsl_scan_io_queue_t *queue, scan_io_t *sio); static dsl_scan_io_queue_t *scan_io_queue_create(vdev_t *vd); static void scan_io_queues_destroy(dsl_scan_t *scn); static kmem_cache_t *sio_cache[SPA_DVAS_PER_BP]; /* sio->sio_nr_dvas must be set so we know which cache to free from */ static void sio_free(scan_io_t *sio) { ASSERT3U(sio->sio_nr_dvas, >, 0); ASSERT3U(sio->sio_nr_dvas, <=, SPA_DVAS_PER_BP); kmem_cache_free(sio_cache[sio->sio_nr_dvas - 1], sio); } /* It is up to the caller to set sio->sio_nr_dvas for freeing */ static scan_io_t * sio_alloc(unsigned short nr_dvas) { ASSERT3U(nr_dvas, >, 0); ASSERT3U(nr_dvas, <=, SPA_DVAS_PER_BP); return (kmem_cache_alloc(sio_cache[nr_dvas - 1], KM_SLEEP)); } void scan_init(void) { /* * This is used in ext_size_compare() to weight segments * based on how sparse they are. This cannot be changed * mid-scan and the tree comparison functions don't currently * have a mechanism for passing additional context to the * compare functions. Thus we store this value globally and * we only allow it to be set at module initialization time */ fill_weight = zfs_scan_fill_weight; for (int i = 0; i < SPA_DVAS_PER_BP; i++) { char name[36]; (void) snprintf(name, sizeof (name), "sio_cache_%d", i); sio_cache[i] = kmem_cache_create(name, (sizeof (scan_io_t) + ((i + 1) * sizeof (dva_t))), 0, NULL, NULL, NULL, NULL, NULL, 0); } } void scan_fini(void) { for (int i = 0; i < SPA_DVAS_PER_BP; i++) { kmem_cache_destroy(sio_cache[i]); } } static inline boolean_t dsl_scan_is_running(const dsl_scan_t *scn) { return (scn->scn_phys.scn_state == DSS_SCANNING); } boolean_t dsl_scan_resilvering(dsl_pool_t *dp) { return (dsl_scan_is_running(dp->dp_scan) && dp->dp_scan->scn_phys.scn_func == POOL_SCAN_RESILVER); } static inline void sio2bp(const scan_io_t *sio, blkptr_t *bp) { bzero(bp, sizeof (*bp)); bp->blk_prop = sio->sio_blk_prop; bp->blk_phys_birth = sio->sio_phys_birth; bp->blk_birth = sio->sio_birth; bp->blk_fill = 1; /* we always only work with data pointers */ bp->blk_cksum = sio->sio_cksum; ASSERT3U(sio->sio_nr_dvas, >, 0); ASSERT3U(sio->sio_nr_dvas, <=, SPA_DVAS_PER_BP); bcopy(sio->sio_dva, bp->blk_dva, sio->sio_nr_dvas * sizeof (dva_t)); } static inline void bp2sio(const blkptr_t *bp, scan_io_t *sio, int dva_i) { sio->sio_blk_prop = bp->blk_prop; sio->sio_phys_birth = bp->blk_phys_birth; sio->sio_birth = bp->blk_birth; sio->sio_cksum = bp->blk_cksum; sio->sio_nr_dvas = BP_GET_NDVAS(bp); /* * Copy the DVAs to the sio. We need all copies of the block so * that the self healing code can use the alternate copies if the * first is corrupted. We want the DVA at index dva_i to be first * in the sio since this is the primary one that we want to issue. */ for (int i = 0, j = dva_i; i < sio->sio_nr_dvas; i++, j++) { sio->sio_dva[i] = bp->blk_dva[j % sio->sio_nr_dvas]; } } int dsl_scan_init(dsl_pool_t *dp, uint64_t txg) { int err; dsl_scan_t *scn; spa_t *spa = dp->dp_spa; uint64_t f; scn = dp->dp_scan = kmem_zalloc(sizeof (dsl_scan_t), KM_SLEEP); scn->scn_dp = dp; /* * It's possible that we're resuming a scan after a reboot so * make sure that the scan_async_destroying flag is initialized * appropriately. */ ASSERT(!scn->scn_async_destroying); scn->scn_async_destroying = spa_feature_is_active(dp->dp_spa, SPA_FEATURE_ASYNC_DESTROY); /* * Calculate the max number of in-flight bytes for pool-wide * scanning operations (minimum 1MB). Limits for the issuing * phase are done per top-level vdev and are handled separately. */ scn->scn_maxinflight_bytes = MAX(zfs_scan_vdev_limit * dsl_scan_count_data_disks(spa->spa_root_vdev), 1ULL << 20); avl_create(&scn->scn_queue, scan_ds_queue_compare, sizeof (scan_ds_t), offsetof(scan_ds_t, sds_node)); avl_create(&scn->scn_prefetch_queue, scan_prefetch_queue_compare, sizeof (scan_prefetch_issue_ctx_t), offsetof(scan_prefetch_issue_ctx_t, spic_avl_node)); err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, "scrub_func", sizeof (uint64_t), 1, &f); if (err == 0) { /* * There was an old-style scrub in progress. Restart a * new-style scrub from the beginning. */ scn->scn_restart_txg = txg; zfs_dbgmsg("old-style scrub was in progress; " "restarting new-style scrub in txg %llu", (longlong_t)scn->scn_restart_txg); /* * Load the queue obj from the old location so that it * can be freed by dsl_scan_done(). */ (void) zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, "scrub_queue", sizeof (uint64_t), 1, &scn->scn_phys.scn_queue_obj); } else { err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS, &scn->scn_phys); /* * Detect if the pool contains the signature of #2094. If it * does properly update the scn->scn_phys structure and notify * the administrator by setting an errata for the pool. */ if (err == EOVERFLOW) { uint64_t zaptmp[SCAN_PHYS_NUMINTS + 1]; VERIFY3S(SCAN_PHYS_NUMINTS, ==, 24); VERIFY3S(offsetof(dsl_scan_phys_t, scn_flags), ==, (23 * sizeof (uint64_t))); err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS + 1, &zaptmp); if (err == 0) { uint64_t overflow = zaptmp[SCAN_PHYS_NUMINTS]; if (overflow & ~DSL_SCAN_FLAGS_MASK || scn->scn_async_destroying) { spa->spa_errata = ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY; return (EOVERFLOW); } bcopy(zaptmp, &scn->scn_phys, SCAN_PHYS_NUMINTS * sizeof (uint64_t)); scn->scn_phys.scn_flags = overflow; /* Required scrub already in progress. */ if (scn->scn_phys.scn_state == DSS_FINISHED || scn->scn_phys.scn_state == DSS_CANCELED) spa->spa_errata = ZPOOL_ERRATA_ZOL_2094_SCRUB; } } if (err == ENOENT) return (0); else if (err) return (err); /* * We might be restarting after a reboot, so jump the issued * counter to how far we've scanned. We know we're consistent * up to here. */ scn->scn_issued_before_pass = scn->scn_phys.scn_examined; if (dsl_scan_is_running(scn) && spa_prev_software_version(dp->dp_spa) < SPA_VERSION_SCAN) { /* * A new-type scrub was in progress on an old * pool, and the pool was accessed by old * software. Restart from the beginning, since * the old software may have changed the pool in * the meantime. */ scn->scn_restart_txg = txg; zfs_dbgmsg("new-style scrub was modified " "by old software; restarting in txg %llu", (longlong_t)scn->scn_restart_txg); } else if (dsl_scan_resilvering(dp)) { /* * If a resilver is in progress and there are already * errors, restart it instead of finishing this scan and * then restarting it. If there haven't been any errors * then remember that the incore DTL is valid. */ if (scn->scn_phys.scn_errors > 0) { scn->scn_restart_txg = txg; zfs_dbgmsg("resilver can't excise DTL_MISSING " "when finished; restarting in txg %llu", (u_longlong_t)scn->scn_restart_txg); } else { /* it's safe to excise DTL when finished */ spa->spa_scrub_started = B_TRUE; } } } bcopy(&scn->scn_phys, &scn->scn_phys_cached, sizeof (scn->scn_phys)); /* reload the queue into the in-core state */ if (scn->scn_phys.scn_queue_obj != 0) { zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, dp->dp_meta_objset, scn->scn_phys.scn_queue_obj); zap_cursor_retrieve(&zc, &za) == 0; (void) zap_cursor_advance(&zc)) { scan_ds_queue_insert(scn, zfs_strtonum(za.za_name, NULL), za.za_first_integer); } zap_cursor_fini(&zc); } spa_scan_stat_init(spa); + vdev_scan_stat_init(spa->spa_root_vdev); + return (0); } void dsl_scan_fini(dsl_pool_t *dp) { if (dp->dp_scan != NULL) { dsl_scan_t *scn = dp->dp_scan; if (scn->scn_taskq != NULL) taskq_destroy(scn->scn_taskq); scan_ds_queue_clear(scn); avl_destroy(&scn->scn_queue); scan_ds_prefetch_queue_clear(scn); avl_destroy(&scn->scn_prefetch_queue); kmem_free(dp->dp_scan, sizeof (dsl_scan_t)); dp->dp_scan = NULL; } } static boolean_t dsl_scan_restarting(dsl_scan_t *scn, dmu_tx_t *tx) { return (scn->scn_restart_txg != 0 && scn->scn_restart_txg <= tx->tx_txg); } boolean_t dsl_scan_resilver_scheduled(dsl_pool_t *dp) { return ((dp->dp_scan && dp->dp_scan->scn_restart_txg != 0) || (spa_async_tasks(dp->dp_spa) & SPA_ASYNC_RESILVER)); } boolean_t dsl_scan_scrubbing(const dsl_pool_t *dp) { dsl_scan_phys_t *scn_phys = &dp->dp_scan->scn_phys; return (scn_phys->scn_state == DSS_SCANNING && scn_phys->scn_func == POOL_SCAN_SCRUB); } boolean_t dsl_scan_is_paused_scrub(const dsl_scan_t *scn) { return (dsl_scan_scrubbing(scn->scn_dp) && scn->scn_phys.scn_flags & DSF_SCRUB_PAUSED); } /* * Writes out a persistent dsl_scan_phys_t record to the pool directory. * Because we can be running in the block sorting algorithm, we do not always * want to write out the record, only when it is "safe" to do so. This safety * condition is achieved by making sure that the sorting queues are empty * (scn_queues_pending == 0). When this condition is not true, the sync'd state * is inconsistent with how much actual scanning progress has been made. The * kind of sync to be performed is specified by the sync_type argument. If the * sync is optional, we only sync if the queues are empty. If the sync is * mandatory, we do a hard ASSERT to make sure that the queues are empty. The * third possible state is a "cached" sync. This is done in response to: * 1) The dataset that was in the last sync'd dsl_scan_phys_t having been * destroyed, so we wouldn't be able to restart scanning from it. * 2) The snapshot that was in the last sync'd dsl_scan_phys_t having been * superseded by a newer snapshot. * 3) The dataset that was in the last sync'd dsl_scan_phys_t having been * swapped with its clone. * In all cases, a cached sync simply rewrites the last record we've written, * just slightly modified. For the modifications that are performed to the * last written dsl_scan_phys_t, see dsl_scan_ds_destroyed, * dsl_scan_ds_snapshotted and dsl_scan_ds_clone_swapped. */ static void dsl_scan_sync_state(dsl_scan_t *scn, dmu_tx_t *tx, state_sync_type_t sync_type) { int i; spa_t *spa = scn->scn_dp->dp_spa; ASSERT(sync_type != SYNC_MANDATORY || scn->scn_queues_pending == 0); if (scn->scn_queues_pending == 0) { for (i = 0; i < spa->spa_root_vdev->vdev_children; i++) { vdev_t *vd = spa->spa_root_vdev->vdev_child[i]; dsl_scan_io_queue_t *q = vd->vdev_scan_io_queue; if (q == NULL) continue; mutex_enter(&vd->vdev_scan_io_queue_lock); ASSERT3P(avl_first(&q->q_sios_by_addr), ==, NULL); ASSERT3P(zfs_btree_first(&q->q_exts_by_size, NULL), ==, NULL); ASSERT3P(range_tree_first(q->q_exts_by_addr), ==, NULL); mutex_exit(&vd->vdev_scan_io_queue_lock); } if (scn->scn_phys.scn_queue_obj != 0) scan_ds_queue_sync(scn, tx); VERIFY0(zap_update(scn->scn_dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS, &scn->scn_phys, tx)); bcopy(&scn->scn_phys, &scn->scn_phys_cached, sizeof (scn->scn_phys)); if (scn->scn_checkpointing) zfs_dbgmsg("finish scan checkpoint"); scn->scn_checkpointing = B_FALSE; scn->scn_last_checkpoint = ddi_get_lbolt(); } else if (sync_type == SYNC_CACHED) { VERIFY0(zap_update(scn->scn_dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS, &scn->scn_phys_cached, tx)); } } int dsl_scan_setup_check(void *arg, dmu_tx_t *tx) { (void) arg; dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev; if (dsl_scan_is_running(scn) || vdev_rebuild_active(rvd)) return (SET_ERROR(EBUSY)); return (0); } void dsl_scan_setup_sync(void *arg, dmu_tx_t *tx) { dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; pool_scan_func_t *funcp = arg; dmu_object_type_t ot = 0; dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; ASSERT(!dsl_scan_is_running(scn)); ASSERT(*funcp > POOL_SCAN_NONE && *funcp < POOL_SCAN_FUNCS); bzero(&scn->scn_phys, sizeof (scn->scn_phys)); scn->scn_phys.scn_func = *funcp; scn->scn_phys.scn_state = DSS_SCANNING; scn->scn_phys.scn_min_txg = 0; scn->scn_phys.scn_max_txg = tx->tx_txg; scn->scn_phys.scn_ddt_class_max = DDT_CLASSES - 1; /* the entire DDT */ scn->scn_phys.scn_start_time = gethrestime_sec(); scn->scn_phys.scn_errors = 0; scn->scn_phys.scn_to_examine = spa->spa_root_vdev->vdev_stat.vs_alloc; scn->scn_issued_before_pass = 0; scn->scn_restart_txg = 0; scn->scn_done_txg = 0; scn->scn_last_checkpoint = 0; scn->scn_checkpointing = B_FALSE; spa_scan_stat_init(spa); + vdev_scan_stat_init(spa->spa_root_vdev); if (DSL_SCAN_IS_SCRUB_RESILVER(scn)) { scn->scn_phys.scn_ddt_class_max = zfs_scrub_ddt_class_max; /* rewrite all disk labels */ vdev_config_dirty(spa->spa_root_vdev); if (vdev_resilver_needed(spa->spa_root_vdev, &scn->scn_phys.scn_min_txg, &scn->scn_phys.scn_max_txg)) { nvlist_t *aux = fnvlist_alloc(); fnvlist_add_string(aux, ZFS_EV_RESILVER_TYPE, "healing"); spa_event_notify(spa, NULL, aux, ESC_ZFS_RESILVER_START); nvlist_free(aux); } else { spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_START); } spa->spa_scrub_started = B_TRUE; /* * If this is an incremental scrub, limit the DDT scrub phase * to just the auto-ditto class (for correctness); the rest * of the scrub should go faster using top-down pruning. */ if (scn->scn_phys.scn_min_txg > TXG_INITIAL) scn->scn_phys.scn_ddt_class_max = DDT_CLASS_DITTO; /* * When starting a resilver clear any existing rebuild state. * This is required to prevent stale rebuild status from * being reported when a rebuild is run, then a resilver and * finally a scrub. In which case only the scrub status * should be reported by 'zpool status'. */ if (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) { vdev_t *rvd = spa->spa_root_vdev; for (uint64_t i = 0; i < rvd->vdev_children; i++) { vdev_t *vd = rvd->vdev_child[i]; vdev_rebuild_clear_sync( (void *)(uintptr_t)vd->vdev_id, tx); } } } /* back to the generic stuff */ if (zfs_scan_blkstats) { if (dp->dp_blkstats == NULL) { dp->dp_blkstats = vmem_alloc(sizeof (zfs_all_blkstats_t), KM_SLEEP); } memset(&dp->dp_blkstats->zab_type, 0, sizeof (dp->dp_blkstats->zab_type)); } else { if (dp->dp_blkstats) { vmem_free(dp->dp_blkstats, sizeof (zfs_all_blkstats_t)); dp->dp_blkstats = NULL; } } if (spa_version(spa) < SPA_VERSION_DSL_SCRUB) ot = DMU_OT_ZAP_OTHER; scn->scn_phys.scn_queue_obj = zap_create(dp->dp_meta_objset, ot ? ot : DMU_OT_SCAN_QUEUE, DMU_OT_NONE, 0, tx); bcopy(&scn->scn_phys, &scn->scn_phys_cached, sizeof (scn->scn_phys)); dsl_scan_sync_state(scn, tx, SYNC_MANDATORY); spa_history_log_internal(spa, "scan setup", tx, "func=%u mintxg=%llu maxtxg=%llu", *funcp, (u_longlong_t)scn->scn_phys.scn_min_txg, (u_longlong_t)scn->scn_phys.scn_max_txg); } /* * Called by the ZFS_IOC_POOL_SCAN ioctl to start a scrub or resilver. * Can also be called to resume a paused scrub. */ int dsl_scan(dsl_pool_t *dp, pool_scan_func_t func) { spa_t *spa = dp->dp_spa; dsl_scan_t *scn = dp->dp_scan; /* * Purge all vdev caches and probe all devices. We do this here * rather than in sync context because this requires a writer lock * on the spa_config lock, which we can't do from sync context. The * spa_scrub_reopen flag indicates that vdev_open() should not * attempt to start another scrub. */ spa_vdev_state_enter(spa, SCL_NONE); spa->spa_scrub_reopen = B_TRUE; vdev_reopen(spa->spa_root_vdev); spa->spa_scrub_reopen = B_FALSE; (void) spa_vdev_state_exit(spa, NULL, 0); if (func == POOL_SCAN_RESILVER) { dsl_scan_restart_resilver(spa->spa_dsl_pool, 0); return (0); } if (func == POOL_SCAN_SCRUB && dsl_scan_is_paused_scrub(scn)) { /* got scrub start cmd, resume paused scrub */ int err = dsl_scrub_set_pause_resume(scn->scn_dp, POOL_SCRUB_NORMAL); if (err == 0) { spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_RESUME); return (SET_ERROR(ECANCELED)); } return (SET_ERROR(err)); } return (dsl_sync_task(spa_name(spa), dsl_scan_setup_check, dsl_scan_setup_sync, &func, 0, ZFS_SPACE_CHECK_EXTRA_RESERVED)); } static void dsl_scan_done(dsl_scan_t *scn, boolean_t complete, dmu_tx_t *tx) { static const char *old_names[] = { "scrub_bookmark", "scrub_ddt_bookmark", "scrub_ddt_class_max", "scrub_queue", "scrub_min_txg", "scrub_max_txg", "scrub_func", "scrub_errors", NULL }; dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; int i; /* Remove any remnants of an old-style scrub. */ for (i = 0; old_names[i]; i++) { (void) zap_remove(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, old_names[i], tx); } if (scn->scn_phys.scn_queue_obj != 0) { VERIFY0(dmu_object_free(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, tx)); scn->scn_phys.scn_queue_obj = 0; } scan_ds_queue_clear(scn); scan_ds_prefetch_queue_clear(scn); scn->scn_phys.scn_flags &= ~DSF_SCRUB_PAUSED; /* * If we were "restarted" from a stopped state, don't bother * with anything else. */ if (!dsl_scan_is_running(scn)) { ASSERT(!scn->scn_is_sorted); return; } if (scn->scn_is_sorted) { scan_io_queues_destroy(scn); scn->scn_is_sorted = B_FALSE; if (scn->scn_taskq != NULL) { taskq_destroy(scn->scn_taskq); scn->scn_taskq = NULL; } } scn->scn_phys.scn_state = complete ? DSS_FINISHED : DSS_CANCELED; spa_notify_waiters(spa); if (dsl_scan_restarting(scn, tx)) spa_history_log_internal(spa, "scan aborted, restarting", tx, "errors=%llu", (u_longlong_t)spa_get_errlog_size(spa)); else if (!complete) spa_history_log_internal(spa, "scan cancelled", tx, "errors=%llu", (u_longlong_t)spa_get_errlog_size(spa)); else spa_history_log_internal(spa, "scan done", tx, "errors=%llu", (u_longlong_t)spa_get_errlog_size(spa)); if (DSL_SCAN_IS_SCRUB_RESILVER(scn)) { spa->spa_scrub_active = B_FALSE; /* * If the scrub/resilver completed, update all DTLs to * reflect this. Whether it succeeded or not, vacate * all temporary scrub DTLs. * * As the scrub does not currently support traversing * data that have been freed but are part of a checkpoint, * we don't mark the scrub as done in the DTLs as faults * may still exist in those vdevs. */ if (complete && !spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) { vdev_dtl_reassess(spa->spa_root_vdev, tx->tx_txg, scn->scn_phys.scn_max_txg, B_TRUE, B_FALSE); if (scn->scn_phys.scn_min_txg) { nvlist_t *aux = fnvlist_alloc(); fnvlist_add_string(aux, ZFS_EV_RESILVER_TYPE, "healing"); spa_event_notify(spa, NULL, aux, ESC_ZFS_RESILVER_FINISH); nvlist_free(aux); } else { spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_FINISH); } } else { vdev_dtl_reassess(spa->spa_root_vdev, tx->tx_txg, 0, B_TRUE, B_FALSE); } spa_errlog_rotate(spa); /* * Don't clear flag until after vdev_dtl_reassess to ensure that * DTL_MISSING will get updated when possible. */ spa->spa_scrub_started = B_FALSE; /* * We may have finished replacing a device. * Let the async thread assess this and handle the detach. */ spa_async_request(spa, SPA_ASYNC_RESILVER_DONE); /* * Clear any resilver_deferred flags in the config. * If there are drives that need resilvering, kick * off an asynchronous request to start resilver. * vdev_clear_resilver_deferred() may update the config * before the resilver can restart. In the event of * a crash during this period, the spa loading code * will find the drives that need to be resilvered * and start the resilver then. */ if (spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER) && vdev_clear_resilver_deferred(spa->spa_root_vdev, tx)) { spa_history_log_internal(spa, "starting deferred resilver", tx, "errors=%llu", (u_longlong_t)spa_get_errlog_size(spa)); spa_async_request(spa, SPA_ASYNC_RESILVER); } /* Clear recent error events (i.e. duplicate events tracking) */ if (complete) zfs_ereport_clear(spa, NULL); } scn->scn_phys.scn_end_time = gethrestime_sec(); if (spa->spa_errata == ZPOOL_ERRATA_ZOL_2094_SCRUB) spa->spa_errata = 0; ASSERT(!dsl_scan_is_running(scn)); } static int dsl_scan_cancel_check(void *arg, dmu_tx_t *tx) { (void) arg; dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; if (!dsl_scan_is_running(scn)) return (SET_ERROR(ENOENT)); return (0); } static void dsl_scan_cancel_sync(void *arg, dmu_tx_t *tx) { (void) arg; dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; dsl_scan_done(scn, B_FALSE, tx); dsl_scan_sync_state(scn, tx, SYNC_MANDATORY); spa_event_notify(scn->scn_dp->dp_spa, NULL, NULL, ESC_ZFS_SCRUB_ABORT); } int dsl_scan_cancel(dsl_pool_t *dp) { return (dsl_sync_task(spa_name(dp->dp_spa), dsl_scan_cancel_check, dsl_scan_cancel_sync, NULL, 3, ZFS_SPACE_CHECK_RESERVED)); } static int dsl_scrub_pause_resume_check(void *arg, dmu_tx_t *tx) { pool_scrub_cmd_t *cmd = arg; dsl_pool_t *dp = dmu_tx_pool(tx); dsl_scan_t *scn = dp->dp_scan; if (*cmd == POOL_SCRUB_PAUSE) { /* can't pause a scrub when there is no in-progress scrub */ if (!dsl_scan_scrubbing(dp)) return (SET_ERROR(ENOENT)); /* can't pause a paused scrub */ if (dsl_scan_is_paused_scrub(scn)) return (SET_ERROR(EBUSY)); } else if (*cmd != POOL_SCRUB_NORMAL) { return (SET_ERROR(ENOTSUP)); } return (0); } static void dsl_scrub_pause_resume_sync(void *arg, dmu_tx_t *tx) { pool_scrub_cmd_t *cmd = arg; dsl_pool_t *dp = dmu_tx_pool(tx); spa_t *spa = dp->dp_spa; dsl_scan_t *scn = dp->dp_scan; if (*cmd == POOL_SCRUB_PAUSE) { /* can't pause a scrub when there is no in-progress scrub */ spa->spa_scan_pass_scrub_pause = gethrestime_sec(); scn->scn_phys.scn_flags |= DSF_SCRUB_PAUSED; scn->scn_phys_cached.scn_flags |= DSF_SCRUB_PAUSED; dsl_scan_sync_state(scn, tx, SYNC_CACHED); spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_PAUSED); spa_notify_waiters(spa); } else { ASSERT3U(*cmd, ==, POOL_SCRUB_NORMAL); if (dsl_scan_is_paused_scrub(scn)) { /* * We need to keep track of how much time we spend * paused per pass so that we can adjust the scrub rate * shown in the output of 'zpool status' */ spa->spa_scan_pass_scrub_spent_paused += gethrestime_sec() - spa->spa_scan_pass_scrub_pause; spa->spa_scan_pass_scrub_pause = 0; scn->scn_phys.scn_flags &= ~DSF_SCRUB_PAUSED; scn->scn_phys_cached.scn_flags &= ~DSF_SCRUB_PAUSED; dsl_scan_sync_state(scn, tx, SYNC_CACHED); } } } /* * Set scrub pause/resume state if it makes sense to do so */ int dsl_scrub_set_pause_resume(const dsl_pool_t *dp, pool_scrub_cmd_t cmd) { return (dsl_sync_task(spa_name(dp->dp_spa), dsl_scrub_pause_resume_check, dsl_scrub_pause_resume_sync, &cmd, 3, ZFS_SPACE_CHECK_RESERVED)); } /* start a new scan, or restart an existing one. */ void dsl_scan_restart_resilver(dsl_pool_t *dp, uint64_t txg) { if (txg == 0) { dmu_tx_t *tx; tx = dmu_tx_create_dd(dp->dp_mos_dir); VERIFY(0 == dmu_tx_assign(tx, TXG_WAIT)); txg = dmu_tx_get_txg(tx); dp->dp_scan->scn_restart_txg = txg; dmu_tx_commit(tx); } else { dp->dp_scan->scn_restart_txg = txg; } zfs_dbgmsg("restarting resilver txg=%llu", (longlong_t)txg); } void dsl_free(dsl_pool_t *dp, uint64_t txg, const blkptr_t *bp) { zio_free(dp->dp_spa, txg, bp); } void dsl_free_sync(zio_t *pio, dsl_pool_t *dp, uint64_t txg, const blkptr_t *bpp) { ASSERT(dsl_pool_sync_context(dp)); zio_nowait(zio_free_sync(pio, dp->dp_spa, txg, bpp, pio->io_flags)); } static int scan_ds_queue_compare(const void *a, const void *b) { const scan_ds_t *sds_a = a, *sds_b = b; if (sds_a->sds_dsobj < sds_b->sds_dsobj) return (-1); if (sds_a->sds_dsobj == sds_b->sds_dsobj) return (0); return (1); } static void scan_ds_queue_clear(dsl_scan_t *scn) { void *cookie = NULL; scan_ds_t *sds; while ((sds = avl_destroy_nodes(&scn->scn_queue, &cookie)) != NULL) { kmem_free(sds, sizeof (*sds)); } } static boolean_t scan_ds_queue_contains(dsl_scan_t *scn, uint64_t dsobj, uint64_t *txg) { scan_ds_t srch, *sds; srch.sds_dsobj = dsobj; sds = avl_find(&scn->scn_queue, &srch, NULL); if (sds != NULL && txg != NULL) *txg = sds->sds_txg; return (sds != NULL); } static void scan_ds_queue_insert(dsl_scan_t *scn, uint64_t dsobj, uint64_t txg) { scan_ds_t *sds; avl_index_t where; sds = kmem_zalloc(sizeof (*sds), KM_SLEEP); sds->sds_dsobj = dsobj; sds->sds_txg = txg; VERIFY3P(avl_find(&scn->scn_queue, sds, &where), ==, NULL); avl_insert(&scn->scn_queue, sds, where); } static void scan_ds_queue_remove(dsl_scan_t *scn, uint64_t dsobj) { scan_ds_t srch, *sds; srch.sds_dsobj = dsobj; sds = avl_find(&scn->scn_queue, &srch, NULL); VERIFY(sds != NULL); avl_remove(&scn->scn_queue, sds); kmem_free(sds, sizeof (*sds)); } static void scan_ds_queue_sync(dsl_scan_t *scn, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; dmu_object_type_t ot = (spa_version(spa) >= SPA_VERSION_DSL_SCRUB) ? DMU_OT_SCAN_QUEUE : DMU_OT_ZAP_OTHER; ASSERT0(scn->scn_queues_pending); ASSERT(scn->scn_phys.scn_queue_obj != 0); VERIFY0(dmu_object_free(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, tx)); scn->scn_phys.scn_queue_obj = zap_create(dp->dp_meta_objset, ot, DMU_OT_NONE, 0, tx); for (scan_ds_t *sds = avl_first(&scn->scn_queue); sds != NULL; sds = AVL_NEXT(&scn->scn_queue, sds)) { VERIFY0(zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, sds->sds_dsobj, sds->sds_txg, tx)); } } /* * Computes the memory limit state that we're currently in. A sorted scan * needs quite a bit of memory to hold the sorting queue, so we need to * reasonably constrain the size so it doesn't impact overall system * performance. We compute two limits: * 1) Hard memory limit: if the amount of memory used by the sorting * queues on a pool gets above this value, we stop the metadata * scanning portion and start issuing the queued up and sorted * I/Os to reduce memory usage. * This limit is calculated as a fraction of physmem (by default 5%). * We constrain the lower bound of the hard limit to an absolute * minimum of zfs_scan_mem_lim_min (default: 16 MiB). We also constrain * the upper bound to 5% of the total pool size - no chance we'll * ever need that much memory, but just to keep the value in check. * 2) Soft memory limit: once we hit the hard memory limit, we start * issuing I/O to reduce queue memory usage, but we don't want to * completely empty out the queues, since we might be able to find I/Os * that will fill in the gaps of our non-sequential IOs at some point * in the future. So we stop the issuing of I/Os once the amount of * memory used drops below the soft limit (at which point we stop issuing * I/O and start scanning metadata again). * * This limit is calculated by subtracting a fraction of the hard * limit from the hard limit. By default this fraction is 5%, so * the soft limit is 95% of the hard limit. We cap the size of the * difference between the hard and soft limits at an absolute * maximum of zfs_scan_mem_lim_soft_max (default: 128 MiB) - this is * sufficient to not cause too frequent switching between the * metadata scan and I/O issue (even at 2k recordsize, 128 MiB's * worth of queues is about 1.2 GiB of on-pool data, so scanning * that should take at least a decent fraction of a second). */ static boolean_t dsl_scan_should_clear(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev; uint64_t alloc, mlim_hard, mlim_soft, mused; alloc = metaslab_class_get_alloc(spa_normal_class(spa)); alloc += metaslab_class_get_alloc(spa_special_class(spa)); alloc += metaslab_class_get_alloc(spa_dedup_class(spa)); mlim_hard = MAX((physmem / zfs_scan_mem_lim_fact) * PAGESIZE, zfs_scan_mem_lim_min); mlim_hard = MIN(mlim_hard, alloc / 20); mlim_soft = mlim_hard - MIN(mlim_hard / zfs_scan_mem_lim_soft_fact, zfs_scan_mem_lim_soft_max); mused = 0; for (uint64_t i = 0; i < rvd->vdev_children; i++) { vdev_t *tvd = rvd->vdev_child[i]; dsl_scan_io_queue_t *queue; mutex_enter(&tvd->vdev_scan_io_queue_lock); queue = tvd->vdev_scan_io_queue; if (queue != NULL) { /* * # of extents in exts_by_addr = # in exts_by_size. * B-tree efficiency is ~75%, but can be as low as 50%. */ mused += zfs_btree_numnodes(&queue->q_exts_by_size) * ((sizeof (range_seg_gap_t) + sizeof (uint64_t)) * 3 / 2) + queue->q_sio_memused; } mutex_exit(&tvd->vdev_scan_io_queue_lock); } dprintf("current scan memory usage: %llu bytes\n", (longlong_t)mused); if (mused == 0) ASSERT0(scn->scn_queues_pending); /* * If we are above our hard limit, we need to clear out memory. * If we are below our soft limit, we need to accumulate sequential IOs. * Otherwise, we should keep doing whatever we are currently doing. */ if (mused >= mlim_hard) return (B_TRUE); else if (mused < mlim_soft) return (B_FALSE); else return (scn->scn_clearing); } static boolean_t dsl_scan_check_suspend(dsl_scan_t *scn, const zbookmark_phys_t *zb) { /* we never skip user/group accounting objects */ if (zb && (int64_t)zb->zb_object < 0) return (B_FALSE); if (scn->scn_suspending) return (B_TRUE); /* we're already suspending */ if (!ZB_IS_ZERO(&scn->scn_phys.scn_bookmark)) return (B_FALSE); /* we're resuming */ /* We only know how to resume from level-0 and objset blocks. */ if (zb && (zb->zb_level != 0 && zb->zb_level != ZB_ROOT_LEVEL)) return (B_FALSE); /* * We suspend if: * - we have scanned for at least the minimum time (default 1 sec * for scrub, 3 sec for resilver), and either we have sufficient * dirty data that we are starting to write more quickly * (default 30%), someone is explicitly waiting for this txg * to complete, or we have used up all of the time in the txg * timeout (default 5 sec). * or * - the spa is shutting down because this pool is being exported * or the machine is rebooting. * or * - the scan queue has reached its memory use limit */ uint64_t curr_time_ns = gethrtime(); uint64_t scan_time_ns = curr_time_ns - scn->scn_sync_start_time; uint64_t sync_time_ns = curr_time_ns - scn->scn_dp->dp_spa->spa_sync_starttime; uint64_t dirty_min_bytes = zfs_dirty_data_max * zfs_vdev_async_write_active_min_dirty_percent / 100; int mintime = (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) ? zfs_resilver_min_time_ms : zfs_scrub_min_time_ms; if ((NSEC2MSEC(scan_time_ns) > mintime && (scn->scn_dp->dp_dirty_total >= dirty_min_bytes || txg_sync_waiting(scn->scn_dp) || NSEC2SEC(sync_time_ns) >= zfs_txg_timeout)) || spa_shutting_down(scn->scn_dp->dp_spa) || (zfs_scan_strict_mem_lim && dsl_scan_should_clear(scn))) { if (zb && zb->zb_level == ZB_ROOT_LEVEL) { dprintf("suspending at first available bookmark " "%llx/%llx/%llx/%llx\n", (longlong_t)zb->zb_objset, (longlong_t)zb->zb_object, (longlong_t)zb->zb_level, (longlong_t)zb->zb_blkid); SET_BOOKMARK(&scn->scn_phys.scn_bookmark, zb->zb_objset, 0, 0, 0); } else if (zb != NULL) { dprintf("suspending at bookmark %llx/%llx/%llx/%llx\n", (longlong_t)zb->zb_objset, (longlong_t)zb->zb_object, (longlong_t)zb->zb_level, (longlong_t)zb->zb_blkid); scn->scn_phys.scn_bookmark = *zb; } else { #ifdef ZFS_DEBUG dsl_scan_phys_t *scnp = &scn->scn_phys; dprintf("suspending at at DDT bookmark " "%llx/%llx/%llx/%llx\n", (longlong_t)scnp->scn_ddt_bookmark.ddb_class, (longlong_t)scnp->scn_ddt_bookmark.ddb_type, (longlong_t)scnp->scn_ddt_bookmark.ddb_checksum, (longlong_t)scnp->scn_ddt_bookmark.ddb_cursor); #endif } scn->scn_suspending = B_TRUE; return (B_TRUE); } return (B_FALSE); } typedef struct zil_scan_arg { dsl_pool_t *zsa_dp; zil_header_t *zsa_zh; } zil_scan_arg_t; static int dsl_scan_zil_block(zilog_t *zilog, const blkptr_t *bp, void *arg, uint64_t claim_txg) { (void) zilog; zil_scan_arg_t *zsa = arg; dsl_pool_t *dp = zsa->zsa_dp; dsl_scan_t *scn = dp->dp_scan; zil_header_t *zh = zsa->zsa_zh; zbookmark_phys_t zb; ASSERT(!BP_IS_REDACTED(bp)); if (BP_IS_HOLE(bp) || bp->blk_birth <= scn->scn_phys.scn_cur_min_txg) return (0); /* * One block ("stubby") can be allocated a long time ago; we * want to visit that one because it has been allocated * (on-disk) even if it hasn't been claimed (even though for * scrub there's nothing to do to it). */ if (claim_txg == 0 && bp->blk_birth >= spa_min_claim_txg(dp->dp_spa)) return (0); SET_BOOKMARK(&zb, zh->zh_log.blk_cksum.zc_word[ZIL_ZC_OBJSET], ZB_ZIL_OBJECT, ZB_ZIL_LEVEL, bp->blk_cksum.zc_word[ZIL_ZC_SEQ]); VERIFY(0 == scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb)); return (0); } static int dsl_scan_zil_record(zilog_t *zilog, const lr_t *lrc, void *arg, uint64_t claim_txg) { (void) zilog; if (lrc->lrc_txtype == TX_WRITE) { zil_scan_arg_t *zsa = arg; dsl_pool_t *dp = zsa->zsa_dp; dsl_scan_t *scn = dp->dp_scan; zil_header_t *zh = zsa->zsa_zh; const lr_write_t *lr = (const lr_write_t *)lrc; const blkptr_t *bp = &lr->lr_blkptr; zbookmark_phys_t zb; ASSERT(!BP_IS_REDACTED(bp)); if (BP_IS_HOLE(bp) || bp->blk_birth <= scn->scn_phys.scn_cur_min_txg) return (0); /* * birth can be < claim_txg if this record's txg is * already txg sync'ed (but this log block contains * other records that are not synced) */ if (claim_txg == 0 || bp->blk_birth < claim_txg) return (0); SET_BOOKMARK(&zb, zh->zh_log.blk_cksum.zc_word[ZIL_ZC_OBJSET], lr->lr_foid, ZB_ZIL_LEVEL, lr->lr_offset / BP_GET_LSIZE(bp)); VERIFY(0 == scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb)); } return (0); } static void dsl_scan_zil(dsl_pool_t *dp, zil_header_t *zh) { uint64_t claim_txg = zh->zh_claim_txg; zil_scan_arg_t zsa = { dp, zh }; zilog_t *zilog; ASSERT(spa_writeable(dp->dp_spa)); /* * We only want to visit blocks that have been claimed but not yet * replayed (or, in read-only mode, blocks that *would* be claimed). */ if (claim_txg == 0) return; zilog = zil_alloc(dp->dp_meta_objset, zh); (void) zil_parse(zilog, dsl_scan_zil_block, dsl_scan_zil_record, &zsa, claim_txg, B_FALSE); zil_free(zilog); } /* * We compare scan_prefetch_issue_ctx_t's based on their bookmarks. The idea * here is to sort the AVL tree by the order each block will be needed. */ static int scan_prefetch_queue_compare(const void *a, const void *b) { const scan_prefetch_issue_ctx_t *spic_a = a, *spic_b = b; const scan_prefetch_ctx_t *spc_a = spic_a->spic_spc; const scan_prefetch_ctx_t *spc_b = spic_b->spic_spc; return (zbookmark_compare(spc_a->spc_datablkszsec, spc_a->spc_indblkshift, spc_b->spc_datablkszsec, spc_b->spc_indblkshift, &spic_a->spic_zb, &spic_b->spic_zb)); } static void scan_prefetch_ctx_rele(scan_prefetch_ctx_t *spc, void *tag) { if (zfs_refcount_remove(&spc->spc_refcnt, tag) == 0) { zfs_refcount_destroy(&spc->spc_refcnt); kmem_free(spc, sizeof (scan_prefetch_ctx_t)); } } static scan_prefetch_ctx_t * scan_prefetch_ctx_create(dsl_scan_t *scn, dnode_phys_t *dnp, void *tag) { scan_prefetch_ctx_t *spc; spc = kmem_alloc(sizeof (scan_prefetch_ctx_t), KM_SLEEP); zfs_refcount_create(&spc->spc_refcnt); zfs_refcount_add(&spc->spc_refcnt, tag); spc->spc_scn = scn; if (dnp != NULL) { spc->spc_datablkszsec = dnp->dn_datablkszsec; spc->spc_indblkshift = dnp->dn_indblkshift; spc->spc_root = B_FALSE; } else { spc->spc_datablkszsec = 0; spc->spc_indblkshift = 0; spc->spc_root = B_TRUE; } return (spc); } static void scan_prefetch_ctx_add_ref(scan_prefetch_ctx_t *spc, void *tag) { zfs_refcount_add(&spc->spc_refcnt, tag); } static void scan_ds_prefetch_queue_clear(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; void *cookie = NULL; scan_prefetch_issue_ctx_t *spic = NULL; mutex_enter(&spa->spa_scrub_lock); while ((spic = avl_destroy_nodes(&scn->scn_prefetch_queue, &cookie)) != NULL) { scan_prefetch_ctx_rele(spic->spic_spc, scn); kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); } mutex_exit(&spa->spa_scrub_lock); } static boolean_t dsl_scan_check_prefetch_resume(scan_prefetch_ctx_t *spc, const zbookmark_phys_t *zb) { zbookmark_phys_t *last_zb = &spc->spc_scn->scn_prefetch_bookmark; dnode_phys_t tmp_dnp; dnode_phys_t *dnp = (spc->spc_root) ? NULL : &tmp_dnp; if (zb->zb_objset != last_zb->zb_objset) return (B_TRUE); if ((int64_t)zb->zb_object < 0) return (B_FALSE); tmp_dnp.dn_datablkszsec = spc->spc_datablkszsec; tmp_dnp.dn_indblkshift = spc->spc_indblkshift; if (zbookmark_subtree_completed(dnp, zb, last_zb)) return (B_TRUE); return (B_FALSE); } static void dsl_scan_prefetch(scan_prefetch_ctx_t *spc, blkptr_t *bp, zbookmark_phys_t *zb) { avl_index_t idx; dsl_scan_t *scn = spc->spc_scn; spa_t *spa = scn->scn_dp->dp_spa; scan_prefetch_issue_ctx_t *spic; if (zfs_no_scrub_prefetch || BP_IS_REDACTED(bp)) return; if (BP_IS_HOLE(bp) || bp->blk_birth <= scn->scn_phys.scn_cur_min_txg || (BP_GET_LEVEL(bp) == 0 && BP_GET_TYPE(bp) != DMU_OT_DNODE && BP_GET_TYPE(bp) != DMU_OT_OBJSET)) return; if (dsl_scan_check_prefetch_resume(spc, zb)) return; scan_prefetch_ctx_add_ref(spc, scn); spic = kmem_alloc(sizeof (scan_prefetch_issue_ctx_t), KM_SLEEP); spic->spic_spc = spc; spic->spic_bp = *bp; spic->spic_zb = *zb; /* * Add the IO to the queue of blocks to prefetch. This allows us to * prioritize blocks that we will need first for the main traversal * thread. */ mutex_enter(&spa->spa_scrub_lock); if (avl_find(&scn->scn_prefetch_queue, spic, &idx) != NULL) { /* this block is already queued for prefetch */ kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); scan_prefetch_ctx_rele(spc, scn); mutex_exit(&spa->spa_scrub_lock); return; } avl_insert(&scn->scn_prefetch_queue, spic, idx); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); } static void dsl_scan_prefetch_dnode(dsl_scan_t *scn, dnode_phys_t *dnp, uint64_t objset, uint64_t object) { int i; zbookmark_phys_t zb; scan_prefetch_ctx_t *spc; if (dnp->dn_nblkptr == 0 && !(dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR)) return; SET_BOOKMARK(&zb, objset, object, 0, 0); spc = scan_prefetch_ctx_create(scn, dnp, FTAG); for (i = 0; i < dnp->dn_nblkptr; i++) { zb.zb_level = BP_GET_LEVEL(&dnp->dn_blkptr[i]); zb.zb_blkid = i; dsl_scan_prefetch(spc, &dnp->dn_blkptr[i], &zb); } if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) { zb.zb_level = 0; zb.zb_blkid = DMU_SPILL_BLKID; dsl_scan_prefetch(spc, DN_SPILL_BLKPTR(dnp), &zb); } scan_prefetch_ctx_rele(spc, FTAG); } static void dsl_scan_prefetch_cb(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp, arc_buf_t *buf, void *private) { (void) zio; scan_prefetch_ctx_t *spc = private; dsl_scan_t *scn = spc->spc_scn; spa_t *spa = scn->scn_dp->dp_spa; /* broadcast that the IO has completed for rate limiting purposes */ mutex_enter(&spa->spa_scrub_lock); ASSERT3U(spa->spa_scrub_inflight, >=, BP_GET_PSIZE(bp)); spa->spa_scrub_inflight -= BP_GET_PSIZE(bp); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); /* if there was an error or we are done prefetching, just cleanup */ if (buf == NULL || scn->scn_prefetch_stop) goto out; if (BP_GET_LEVEL(bp) > 0) { int i; blkptr_t *cbp; int epb = BP_GET_LSIZE(bp) >> SPA_BLKPTRSHIFT; zbookmark_phys_t czb; for (i = 0, cbp = buf->b_data; i < epb; i++, cbp++) { SET_BOOKMARK(&czb, zb->zb_objset, zb->zb_object, zb->zb_level - 1, zb->zb_blkid * epb + i); dsl_scan_prefetch(spc, cbp, &czb); } } else if (BP_GET_TYPE(bp) == DMU_OT_DNODE) { dnode_phys_t *cdnp; int i; int epb = BP_GET_LSIZE(bp) >> DNODE_SHIFT; for (i = 0, cdnp = buf->b_data; i < epb; i += cdnp->dn_extra_slots + 1, cdnp += cdnp->dn_extra_slots + 1) { dsl_scan_prefetch_dnode(scn, cdnp, zb->zb_objset, zb->zb_blkid * epb + i); } } else if (BP_GET_TYPE(bp) == DMU_OT_OBJSET) { objset_phys_t *osp = buf->b_data; dsl_scan_prefetch_dnode(scn, &osp->os_meta_dnode, zb->zb_objset, DMU_META_DNODE_OBJECT); if (OBJSET_BUF_HAS_USERUSED(buf)) { dsl_scan_prefetch_dnode(scn, &osp->os_groupused_dnode, zb->zb_objset, DMU_GROUPUSED_OBJECT); dsl_scan_prefetch_dnode(scn, &osp->os_userused_dnode, zb->zb_objset, DMU_USERUSED_OBJECT); } } out: if (buf != NULL) arc_buf_destroy(buf, private); scan_prefetch_ctx_rele(spc, scn); } static void dsl_scan_prefetch_thread(void *arg) { dsl_scan_t *scn = arg; spa_t *spa = scn->scn_dp->dp_spa; scan_prefetch_issue_ctx_t *spic; /* loop until we are told to stop */ while (!scn->scn_prefetch_stop) { arc_flags_t flags = ARC_FLAG_NOWAIT | ARC_FLAG_PRESCIENT_PREFETCH | ARC_FLAG_PREFETCH; int zio_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCAN_THREAD; mutex_enter(&spa->spa_scrub_lock); /* * Wait until we have an IO to issue and are not above our * maximum in flight limit. */ while (!scn->scn_prefetch_stop && (avl_numnodes(&scn->scn_prefetch_queue) == 0 || spa->spa_scrub_inflight >= scn->scn_maxinflight_bytes)) { cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock); } /* recheck if we should stop since we waited for the cv */ if (scn->scn_prefetch_stop) { mutex_exit(&spa->spa_scrub_lock); break; } /* remove the prefetch IO from the tree */ spic = avl_first(&scn->scn_prefetch_queue); spa->spa_scrub_inflight += BP_GET_PSIZE(&spic->spic_bp); avl_remove(&scn->scn_prefetch_queue, spic); mutex_exit(&spa->spa_scrub_lock); if (BP_IS_PROTECTED(&spic->spic_bp)) { ASSERT(BP_GET_TYPE(&spic->spic_bp) == DMU_OT_DNODE || BP_GET_TYPE(&spic->spic_bp) == DMU_OT_OBJSET); ASSERT3U(BP_GET_LEVEL(&spic->spic_bp), ==, 0); zio_flags |= ZIO_FLAG_RAW; } /* issue the prefetch asynchronously */ (void) arc_read(scn->scn_zio_root, scn->scn_dp->dp_spa, &spic->spic_bp, dsl_scan_prefetch_cb, spic->spic_spc, ZIO_PRIORITY_SCRUB, zio_flags, &flags, &spic->spic_zb); kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); } ASSERT(scn->scn_prefetch_stop); /* free any prefetches we didn't get to complete */ mutex_enter(&spa->spa_scrub_lock); while ((spic = avl_first(&scn->scn_prefetch_queue)) != NULL) { avl_remove(&scn->scn_prefetch_queue, spic); scan_prefetch_ctx_rele(spic->spic_spc, scn); kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); } ASSERT0(avl_numnodes(&scn->scn_prefetch_queue)); mutex_exit(&spa->spa_scrub_lock); } static boolean_t dsl_scan_check_resume(dsl_scan_t *scn, const dnode_phys_t *dnp, const zbookmark_phys_t *zb) { /* * We never skip over user/group accounting objects (obj<0) */ if (!ZB_IS_ZERO(&scn->scn_phys.scn_bookmark) && (int64_t)zb->zb_object >= 0) { /* * If we already visited this bp & everything below (in * a prior txg sync), don't bother doing it again. */ if (zbookmark_subtree_completed(dnp, zb, &scn->scn_phys.scn_bookmark)) return (B_TRUE); /* * If we found the block we're trying to resume from, or * we went past it, zero it out to indicate that it's OK * to start checking for suspending again. */ if (zbookmark_subtree_tbd(dnp, zb, &scn->scn_phys.scn_bookmark)) { dprintf("resuming at %llx/%llx/%llx/%llx\n", (longlong_t)zb->zb_objset, (longlong_t)zb->zb_object, (longlong_t)zb->zb_level, (longlong_t)zb->zb_blkid); bzero(&scn->scn_phys.scn_bookmark, sizeof (*zb)); } } return (B_FALSE); } static void dsl_scan_visitbp(blkptr_t *bp, const zbookmark_phys_t *zb, dnode_phys_t *dnp, dsl_dataset_t *ds, dsl_scan_t *scn, dmu_objset_type_t ostype, dmu_tx_t *tx); inline __attribute__((always_inline)) static void dsl_scan_visitdnode( dsl_scan_t *, dsl_dataset_t *ds, dmu_objset_type_t ostype, dnode_phys_t *dnp, uint64_t object, dmu_tx_t *tx); /* * Return nonzero on i/o error. * Return new buf to write out in *bufp. */ inline __attribute__((always_inline)) static int dsl_scan_recurse(dsl_scan_t *scn, dsl_dataset_t *ds, dmu_objset_type_t ostype, dnode_phys_t *dnp, const blkptr_t *bp, const zbookmark_phys_t *zb, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; int zio_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCAN_THREAD; int err; ASSERT(!BP_IS_REDACTED(bp)); /* * There is an unlikely case of encountering dnodes with contradicting * dn_bonuslen and DNODE_FLAG_SPILL_BLKPTR flag before in files created * or modified before commit 4254acb was merged. As it is not possible * to know which of the two is correct, report an error. */ if (dnp != NULL && dnp->dn_bonuslen > DN_MAX_BONUS_LEN(dnp)) { scn->scn_phys.scn_errors++; spa_log_error(dp->dp_spa, zb); return (SET_ERROR(EINVAL)); } if (BP_GET_LEVEL(bp) > 0) { arc_flags_t flags = ARC_FLAG_WAIT; int i; blkptr_t *cbp; int epb = BP_GET_LSIZE(bp) >> SPA_BLKPTRSHIFT; arc_buf_t *buf; err = arc_read(NULL, dp->dp_spa, bp, arc_getbuf_func, &buf, ZIO_PRIORITY_SCRUB, zio_flags, &flags, zb); if (err) { scn->scn_phys.scn_errors++; return (err); } for (i = 0, cbp = buf->b_data; i < epb; i++, cbp++) { zbookmark_phys_t czb; SET_BOOKMARK(&czb, zb->zb_objset, zb->zb_object, zb->zb_level - 1, zb->zb_blkid * epb + i); dsl_scan_visitbp(cbp, &czb, dnp, ds, scn, ostype, tx); } arc_buf_destroy(buf, &buf); } else if (BP_GET_TYPE(bp) == DMU_OT_DNODE) { arc_flags_t flags = ARC_FLAG_WAIT; dnode_phys_t *cdnp; int i; int epb = BP_GET_LSIZE(bp) >> DNODE_SHIFT; arc_buf_t *buf; if (BP_IS_PROTECTED(bp)) { ASSERT3U(BP_GET_COMPRESS(bp), ==, ZIO_COMPRESS_OFF); zio_flags |= ZIO_FLAG_RAW; } err = arc_read(NULL, dp->dp_spa, bp, arc_getbuf_func, &buf, ZIO_PRIORITY_SCRUB, zio_flags, &flags, zb); if (err) { scn->scn_phys.scn_errors++; return (err); } for (i = 0, cdnp = buf->b_data; i < epb; i += cdnp->dn_extra_slots + 1, cdnp += cdnp->dn_extra_slots + 1) { dsl_scan_visitdnode(scn, ds, ostype, cdnp, zb->zb_blkid * epb + i, tx); } arc_buf_destroy(buf, &buf); } else if (BP_GET_TYPE(bp) == DMU_OT_OBJSET) { arc_flags_t flags = ARC_FLAG_WAIT; objset_phys_t *osp; arc_buf_t *buf; err = arc_read(NULL, dp->dp_spa, bp, arc_getbuf_func, &buf, ZIO_PRIORITY_SCRUB, zio_flags, &flags, zb); if (err) { scn->scn_phys.scn_errors++; return (err); } osp = buf->b_data; dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_meta_dnode, DMU_META_DNODE_OBJECT, tx); if (OBJSET_BUF_HAS_USERUSED(buf)) { /* * We also always visit user/group/project accounting * objects, and never skip them, even if we are * suspending. This is necessary so that the * space deltas from this txg get integrated. */ if (OBJSET_BUF_HAS_PROJECTUSED(buf)) dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_projectused_dnode, DMU_PROJECTUSED_OBJECT, tx); dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_groupused_dnode, DMU_GROUPUSED_OBJECT, tx); dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_userused_dnode, DMU_USERUSED_OBJECT, tx); } arc_buf_destroy(buf, &buf); } return (0); } inline __attribute__((always_inline)) static void dsl_scan_visitdnode(dsl_scan_t *scn, dsl_dataset_t *ds, dmu_objset_type_t ostype, dnode_phys_t *dnp, uint64_t object, dmu_tx_t *tx) { int j; for (j = 0; j < dnp->dn_nblkptr; j++) { zbookmark_phys_t czb; SET_BOOKMARK(&czb, ds ? ds->ds_object : 0, object, dnp->dn_nlevels - 1, j); dsl_scan_visitbp(&dnp->dn_blkptr[j], &czb, dnp, ds, scn, ostype, tx); } if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) { zbookmark_phys_t czb; SET_BOOKMARK(&czb, ds ? ds->ds_object : 0, object, 0, DMU_SPILL_BLKID); dsl_scan_visitbp(DN_SPILL_BLKPTR(dnp), &czb, dnp, ds, scn, ostype, tx); } } /* * The arguments are in this order because mdb can only print the * first 5; we want them to be useful. */ static void dsl_scan_visitbp(blkptr_t *bp, const zbookmark_phys_t *zb, dnode_phys_t *dnp, dsl_dataset_t *ds, dsl_scan_t *scn, dmu_objset_type_t ostype, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; blkptr_t *bp_toread = NULL; if (dsl_scan_check_suspend(scn, zb)) return; if (dsl_scan_check_resume(scn, dnp, zb)) return; scn->scn_visited_this_txg++; /* * This debugging is commented out to conserve stack space. This * function is called recursively and the debugging adds several * bytes to the stack for each call. It can be commented back in * if required to debug an issue in dsl_scan_visitbp(). * * dprintf_bp(bp, * "visiting ds=%p/%llu zb=%llx/%llx/%llx/%llx bp=%p", * ds, ds ? ds->ds_object : 0, * zb->zb_objset, zb->zb_object, zb->zb_level, zb->zb_blkid, * bp); */ if (BP_IS_HOLE(bp)) { scn->scn_holes_this_txg++; return; } if (BP_IS_REDACTED(bp)) { ASSERT(dsl_dataset_feature_is_active(ds, SPA_FEATURE_REDACTED_DATASETS)); return; } /* * Check if this block contradicts any filesystem flags. */ spa_feature_t f = SPA_FEATURE_LARGE_BLOCKS; if (BP_GET_LSIZE(bp) > SPA_OLD_MAXBLOCKSIZE) ASSERT(dsl_dataset_feature_is_active(ds, f)); f = zio_checksum_to_feature(BP_GET_CHECKSUM(bp)); if (f != SPA_FEATURE_NONE) ASSERT(dsl_dataset_feature_is_active(ds, f)); f = zio_compress_to_feature(BP_GET_COMPRESS(bp)); if (f != SPA_FEATURE_NONE) ASSERT(dsl_dataset_feature_is_active(ds, f)); if (bp->blk_birth <= scn->scn_phys.scn_cur_min_txg) { scn->scn_lt_min_this_txg++; return; } bp_toread = kmem_alloc(sizeof (blkptr_t), KM_SLEEP); *bp_toread = *bp; if (dsl_scan_recurse(scn, ds, ostype, dnp, bp_toread, zb, tx) != 0) goto out; /* * If dsl_scan_ddt() has already visited this block, it will have * already done any translations or scrubbing, so don't call the * callback again. */ if (ddt_class_contains(dp->dp_spa, scn->scn_phys.scn_ddt_class_max, bp)) { scn->scn_ddt_contained_this_txg++; goto out; } /* * If this block is from the future (after cur_max_txg), then we * are doing this on behalf of a deleted snapshot, and we will * revisit the future block on the next pass of this dataset. * Don't scan it now unless we need to because something * under it was modified. */ if (BP_PHYSICAL_BIRTH(bp) > scn->scn_phys.scn_cur_max_txg) { scn->scn_gt_max_this_txg++; goto out; } scan_funcs[scn->scn_phys.scn_func](dp, bp, zb); out: kmem_free(bp_toread, sizeof (blkptr_t)); } static void dsl_scan_visit_rootbp(dsl_scan_t *scn, dsl_dataset_t *ds, blkptr_t *bp, dmu_tx_t *tx) { zbookmark_phys_t zb; scan_prefetch_ctx_t *spc; SET_BOOKMARK(&zb, ds ? ds->ds_object : DMU_META_OBJSET, ZB_ROOT_OBJECT, ZB_ROOT_LEVEL, ZB_ROOT_BLKID); if (ZB_IS_ZERO(&scn->scn_phys.scn_bookmark)) { SET_BOOKMARK(&scn->scn_prefetch_bookmark, zb.zb_objset, 0, 0, 0); } else { scn->scn_prefetch_bookmark = scn->scn_phys.scn_bookmark; } scn->scn_objsets_visited_this_txg++; spc = scan_prefetch_ctx_create(scn, NULL, FTAG); dsl_scan_prefetch(spc, bp, &zb); scan_prefetch_ctx_rele(spc, FTAG); dsl_scan_visitbp(bp, &zb, NULL, ds, scn, DMU_OST_NONE, tx); dprintf_ds(ds, "finished scan%s", ""); } static void ds_destroyed_scn_phys(dsl_dataset_t *ds, dsl_scan_phys_t *scn_phys) { if (scn_phys->scn_bookmark.zb_objset == ds->ds_object) { if (ds->ds_is_snapshot) { /* * Note: * - scn_cur_{min,max}_txg stays the same. * - Setting the flag is not really necessary if * scn_cur_max_txg == scn_max_txg, because there * is nothing after this snapshot that we care * about. However, we set it anyway and then * ignore it when we retraverse it in * dsl_scan_visitds(). */ scn_phys->scn_bookmark.zb_objset = dsl_dataset_phys(ds)->ds_next_snap_obj; zfs_dbgmsg("destroying ds %llu; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds->ds_object, (u_longlong_t)dsl_dataset_phys(ds)-> ds_next_snap_obj); scn_phys->scn_flags |= DSF_VISIT_DS_AGAIN; } else { SET_BOOKMARK(&scn_phys->scn_bookmark, ZB_DESTROYED_OBJSET, 0, 0, 0); zfs_dbgmsg("destroying ds %llu; currently traversing; " "reset bookmark to -1,0,0,0", (u_longlong_t)ds->ds_object); } } } /* * Invoked when a dataset is destroyed. We need to make sure that: * * 1) If it is the dataset that was currently being scanned, we write * a new dsl_scan_phys_t and marking the objset reference in it * as destroyed. * 2) Remove it from the work queue, if it was present. * * If the dataset was actually a snapshot, instead of marking the dataset * as destroyed, we instead substitute the next snapshot in line. */ void dsl_scan_ds_destroyed(dsl_dataset_t *ds, dmu_tx_t *tx) { dsl_pool_t *dp = ds->ds_dir->dd_pool; dsl_scan_t *scn = dp->dp_scan; uint64_t mintxg; if (!dsl_scan_is_running(scn)) return; ds_destroyed_scn_phys(ds, &scn->scn_phys); ds_destroyed_scn_phys(ds, &scn->scn_phys_cached); if (scan_ds_queue_contains(scn, ds->ds_object, &mintxg)) { scan_ds_queue_remove(scn, ds->ds_object); if (ds->ds_is_snapshot) scan_ds_queue_insert(scn, dsl_dataset_phys(ds)->ds_next_snap_obj, mintxg); } if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, &mintxg) == 0) { ASSERT3U(dsl_dataset_phys(ds)->ds_num_children, <=, 1); VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, tx)); if (ds->ds_is_snapshot) { /* * We keep the same mintxg; it could be > * ds_creation_txg if the previous snapshot was * deleted too. */ VERIFY(zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, dsl_dataset_phys(ds)->ds_next_snap_obj, mintxg, tx) == 0); zfs_dbgmsg("destroying ds %llu; in queue; " "replacing with %llu", (u_longlong_t)ds->ds_object, (u_longlong_t)dsl_dataset_phys(ds)-> ds_next_snap_obj); } else { zfs_dbgmsg("destroying ds %llu; in queue; removing", (u_longlong_t)ds->ds_object); } } /* * dsl_scan_sync() should be called after this, and should sync * out our changed state, but just to be safe, do it here. */ dsl_scan_sync_state(scn, tx, SYNC_CACHED); } static void ds_snapshotted_bookmark(dsl_dataset_t *ds, zbookmark_phys_t *scn_bookmark) { if (scn_bookmark->zb_objset == ds->ds_object) { scn_bookmark->zb_objset = dsl_dataset_phys(ds)->ds_prev_snap_obj; zfs_dbgmsg("snapshotting ds %llu; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds->ds_object, (u_longlong_t)dsl_dataset_phys(ds)->ds_prev_snap_obj); } } /* * Called when a dataset is snapshotted. If we were currently traversing * this snapshot, we reset our bookmark to point at the newly created * snapshot. We also modify our work queue to remove the old snapshot and * replace with the new one. */ void dsl_scan_ds_snapshotted(dsl_dataset_t *ds, dmu_tx_t *tx) { dsl_pool_t *dp = ds->ds_dir->dd_pool; dsl_scan_t *scn = dp->dp_scan; uint64_t mintxg; if (!dsl_scan_is_running(scn)) return; ASSERT(dsl_dataset_phys(ds)->ds_prev_snap_obj != 0); ds_snapshotted_bookmark(ds, &scn->scn_phys.scn_bookmark); ds_snapshotted_bookmark(ds, &scn->scn_phys_cached.scn_bookmark); if (scan_ds_queue_contains(scn, ds->ds_object, &mintxg)) { scan_ds_queue_remove(scn, ds->ds_object); scan_ds_queue_insert(scn, dsl_dataset_phys(ds)->ds_prev_snap_obj, mintxg); } if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, &mintxg) == 0) { VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, tx)); VERIFY(zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, dsl_dataset_phys(ds)->ds_prev_snap_obj, mintxg, tx) == 0); zfs_dbgmsg("snapshotting ds %llu; in queue; " "replacing with %llu", (u_longlong_t)ds->ds_object, (u_longlong_t)dsl_dataset_phys(ds)->ds_prev_snap_obj); } dsl_scan_sync_state(scn, tx, SYNC_CACHED); } static void ds_clone_swapped_bookmark(dsl_dataset_t *ds1, dsl_dataset_t *ds2, zbookmark_phys_t *scn_bookmark) { if (scn_bookmark->zb_objset == ds1->ds_object) { scn_bookmark->zb_objset = ds2->ds_object; zfs_dbgmsg("clone_swap ds %llu; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds1->ds_object, (u_longlong_t)ds2->ds_object); } else if (scn_bookmark->zb_objset == ds2->ds_object) { scn_bookmark->zb_objset = ds1->ds_object; zfs_dbgmsg("clone_swap ds %llu; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds2->ds_object, (u_longlong_t)ds1->ds_object); } } /* * Called when an origin dataset and its clone are swapped. If we were * currently traversing the dataset, we need to switch to traversing the * newly promoted clone. */ void dsl_scan_ds_clone_swapped(dsl_dataset_t *ds1, dsl_dataset_t *ds2, dmu_tx_t *tx) { dsl_pool_t *dp = ds1->ds_dir->dd_pool; dsl_scan_t *scn = dp->dp_scan; uint64_t mintxg1, mintxg2; boolean_t ds1_queued, ds2_queued; if (!dsl_scan_is_running(scn)) return; ds_clone_swapped_bookmark(ds1, ds2, &scn->scn_phys.scn_bookmark); ds_clone_swapped_bookmark(ds1, ds2, &scn->scn_phys_cached.scn_bookmark); /* * Handle the in-memory scan queue. */ ds1_queued = scan_ds_queue_contains(scn, ds1->ds_object, &mintxg1); ds2_queued = scan_ds_queue_contains(scn, ds2->ds_object, &mintxg2); /* Sanity checking. */ if (ds1_queued) { ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); } if (ds2_queued) { ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); } if (ds1_queued && ds2_queued) { /* * If both are queued, we don't need to do anything. * The swapping code below would not handle this case correctly, * since we can't insert ds2 if it is already there. That's * because scan_ds_queue_insert() prohibits a duplicate insert * and panics. */ } else if (ds1_queued) { scan_ds_queue_remove(scn, ds1->ds_object); scan_ds_queue_insert(scn, ds2->ds_object, mintxg1); } else if (ds2_queued) { scan_ds_queue_remove(scn, ds2->ds_object); scan_ds_queue_insert(scn, ds1->ds_object, mintxg2); } /* * Handle the on-disk scan queue. * The on-disk state is an out-of-date version of the in-memory state, * so the in-memory and on-disk values for ds1_queued and ds2_queued may * be different. Therefore we need to apply the swap logic to the * on-disk state independently of the in-memory state. */ ds1_queued = zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, &mintxg1) == 0; ds2_queued = zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds2->ds_object, &mintxg2) == 0; /* Sanity checking. */ if (ds1_queued) { ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); } if (ds2_queued) { ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); } if (ds1_queued && ds2_queued) { /* * If both are queued, we don't need to do anything. * Alternatively, we could check for EEXIST from * zap_add_int_key() and back out to the original state, but * that would be more work than checking for this case upfront. */ } else if (ds1_queued) { VERIFY3S(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, tx)); VERIFY3S(0, ==, zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds2->ds_object, mintxg1, tx)); zfs_dbgmsg("clone_swap ds %llu; in queue; " "replacing with %llu", (u_longlong_t)ds1->ds_object, (u_longlong_t)ds2->ds_object); } else if (ds2_queued) { VERIFY3S(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds2->ds_object, tx)); VERIFY3S(0, ==, zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, mintxg2, tx)); zfs_dbgmsg("clone_swap ds %llu; in queue; " "replacing with %llu", (u_longlong_t)ds2->ds_object, (u_longlong_t)ds1->ds_object); } dsl_scan_sync_state(scn, tx, SYNC_CACHED); } static int enqueue_clones_cb(dsl_pool_t *dp, dsl_dataset_t *hds, void *arg) { uint64_t originobj = *(uint64_t *)arg; dsl_dataset_t *ds; int err; dsl_scan_t *scn = dp->dp_scan; if (dsl_dir_phys(hds->ds_dir)->dd_origin_obj != originobj) return (0); err = dsl_dataset_hold_obj(dp, hds->ds_object, FTAG, &ds); if (err) return (err); while (dsl_dataset_phys(ds)->ds_prev_snap_obj != originobj) { dsl_dataset_t *prev; err = dsl_dataset_hold_obj(dp, dsl_dataset_phys(ds)->ds_prev_snap_obj, FTAG, &prev); dsl_dataset_rele(ds, FTAG); if (err) return (err); ds = prev; } scan_ds_queue_insert(scn, ds->ds_object, dsl_dataset_phys(ds)->ds_prev_snap_txg); dsl_dataset_rele(ds, FTAG); return (0); } static void dsl_scan_visitds(dsl_scan_t *scn, uint64_t dsobj, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; dsl_dataset_t *ds; VERIFY3U(0, ==, dsl_dataset_hold_obj(dp, dsobj, FTAG, &ds)); if (scn->scn_phys.scn_cur_min_txg >= scn->scn_phys.scn_max_txg) { /* * This can happen if this snapshot was created after the * scan started, and we already completed a previous snapshot * that was created after the scan started. This snapshot * only references blocks with: * * birth < our ds_creation_txg * cur_min_txg is no less than ds_creation_txg. * We have already visited these blocks. * or * birth > scn_max_txg * The scan requested not to visit these blocks. * * Subsequent snapshots (and clones) can reference our * blocks, or blocks with even higher birth times. * Therefore we do not need to visit them either, * so we do not add them to the work queue. * * Note that checking for cur_min_txg >= cur_max_txg * is not sufficient, because in that case we may need to * visit subsequent snapshots. This happens when min_txg > 0, * which raises cur_min_txg. In this case we will visit * this dataset but skip all of its blocks, because the * rootbp's birth time is < cur_min_txg. Then we will * add the next snapshots/clones to the work queue. */ char *dsname = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP); dsl_dataset_name(ds, dsname); zfs_dbgmsg("scanning dataset %llu (%s) is unnecessary because " "cur_min_txg (%llu) >= max_txg (%llu)", (longlong_t)dsobj, dsname, (longlong_t)scn->scn_phys.scn_cur_min_txg, (longlong_t)scn->scn_phys.scn_max_txg); kmem_free(dsname, MAXNAMELEN); goto out; } /* * Only the ZIL in the head (non-snapshot) is valid. Even though * snapshots can have ZIL block pointers (which may be the same * BP as in the head), they must be ignored. In addition, $ORIGIN * doesn't have a objset (i.e. its ds_bp is a hole) so we don't * need to look for a ZIL in it either. So we traverse the ZIL here, * rather than in scan_recurse(), because the regular snapshot * block-sharing rules don't apply to it. */ if (!dsl_dataset_is_snapshot(ds) && (dp->dp_origin_snap == NULL || ds->ds_dir != dp->dp_origin_snap->ds_dir)) { objset_t *os; if (dmu_objset_from_ds(ds, &os) != 0) { goto out; } dsl_scan_zil(dp, &os->os_zil_header); } /* * Iterate over the bps in this ds. */ dmu_buf_will_dirty(ds->ds_dbuf, tx); rrw_enter(&ds->ds_bp_rwlock, RW_READER, FTAG); dsl_scan_visit_rootbp(scn, ds, &dsl_dataset_phys(ds)->ds_bp, tx); rrw_exit(&ds->ds_bp_rwlock, FTAG); char *dsname = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP); dsl_dataset_name(ds, dsname); zfs_dbgmsg("scanned dataset %llu (%s) with min=%llu max=%llu; " "suspending=%u", (longlong_t)dsobj, dsname, (longlong_t)scn->scn_phys.scn_cur_min_txg, (longlong_t)scn->scn_phys.scn_cur_max_txg, (int)scn->scn_suspending); kmem_free(dsname, ZFS_MAX_DATASET_NAME_LEN); if (scn->scn_suspending) goto out; /* * We've finished this pass over this dataset. */ /* * If we did not completely visit this dataset, do another pass. */ if (scn->scn_phys.scn_flags & DSF_VISIT_DS_AGAIN) { zfs_dbgmsg("incomplete pass; visiting again"); scn->scn_phys.scn_flags &= ~DSF_VISIT_DS_AGAIN; scan_ds_queue_insert(scn, ds->ds_object, scn->scn_phys.scn_cur_max_txg); goto out; } /* * Add descendant datasets to work queue. */ if (dsl_dataset_phys(ds)->ds_next_snap_obj != 0) { scan_ds_queue_insert(scn, dsl_dataset_phys(ds)->ds_next_snap_obj, dsl_dataset_phys(ds)->ds_creation_txg); } if (dsl_dataset_phys(ds)->ds_num_children > 1) { boolean_t usenext = B_FALSE; if (dsl_dataset_phys(ds)->ds_next_clones_obj != 0) { uint64_t count; /* * A bug in a previous version of the code could * cause upgrade_clones_cb() to not set * ds_next_snap_obj when it should, leading to a * missing entry. Therefore we can only use the * next_clones_obj when its count is correct. */ int err = zap_count(dp->dp_meta_objset, dsl_dataset_phys(ds)->ds_next_clones_obj, &count); if (err == 0 && count == dsl_dataset_phys(ds)->ds_num_children - 1) usenext = B_TRUE; } if (usenext) { zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, dp->dp_meta_objset, dsl_dataset_phys(ds)->ds_next_clones_obj); zap_cursor_retrieve(&zc, &za) == 0; (void) zap_cursor_advance(&zc)) { scan_ds_queue_insert(scn, zfs_strtonum(za.za_name, NULL), dsl_dataset_phys(ds)->ds_creation_txg); } zap_cursor_fini(&zc); } else { VERIFY0(dmu_objset_find_dp(dp, dp->dp_root_dir_obj, enqueue_clones_cb, &ds->ds_object, DS_FIND_CHILDREN)); } } out: dsl_dataset_rele(ds, FTAG); } static int enqueue_cb(dsl_pool_t *dp, dsl_dataset_t *hds, void *arg) { (void) arg; dsl_dataset_t *ds; int err; dsl_scan_t *scn = dp->dp_scan; err = dsl_dataset_hold_obj(dp, hds->ds_object, FTAG, &ds); if (err) return (err); while (dsl_dataset_phys(ds)->ds_prev_snap_obj != 0) { dsl_dataset_t *prev; err = dsl_dataset_hold_obj(dp, dsl_dataset_phys(ds)->ds_prev_snap_obj, FTAG, &prev); if (err) { dsl_dataset_rele(ds, FTAG); return (err); } /* * If this is a clone, we don't need to worry about it for now. */ if (dsl_dataset_phys(prev)->ds_next_snap_obj != ds->ds_object) { dsl_dataset_rele(ds, FTAG); dsl_dataset_rele(prev, FTAG); return (0); } dsl_dataset_rele(ds, FTAG); ds = prev; } scan_ds_queue_insert(scn, ds->ds_object, dsl_dataset_phys(ds)->ds_prev_snap_txg); dsl_dataset_rele(ds, FTAG); return (0); } void dsl_scan_ddt_entry(dsl_scan_t *scn, enum zio_checksum checksum, ddt_entry_t *dde, dmu_tx_t *tx) { (void) tx; const ddt_key_t *ddk = &dde->dde_key; ddt_phys_t *ddp = dde->dde_phys; blkptr_t bp; zbookmark_phys_t zb = { 0 }; if (!dsl_scan_is_running(scn)) return; /* * This function is special because it is the only thing * that can add scan_io_t's to the vdev scan queues from * outside dsl_scan_sync(). For the most part this is ok * as long as it is called from within syncing context. * However, dsl_scan_sync() expects that no new sio's will * be added between when all the work for a scan is done * and the next txg when the scan is actually marked as * completed. This check ensures we do not issue new sio's * during this period. */ if (scn->scn_done_txg != 0) return; for (int p = 0; p < DDT_PHYS_TYPES; p++, ddp++) { if (ddp->ddp_phys_birth == 0 || ddp->ddp_phys_birth > scn->scn_phys.scn_max_txg) continue; ddt_bp_create(checksum, ddk, ddp, &bp); scn->scn_visited_this_txg++; scan_funcs[scn->scn_phys.scn_func](scn->scn_dp, &bp, &zb); } } /* * Scrub/dedup interaction. * * If there are N references to a deduped block, we don't want to scrub it * N times -- ideally, we should scrub it exactly once. * * We leverage the fact that the dde's replication class (enum ddt_class) * is ordered from highest replication class (DDT_CLASS_DITTO) to lowest * (DDT_CLASS_UNIQUE) so that we may walk the DDT in that order. * * To prevent excess scrubbing, the scrub begins by walking the DDT * to find all blocks with refcnt > 1, and scrubs each of these once. * Since there are two replication classes which contain blocks with * refcnt > 1, we scrub the highest replication class (DDT_CLASS_DITTO) first. * Finally the top-down scrub begins, only visiting blocks with refcnt == 1. * * There would be nothing more to say if a block's refcnt couldn't change * during a scrub, but of course it can so we must account for changes * in a block's replication class. * * Here's an example of what can occur: * * If a block has refcnt > 1 during the DDT scrub phase, but has refcnt == 1 * when visited during the top-down scrub phase, it will be scrubbed twice. * This negates our scrub optimization, but is otherwise harmless. * * If a block has refcnt == 1 during the DDT scrub phase, but has refcnt > 1 * on each visit during the top-down scrub phase, it will never be scrubbed. * To catch this, ddt_sync_entry() notifies the scrub code whenever a block's * reference class transitions to a higher level (i.e DDT_CLASS_UNIQUE to * DDT_CLASS_DUPLICATE); if it transitions from refcnt == 1 to refcnt > 1 * while a scrub is in progress, it scrubs the block right then. */ static void dsl_scan_ddt(dsl_scan_t *scn, dmu_tx_t *tx) { ddt_bookmark_t *ddb = &scn->scn_phys.scn_ddt_bookmark; ddt_entry_t dde; int error; uint64_t n = 0; bzero(&dde, sizeof (ddt_entry_t)); while ((error = ddt_walk(scn->scn_dp->dp_spa, ddb, &dde)) == 0) { ddt_t *ddt; if (ddb->ddb_class > scn->scn_phys.scn_ddt_class_max) break; dprintf("visiting ddb=%llu/%llu/%llu/%llx\n", (longlong_t)ddb->ddb_class, (longlong_t)ddb->ddb_type, (longlong_t)ddb->ddb_checksum, (longlong_t)ddb->ddb_cursor); /* There should be no pending changes to the dedup table */ ddt = scn->scn_dp->dp_spa->spa_ddt[ddb->ddb_checksum]; ASSERT(avl_first(&ddt->ddt_tree) == NULL); dsl_scan_ddt_entry(scn, ddb->ddb_checksum, &dde, tx); n++; if (dsl_scan_check_suspend(scn, NULL)) break; } zfs_dbgmsg("scanned %llu ddt entries with class_max = %u; " "suspending=%u", (longlong_t)n, (int)scn->scn_phys.scn_ddt_class_max, (int)scn->scn_suspending); ASSERT(error == 0 || error == ENOENT); ASSERT(error != ENOENT || ddb->ddb_class > scn->scn_phys.scn_ddt_class_max); } static uint64_t dsl_scan_ds_maxtxg(dsl_dataset_t *ds) { uint64_t smt = ds->ds_dir->dd_pool->dp_scan->scn_phys.scn_max_txg; if (ds->ds_is_snapshot) return (MIN(smt, dsl_dataset_phys(ds)->ds_creation_txg)); return (smt); } static void dsl_scan_visit(dsl_scan_t *scn, dmu_tx_t *tx) { scan_ds_t *sds; dsl_pool_t *dp = scn->scn_dp; if (scn->scn_phys.scn_ddt_bookmark.ddb_class <= scn->scn_phys.scn_ddt_class_max) { scn->scn_phys.scn_cur_min_txg = scn->scn_phys.scn_min_txg; scn->scn_phys.scn_cur_max_txg = scn->scn_phys.scn_max_txg; dsl_scan_ddt(scn, tx); if (scn->scn_suspending) return; } if (scn->scn_phys.scn_bookmark.zb_objset == DMU_META_OBJSET) { /* First do the MOS & ORIGIN */ scn->scn_phys.scn_cur_min_txg = scn->scn_phys.scn_min_txg; scn->scn_phys.scn_cur_max_txg = scn->scn_phys.scn_max_txg; dsl_scan_visit_rootbp(scn, NULL, &dp->dp_meta_rootbp, tx); spa_set_rootblkptr(dp->dp_spa, &dp->dp_meta_rootbp); if (scn->scn_suspending) return; if (spa_version(dp->dp_spa) < SPA_VERSION_DSL_SCRUB) { VERIFY0(dmu_objset_find_dp(dp, dp->dp_root_dir_obj, enqueue_cb, NULL, DS_FIND_CHILDREN)); } else { dsl_scan_visitds(scn, dp->dp_origin_snap->ds_object, tx); } ASSERT(!scn->scn_suspending); } else if (scn->scn_phys.scn_bookmark.zb_objset != ZB_DESTROYED_OBJSET) { uint64_t dsobj = scn->scn_phys.scn_bookmark.zb_objset; /* * If we were suspended, continue from here. Note if the * ds we were suspended on was deleted, the zb_objset may * be -1, so we will skip this and find a new objset * below. */ dsl_scan_visitds(scn, dsobj, tx); if (scn->scn_suspending) return; } /* * In case we suspended right at the end of the ds, zero the * bookmark so we don't think that we're still trying to resume. */ bzero(&scn->scn_phys.scn_bookmark, sizeof (zbookmark_phys_t)); /* * Keep pulling things out of the dataset avl queue. Updates to the * persistent zap-object-as-queue happen only at checkpoints. */ while ((sds = avl_first(&scn->scn_queue)) != NULL) { dsl_dataset_t *ds; uint64_t dsobj = sds->sds_dsobj; uint64_t txg = sds->sds_txg; /* dequeue and free the ds from the queue */ scan_ds_queue_remove(scn, dsobj); sds = NULL; /* set up min / max txg */ VERIFY3U(0, ==, dsl_dataset_hold_obj(dp, dsobj, FTAG, &ds)); if (txg != 0) { scn->scn_phys.scn_cur_min_txg = MAX(scn->scn_phys.scn_min_txg, txg); } else { scn->scn_phys.scn_cur_min_txg = MAX(scn->scn_phys.scn_min_txg, dsl_dataset_phys(ds)->ds_prev_snap_txg); } scn->scn_phys.scn_cur_max_txg = dsl_scan_ds_maxtxg(ds); dsl_dataset_rele(ds, FTAG); dsl_scan_visitds(scn, dsobj, tx); if (scn->scn_suspending) return; } /* No more objsets to fetch, we're done */ scn->scn_phys.scn_bookmark.zb_objset = ZB_DESTROYED_OBJSET; ASSERT0(scn->scn_suspending); } static uint64_t dsl_scan_count_data_disks(vdev_t *rvd) { uint64_t i, leaves = 0; for (i = 0; i < rvd->vdev_children; i++) { vdev_t *vd = rvd->vdev_child[i]; if (vd->vdev_islog || vd->vdev_isspare || vd->vdev_isl2cache) continue; leaves += vdev_get_ndisks(vd) - vdev_get_nparity(vd); } return (leaves); } static void scan_io_queues_update_zio_stats(dsl_scan_io_queue_t *q, const blkptr_t *bp) { int i; uint64_t cur_size = 0; for (i = 0; i < BP_GET_NDVAS(bp); i++) { cur_size += DVA_GET_ASIZE(&bp->blk_dva[i]); } q->q_total_zio_size_this_txg += cur_size; q->q_zios_this_txg++; } static void scan_io_queues_update_seg_stats(dsl_scan_io_queue_t *q, uint64_t start, uint64_t end) { q->q_total_seg_size_this_txg += end - start; q->q_segs_this_txg++; } static boolean_t scan_io_queue_check_suspend(dsl_scan_t *scn) { /* See comment in dsl_scan_check_suspend() */ uint64_t curr_time_ns = gethrtime(); uint64_t scan_time_ns = curr_time_ns - scn->scn_sync_start_time; uint64_t sync_time_ns = curr_time_ns - scn->scn_dp->dp_spa->spa_sync_starttime; uint64_t dirty_min_bytes = zfs_dirty_data_max * zfs_vdev_async_write_active_min_dirty_percent / 100; int mintime = (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) ? zfs_resilver_min_time_ms : zfs_scrub_min_time_ms; return ((NSEC2MSEC(scan_time_ns) > mintime && (scn->scn_dp->dp_dirty_total >= dirty_min_bytes || txg_sync_waiting(scn->scn_dp) || NSEC2SEC(sync_time_ns) >= zfs_txg_timeout)) || spa_shutting_down(scn->scn_dp->dp_spa)); } /* * Given a list of scan_io_t's in io_list, this issues the I/Os out to * disk. This consumes the io_list and frees the scan_io_t's. This is * called when emptying queues, either when we're up against the memory * limit or when we have finished scanning. Returns B_TRUE if we stopped * processing the list before we finished. Any sios that were not issued * will remain in the io_list. */ static boolean_t scan_io_queue_issue(dsl_scan_io_queue_t *queue, list_t *io_list) { dsl_scan_t *scn = queue->q_scn; scan_io_t *sio; boolean_t suspended = B_FALSE; while ((sio = list_head(io_list)) != NULL) { blkptr_t bp; if (scan_io_queue_check_suspend(scn)) { suspended = B_TRUE; break; } sio2bp(sio, &bp); scan_exec_io(scn->scn_dp, &bp, sio->sio_flags, &sio->sio_zb, queue); (void) list_remove_head(io_list); scan_io_queues_update_zio_stats(queue, &bp); sio_free(sio); } return (suspended); } /* * This function removes sios from an IO queue which reside within a given * range_seg_t and inserts them (in offset order) into a list. Note that * we only ever return a maximum of 32 sios at once. If there are more sios * to process within this segment that did not make it onto the list we * return B_TRUE and otherwise B_FALSE. */ static boolean_t scan_io_queue_gather(dsl_scan_io_queue_t *queue, range_seg_t *rs, list_t *list) { scan_io_t *srch_sio, *sio, *next_sio; avl_index_t idx; uint_t num_sios = 0; int64_t bytes_issued = 0; ASSERT(rs != NULL); ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); srch_sio = sio_alloc(1); srch_sio->sio_nr_dvas = 1; SIO_SET_OFFSET(srch_sio, rs_get_start(rs, queue->q_exts_by_addr)); /* * The exact start of the extent might not contain any matching zios, * so if that's the case, examine the next one in the tree. */ sio = avl_find(&queue->q_sios_by_addr, srch_sio, &idx); sio_free(srch_sio); if (sio == NULL) sio = avl_nearest(&queue->q_sios_by_addr, idx, AVL_AFTER); while (sio != NULL && SIO_GET_OFFSET(sio) < rs_get_end(rs, queue->q_exts_by_addr) && num_sios <= 32) { ASSERT3U(SIO_GET_OFFSET(sio), >=, rs_get_start(rs, queue->q_exts_by_addr)); ASSERT3U(SIO_GET_END_OFFSET(sio), <=, rs_get_end(rs, queue->q_exts_by_addr)); next_sio = AVL_NEXT(&queue->q_sios_by_addr, sio); avl_remove(&queue->q_sios_by_addr, sio); if (avl_is_empty(&queue->q_sios_by_addr)) atomic_add_64(&queue->q_scn->scn_queues_pending, -1); queue->q_sio_memused -= SIO_GET_MUSED(sio); bytes_issued += SIO_GET_ASIZE(sio); num_sios++; list_insert_tail(list, sio); sio = next_sio; } /* * We limit the number of sios we process at once to 32 to avoid * biting off more than we can chew. If we didn't take everything * in the segment we update it to reflect the work we were able to * complete. Otherwise, we remove it from the range tree entirely. */ if (sio != NULL && SIO_GET_OFFSET(sio) < rs_get_end(rs, queue->q_exts_by_addr)) { range_tree_adjust_fill(queue->q_exts_by_addr, rs, -bytes_issued); range_tree_resize_segment(queue->q_exts_by_addr, rs, SIO_GET_OFFSET(sio), rs_get_end(rs, queue->q_exts_by_addr) - SIO_GET_OFFSET(sio)); queue->q_last_ext_addr = SIO_GET_OFFSET(sio); return (B_TRUE); } else { uint64_t rstart = rs_get_start(rs, queue->q_exts_by_addr); uint64_t rend = rs_get_end(rs, queue->q_exts_by_addr); range_tree_remove(queue->q_exts_by_addr, rstart, rend - rstart); queue->q_last_ext_addr = -1; return (B_FALSE); } } /* * This is called from the queue emptying thread and selects the next * extent from which we are to issue I/Os. The behavior of this function * depends on the state of the scan, the current memory consumption and * whether or not we are performing a scan shutdown. * 1) We select extents in an elevator algorithm (LBA-order) if the scan * needs to perform a checkpoint * 2) We select the largest available extent if we are up against the * memory limit. * 3) Otherwise we don't select any extents. */ static range_seg_t * scan_io_queue_fetch_ext(dsl_scan_io_queue_t *queue) { dsl_scan_t *scn = queue->q_scn; range_tree_t *rt = queue->q_exts_by_addr; ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); ASSERT(scn->scn_is_sorted); if (!scn->scn_checkpointing && !scn->scn_clearing) return (NULL); /* * During normal clearing, we want to issue our largest segments * first, keeping IO as sequential as possible, and leaving the * smaller extents for later with the hope that they might eventually * grow to larger sequential segments. However, when the scan is * checkpointing, no new extents will be added to the sorting queue, * so the way we are sorted now is as good as it will ever get. * In this case, we instead switch to issuing extents in LBA order. */ if ((zfs_scan_issue_strategy < 1 && scn->scn_checkpointing) || zfs_scan_issue_strategy == 1) return (range_tree_first(rt)); /* * Try to continue previous extent if it is not completed yet. After * shrink in scan_io_queue_gather() it may no longer be the best, but * otherwise we leave shorter remnant every txg. */ uint64_t start; uint64_t size = 1 << rt->rt_shift; range_seg_t *addr_rs; if (queue->q_last_ext_addr != -1) { start = queue->q_last_ext_addr; addr_rs = range_tree_find(rt, start, size); if (addr_rs != NULL) return (addr_rs); } /* * Nothing to continue, so find new best extent. */ uint64_t *v = zfs_btree_first(&queue->q_exts_by_size, NULL); if (v == NULL) return (NULL); queue->q_last_ext_addr = start = *v << rt->rt_shift; /* * We need to get the original entry in the by_addr tree so we can * modify it. */ addr_rs = range_tree_find(rt, start, size); ASSERT3P(addr_rs, !=, NULL); ASSERT3U(rs_get_start(addr_rs, rt), ==, start); ASSERT3U(rs_get_end(addr_rs, rt), >, start); return (addr_rs); } static void scan_io_queues_run_one(void *arg) { dsl_scan_io_queue_t *queue = arg; kmutex_t *q_lock = &queue->q_vd->vdev_scan_io_queue_lock; boolean_t suspended = B_FALSE; range_seg_t *rs; scan_io_t *sio; zio_t *zio; list_t sio_list; ASSERT(queue->q_scn->scn_is_sorted); list_create(&sio_list, sizeof (scan_io_t), offsetof(scan_io_t, sio_nodes.sio_list_node)); zio = zio_null(queue->q_scn->scn_zio_root, queue->q_scn->scn_dp->dp_spa, NULL, NULL, NULL, ZIO_FLAG_CANFAIL); mutex_enter(q_lock); queue->q_zio = zio; /* Calculate maximum in-flight bytes for this vdev. */ queue->q_maxinflight_bytes = MAX(1, zfs_scan_vdev_limit * (vdev_get_ndisks(queue->q_vd) - vdev_get_nparity(queue->q_vd))); /* reset per-queue scan statistics for this txg */ queue->q_total_seg_size_this_txg = 0; queue->q_segs_this_txg = 0; queue->q_total_zio_size_this_txg = 0; queue->q_zios_this_txg = 0; /* loop until we run out of time or sios */ while ((rs = scan_io_queue_fetch_ext(queue)) != NULL) { uint64_t seg_start = 0, seg_end = 0; boolean_t more_left; ASSERT(list_is_empty(&sio_list)); /* loop while we still have sios left to process in this rs */ do { scan_io_t *first_sio, *last_sio; /* * We have selected which extent needs to be * processed next. Gather up the corresponding sios. */ more_left = scan_io_queue_gather(queue, rs, &sio_list); ASSERT(!list_is_empty(&sio_list)); first_sio = list_head(&sio_list); last_sio = list_tail(&sio_list); seg_end = SIO_GET_END_OFFSET(last_sio); if (seg_start == 0) seg_start = SIO_GET_OFFSET(first_sio); /* * Issuing sios can take a long time so drop the * queue lock. The sio queue won't be updated by * other threads since we're in syncing context so * we can be sure that our trees will remain exactly * as we left them. */ mutex_exit(q_lock); suspended = scan_io_queue_issue(queue, &sio_list); mutex_enter(q_lock); if (suspended) break; } while (more_left); /* update statistics for debugging purposes */ scan_io_queues_update_seg_stats(queue, seg_start, seg_end); if (suspended) break; } /* * If we were suspended in the middle of processing, * requeue any unfinished sios and exit. */ while ((sio = list_head(&sio_list)) != NULL) { list_remove(&sio_list, sio); scan_io_queue_insert_impl(queue, sio); } queue->q_zio = NULL; mutex_exit(q_lock); zio_nowait(zio); list_destroy(&sio_list); } /* * Performs an emptying run on all scan queues in the pool. This just * punches out one thread per top-level vdev, each of which processes * only that vdev's scan queue. We can parallelize the I/O here because * we know that each queue's I/Os only affect its own top-level vdev. * * This function waits for the queue runs to complete, and must be * called from dsl_scan_sync (or in general, syncing context). */ static void scan_io_queues_run(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; ASSERT(scn->scn_is_sorted); ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER)); if (scn->scn_queues_pending == 0) return; if (scn->scn_taskq == NULL) { int nthreads = spa->spa_root_vdev->vdev_children; /* * We need to make this taskq *always* execute as many * threads in parallel as we have top-level vdevs and no * less, otherwise strange serialization of the calls to * scan_io_queues_run_one can occur during spa_sync runs * and that significantly impacts performance. */ scn->scn_taskq = taskq_create("dsl_scan_iss", nthreads, minclsyspri, nthreads, nthreads, TASKQ_PREPOPULATE); } for (uint64_t i = 0; i < spa->spa_root_vdev->vdev_children; i++) { vdev_t *vd = spa->spa_root_vdev->vdev_child[i]; mutex_enter(&vd->vdev_scan_io_queue_lock); if (vd->vdev_scan_io_queue != NULL) { VERIFY(taskq_dispatch(scn->scn_taskq, scan_io_queues_run_one, vd->vdev_scan_io_queue, TQ_SLEEP) != TASKQID_INVALID); } mutex_exit(&vd->vdev_scan_io_queue_lock); } /* * Wait for the queues to finish issuing their IOs for this run * before we return. There may still be IOs in flight at this * point. */ taskq_wait(scn->scn_taskq); } static boolean_t dsl_scan_async_block_should_pause(dsl_scan_t *scn) { uint64_t elapsed_nanosecs; if (zfs_recover) return (B_FALSE); if (zfs_async_block_max_blocks != 0 && scn->scn_visited_this_txg >= zfs_async_block_max_blocks) { return (B_TRUE); } if (zfs_max_async_dedup_frees != 0 && scn->scn_dedup_frees_this_txg >= zfs_max_async_dedup_frees) { return (B_TRUE); } elapsed_nanosecs = gethrtime() - scn->scn_sync_start_time; return (elapsed_nanosecs / NANOSEC > zfs_txg_timeout || (NSEC2MSEC(elapsed_nanosecs) > scn->scn_async_block_min_time_ms && txg_sync_waiting(scn->scn_dp)) || spa_shutting_down(scn->scn_dp->dp_spa)); } static int dsl_scan_free_block_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { dsl_scan_t *scn = arg; if (!scn->scn_is_bptree || (BP_GET_LEVEL(bp) == 0 && BP_GET_TYPE(bp) != DMU_OT_OBJSET)) { if (dsl_scan_async_block_should_pause(scn)) return (SET_ERROR(ERESTART)); } zio_nowait(zio_free_sync(scn->scn_zio_root, scn->scn_dp->dp_spa, dmu_tx_get_txg(tx), bp, 0)); dsl_dir_diduse_space(tx->tx_pool->dp_free_dir, DD_USED_HEAD, -bp_get_dsize_sync(scn->scn_dp->dp_spa, bp), -BP_GET_PSIZE(bp), -BP_GET_UCSIZE(bp), tx); scn->scn_visited_this_txg++; if (BP_GET_DEDUP(bp)) scn->scn_dedup_frees_this_txg++; return (0); } static void dsl_scan_update_stats(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; uint64_t i; uint64_t seg_size_total = 0, zio_size_total = 0; uint64_t seg_count_total = 0, zio_count_total = 0; for (i = 0; i < spa->spa_root_vdev->vdev_children; i++) { vdev_t *vd = spa->spa_root_vdev->vdev_child[i]; dsl_scan_io_queue_t *queue = vd->vdev_scan_io_queue; if (queue == NULL) continue; seg_size_total += queue->q_total_seg_size_this_txg; zio_size_total += queue->q_total_zio_size_this_txg; seg_count_total += queue->q_segs_this_txg; zio_count_total += queue->q_zios_this_txg; } if (seg_count_total == 0 || zio_count_total == 0) { scn->scn_avg_seg_size_this_txg = 0; scn->scn_avg_zio_size_this_txg = 0; scn->scn_segs_this_txg = 0; scn->scn_zios_this_txg = 0; return; } scn->scn_avg_seg_size_this_txg = seg_size_total / seg_count_total; scn->scn_avg_zio_size_this_txg = zio_size_total / zio_count_total; scn->scn_segs_this_txg = seg_count_total; scn->scn_zios_this_txg = zio_count_total; } static int bpobj_dsl_scan_free_block_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed, dmu_tx_t *tx) { ASSERT(!bp_freed); return (dsl_scan_free_block_cb(arg, bp, tx)); } static int dsl_scan_obsolete_block_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed, dmu_tx_t *tx) { ASSERT(!bp_freed); dsl_scan_t *scn = arg; const dva_t *dva = &bp->blk_dva[0]; if (dsl_scan_async_block_should_pause(scn)) return (SET_ERROR(ERESTART)); spa_vdev_indirect_mark_obsolete(scn->scn_dp->dp_spa, DVA_GET_VDEV(dva), DVA_GET_OFFSET(dva), DVA_GET_ASIZE(dva), tx); scn->scn_visited_this_txg++; return (0); } boolean_t dsl_scan_active(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; uint64_t used = 0, comp, uncomp; boolean_t clones_left; if (spa->spa_load_state != SPA_LOAD_NONE) return (B_FALSE); if (spa_shutting_down(spa)) return (B_FALSE); if ((dsl_scan_is_running(scn) && !dsl_scan_is_paused_scrub(scn)) || (scn->scn_async_destroying && !scn->scn_async_stalled)) return (B_TRUE); if (spa_version(scn->scn_dp->dp_spa) >= SPA_VERSION_DEADLISTS) { (void) bpobj_space(&scn->scn_dp->dp_free_bpobj, &used, &comp, &uncomp); } clones_left = spa_livelist_delete_check(spa); return ((used != 0) || (clones_left)); } static boolean_t dsl_scan_check_deferred(vdev_t *vd) { boolean_t need_resilver = B_FALSE; for (int c = 0; c < vd->vdev_children; c++) { need_resilver |= dsl_scan_check_deferred(vd->vdev_child[c]); } if (!vdev_is_concrete(vd) || vd->vdev_aux || !vd->vdev_ops->vdev_op_leaf) return (need_resilver); if (!vd->vdev_resilver_deferred) need_resilver = B_TRUE; return (need_resilver); } static boolean_t dsl_scan_need_resilver(spa_t *spa, const dva_t *dva, size_t psize, uint64_t phys_birth) { vdev_t *vd; vd = vdev_lookup_top(spa, DVA_GET_VDEV(dva)); if (vd->vdev_ops == &vdev_indirect_ops) { /* * The indirect vdev can point to multiple * vdevs. For simplicity, always create * the resilver zio_t. zio_vdev_io_start() * will bypass the child resilver i/o's if * they are on vdevs that don't have DTL's. */ return (B_TRUE); } if (DVA_GET_GANG(dva)) { /* * Gang members may be spread across multiple * vdevs, so the best estimate we have is the * scrub range, which has already been checked. * XXX -- it would be better to change our * allocation policy to ensure that all * gang members reside on the same vdev. */ return (B_TRUE); } /* * Check if the top-level vdev must resilver this offset. * When the offset does not intersect with a dirty leaf DTL * then it may be possible to skip the resilver IO. The psize * is provided instead of asize to simplify the check for RAIDZ. */ if (!vdev_dtl_need_resilver(vd, dva, psize, phys_birth)) return (B_FALSE); /* * Check that this top-level vdev has a device under it which * is resilvering and is not deferred. */ if (!dsl_scan_check_deferred(vd)) return (B_FALSE); return (B_TRUE); } static int dsl_process_async_destroys(dsl_pool_t *dp, dmu_tx_t *tx) { dsl_scan_t *scn = dp->dp_scan; spa_t *spa = dp->dp_spa; int err = 0; if (spa_suspend_async_destroy(spa)) return (0); if (zfs_free_bpobj_enabled && spa_version(spa) >= SPA_VERSION_DEADLISTS) { scn->scn_is_bptree = B_FALSE; scn->scn_async_block_min_time_ms = zfs_free_min_time_ms; scn->scn_zio_root = zio_root(spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED); err = bpobj_iterate(&dp->dp_free_bpobj, bpobj_dsl_scan_free_block_cb, scn, tx); VERIFY0(zio_wait(scn->scn_zio_root)); scn->scn_zio_root = NULL; if (err != 0 && err != ERESTART) zfs_panic_recover("error %u from bpobj_iterate()", err); } if (err == 0 && spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY)) { ASSERT(scn->scn_async_destroying); scn->scn_is_bptree = B_TRUE; scn->scn_zio_root = zio_root(spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED); err = bptree_iterate(dp->dp_meta_objset, dp->dp_bptree_obj, B_TRUE, dsl_scan_free_block_cb, scn, tx); VERIFY0(zio_wait(scn->scn_zio_root)); scn->scn_zio_root = NULL; if (err == EIO || err == ECKSUM) { err = 0; } else if (err != 0 && err != ERESTART) { zfs_panic_recover("error %u from " "traverse_dataset_destroyed()", err); } if (bptree_is_empty(dp->dp_meta_objset, dp->dp_bptree_obj)) { /* finished; deactivate async destroy feature */ spa_feature_decr(spa, SPA_FEATURE_ASYNC_DESTROY, tx); ASSERT(!spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY)); VERIFY0(zap_remove(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_BPTREE_OBJ, tx)); VERIFY0(bptree_free(dp->dp_meta_objset, dp->dp_bptree_obj, tx)); dp->dp_bptree_obj = 0; scn->scn_async_destroying = B_FALSE; scn->scn_async_stalled = B_FALSE; } else { /* * If we didn't make progress, mark the async * destroy as stalled, so that we will not initiate * a spa_sync() on its behalf. Note that we only * check this if we are not finished, because if the * bptree had no blocks for us to visit, we can * finish without "making progress". */ scn->scn_async_stalled = (scn->scn_visited_this_txg == 0); } } if (scn->scn_visited_this_txg) { zfs_dbgmsg("freed %llu blocks in %llums from " "free_bpobj/bptree txg %llu; err=%u", (longlong_t)scn->scn_visited_this_txg, (longlong_t) NSEC2MSEC(gethrtime() - scn->scn_sync_start_time), (longlong_t)tx->tx_txg, err); scn->scn_visited_this_txg = 0; scn->scn_dedup_frees_this_txg = 0; /* * Write out changes to the DDT that may be required as a * result of the blocks freed. This ensures that the DDT * is clean when a scrub/resilver runs. */ ddt_sync(spa, tx->tx_txg); } if (err != 0) return (err); if (dp->dp_free_dir != NULL && !scn->scn_async_destroying && zfs_free_leak_on_eio && (dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes != 0 || dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes != 0 || dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes != 0)) { /* * We have finished background destroying, but there is still * some space left in the dp_free_dir. Transfer this leaked * space to the dp_leak_dir. */ if (dp->dp_leak_dir == NULL) { rrw_enter(&dp->dp_config_rwlock, RW_WRITER, FTAG); (void) dsl_dir_create_sync(dp, dp->dp_root_dir, LEAK_DIR_NAME, tx); VERIFY0(dsl_pool_open_special_dir(dp, LEAK_DIR_NAME, &dp->dp_leak_dir)); rrw_exit(&dp->dp_config_rwlock, FTAG); } dsl_dir_diduse_space(dp->dp_leak_dir, DD_USED_HEAD, dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes, dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes, dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes, tx); dsl_dir_diduse_space(dp->dp_free_dir, DD_USED_HEAD, -dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes, -dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes, -dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes, tx); } if (dp->dp_free_dir != NULL && !scn->scn_async_destroying && !spa_livelist_delete_check(spa)) { /* finished; verify that space accounting went to zero */ ASSERT0(dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes); ASSERT0(dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes); ASSERT0(dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes); } spa_notify_waiters(spa); EQUIV(bpobj_is_open(&dp->dp_obsolete_bpobj), 0 == zap_contains(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_OBSOLETE_BPOBJ)); if (err == 0 && bpobj_is_open(&dp->dp_obsolete_bpobj)) { ASSERT(spa_feature_is_active(dp->dp_spa, SPA_FEATURE_OBSOLETE_COUNTS)); scn->scn_is_bptree = B_FALSE; scn->scn_async_block_min_time_ms = zfs_obsolete_min_time_ms; err = bpobj_iterate(&dp->dp_obsolete_bpobj, dsl_scan_obsolete_block_cb, scn, tx); if (err != 0 && err != ERESTART) zfs_panic_recover("error %u from bpobj_iterate()", err); if (bpobj_is_empty(&dp->dp_obsolete_bpobj)) dsl_pool_destroy_obsolete_bpobj(dp, tx); } return (0); } /* * This is the primary entry point for scans that is called from syncing * context. Scans must happen entirely during syncing context so that we * can guarantee that blocks we are currently scanning will not change out * from under us. While a scan is active, this function controls how quickly * transaction groups proceed, instead of the normal handling provided by * txg_sync_thread(). */ void dsl_scan_sync(dsl_pool_t *dp, dmu_tx_t *tx) { int err = 0; dsl_scan_t *scn = dp->dp_scan; spa_t *spa = dp->dp_spa; state_sync_type_t sync_type = SYNC_OPTIONAL; if (spa->spa_resilver_deferred && !spa_feature_is_active(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER)) spa_feature_incr(spa, SPA_FEATURE_RESILVER_DEFER, tx); /* * Check for scn_restart_txg before checking spa_load_state, so * that we can restart an old-style scan while the pool is being * imported (see dsl_scan_init). We also restart scans if there * is a deferred resilver and the user has manually disabled * deferred resilvers via the tunable. */ if (dsl_scan_restarting(scn, tx) || (spa->spa_resilver_deferred && zfs_resilver_disable_defer)) { pool_scan_func_t func = POOL_SCAN_SCRUB; dsl_scan_done(scn, B_FALSE, tx); if (vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) func = POOL_SCAN_RESILVER; zfs_dbgmsg("restarting scan func=%u txg=%llu", func, (longlong_t)tx->tx_txg); dsl_scan_setup_sync(&func, tx); } /* * Only process scans in sync pass 1. */ if (spa_sync_pass(spa) > 1) return; /* * If the spa is shutting down, then stop scanning. This will * ensure that the scan does not dirty any new data during the * shutdown phase. */ if (spa_shutting_down(spa)) return; /* * If the scan is inactive due to a stalled async destroy, try again. */ if (!scn->scn_async_stalled && !dsl_scan_active(scn)) return; /* reset scan statistics */ scn->scn_visited_this_txg = 0; scn->scn_dedup_frees_this_txg = 0; scn->scn_holes_this_txg = 0; scn->scn_lt_min_this_txg = 0; scn->scn_gt_max_this_txg = 0; scn->scn_ddt_contained_this_txg = 0; scn->scn_objsets_visited_this_txg = 0; scn->scn_avg_seg_size_this_txg = 0; scn->scn_segs_this_txg = 0; scn->scn_avg_zio_size_this_txg = 0; scn->scn_zios_this_txg = 0; scn->scn_suspending = B_FALSE; scn->scn_sync_start_time = gethrtime(); spa->spa_scrub_active = B_TRUE; /* * First process the async destroys. If we suspend, don't do * any scrubbing or resilvering. This ensures that there are no * async destroys while we are scanning, so the scan code doesn't * have to worry about traversing it. It is also faster to free the * blocks than to scrub them. */ err = dsl_process_async_destroys(dp, tx); if (err != 0) return; if (!dsl_scan_is_running(scn) || dsl_scan_is_paused_scrub(scn)) return; /* * Wait a few txgs after importing to begin scanning so that * we can get the pool imported quickly. */ if (spa->spa_syncing_txg < spa->spa_first_txg + SCAN_IMPORT_WAIT_TXGS) return; /* * zfs_scan_suspend_progress can be set to disable scan progress. * We don't want to spin the txg_sync thread, so we add a delay * here to simulate the time spent doing a scan. This is mostly * useful for testing and debugging. */ if (zfs_scan_suspend_progress) { uint64_t scan_time_ns = gethrtime() - scn->scn_sync_start_time; int mintime = (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) ? zfs_resilver_min_time_ms : zfs_scrub_min_time_ms; while (zfs_scan_suspend_progress && !txg_sync_waiting(scn->scn_dp) && !spa_shutting_down(scn->scn_dp->dp_spa) && NSEC2MSEC(scan_time_ns) < mintime) { delay(hz); scan_time_ns = gethrtime() - scn->scn_sync_start_time; } return; } + /* + * Disabled by default, set zfs_scan_report_txgs to report + * average performance over the last zfs_scan_report_txgs TXGs. + */ + if (!dsl_scan_is_paused_scrub(scn) && zfs_scan_report_txgs != 0 && + tx->tx_txg % zfs_scan_report_txgs == 0) { + scn->scn_issued_before_pass += spa->spa_scan_pass_issued; + spa_scan_stat_init(spa); + } + /* * It is possible to switch from unsorted to sorted at any time, * but afterwards the scan will remain sorted unless reloaded from * a checkpoint after a reboot. */ if (!zfs_scan_legacy) { scn->scn_is_sorted = B_TRUE; if (scn->scn_last_checkpoint == 0) scn->scn_last_checkpoint = ddi_get_lbolt(); } /* * For sorted scans, determine what kind of work we will be doing * this txg based on our memory limitations and whether or not we * need to perform a checkpoint. */ if (scn->scn_is_sorted) { /* * If we are over our checkpoint interval, set scn_clearing * so that we can begin checkpointing immediately. The * checkpoint allows us to save a consistent bookmark * representing how much data we have scrubbed so far. * Otherwise, use the memory limit to determine if we should * scan for metadata or start issue scrub IOs. We accumulate * metadata until we hit our hard memory limit at which point * we issue scrub IOs until we are at our soft memory limit. */ if (scn->scn_checkpointing || ddi_get_lbolt() - scn->scn_last_checkpoint > SEC_TO_TICK(zfs_scan_checkpoint_intval)) { if (!scn->scn_checkpointing) zfs_dbgmsg("begin scan checkpoint"); scn->scn_checkpointing = B_TRUE; scn->scn_clearing = B_TRUE; } else { boolean_t should_clear = dsl_scan_should_clear(scn); if (should_clear && !scn->scn_clearing) { zfs_dbgmsg("begin scan clearing"); scn->scn_clearing = B_TRUE; } else if (!should_clear && scn->scn_clearing) { zfs_dbgmsg("finish scan clearing"); scn->scn_clearing = B_FALSE; } } } else { ASSERT0(scn->scn_checkpointing); ASSERT0(scn->scn_clearing); } if (!scn->scn_clearing && scn->scn_done_txg == 0) { /* Need to scan metadata for more blocks to scrub */ dsl_scan_phys_t *scnp = &scn->scn_phys; taskqid_t prefetch_tqid; /* * Recalculate the max number of in-flight bytes for pool-wide * scanning operations (minimum 1MB). Limits for the issuing * phase are done per top-level vdev and are handled separately. */ scn->scn_maxinflight_bytes = MAX(zfs_scan_vdev_limit * dsl_scan_count_data_disks(spa->spa_root_vdev), 1ULL << 20); if (scnp->scn_ddt_bookmark.ddb_class <= scnp->scn_ddt_class_max) { ASSERT(ZB_IS_ZERO(&scnp->scn_bookmark)); zfs_dbgmsg("doing scan sync txg %llu; " "ddt bm=%llu/%llu/%llu/%llx", (longlong_t)tx->tx_txg, (longlong_t)scnp->scn_ddt_bookmark.ddb_class, (longlong_t)scnp->scn_ddt_bookmark.ddb_type, (longlong_t)scnp->scn_ddt_bookmark.ddb_checksum, (longlong_t)scnp->scn_ddt_bookmark.ddb_cursor); } else { zfs_dbgmsg("doing scan sync txg %llu; " "bm=%llu/%llu/%llu/%llu", (longlong_t)tx->tx_txg, (longlong_t)scnp->scn_bookmark.zb_objset, (longlong_t)scnp->scn_bookmark.zb_object, (longlong_t)scnp->scn_bookmark.zb_level, (longlong_t)scnp->scn_bookmark.zb_blkid); } scn->scn_zio_root = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_CANFAIL); scn->scn_prefetch_stop = B_FALSE; prefetch_tqid = taskq_dispatch(dp->dp_sync_taskq, dsl_scan_prefetch_thread, scn, TQ_SLEEP); ASSERT(prefetch_tqid != TASKQID_INVALID); dsl_pool_config_enter(dp, FTAG); dsl_scan_visit(scn, tx); dsl_pool_config_exit(dp, FTAG); mutex_enter(&dp->dp_spa->spa_scrub_lock); scn->scn_prefetch_stop = B_TRUE; cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&dp->dp_spa->spa_scrub_lock); taskq_wait_id(dp->dp_sync_taskq, prefetch_tqid); (void) zio_wait(scn->scn_zio_root); scn->scn_zio_root = NULL; zfs_dbgmsg("scan visited %llu blocks in %llums " "(%llu os's, %llu holes, %llu < mintxg, " "%llu in ddt, %llu > maxtxg)", (longlong_t)scn->scn_visited_this_txg, (longlong_t)NSEC2MSEC(gethrtime() - scn->scn_sync_start_time), (longlong_t)scn->scn_objsets_visited_this_txg, (longlong_t)scn->scn_holes_this_txg, (longlong_t)scn->scn_lt_min_this_txg, (longlong_t)scn->scn_ddt_contained_this_txg, (longlong_t)scn->scn_gt_max_this_txg); if (!scn->scn_suspending) { ASSERT0(avl_numnodes(&scn->scn_queue)); scn->scn_done_txg = tx->tx_txg + 1; if (scn->scn_is_sorted) { scn->scn_checkpointing = B_TRUE; scn->scn_clearing = B_TRUE; + scn->scn_issued_before_pass += + spa->spa_scan_pass_issued; + spa_scan_stat_init(spa); } zfs_dbgmsg("scan complete txg %llu", (longlong_t)tx->tx_txg); } } else if (scn->scn_is_sorted && scn->scn_queues_pending != 0) { ASSERT(scn->scn_clearing); /* need to issue scrubbing IOs from per-vdev queues */ scn->scn_zio_root = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_CANFAIL); scan_io_queues_run(scn); (void) zio_wait(scn->scn_zio_root); scn->scn_zio_root = NULL; /* calculate and dprintf the current memory usage */ (void) dsl_scan_should_clear(scn); dsl_scan_update_stats(scn); zfs_dbgmsg("scan issued %llu blocks (%llu segs) in %llums " "(avg_block_size = %llu, avg_seg_size = %llu)", (longlong_t)scn->scn_zios_this_txg, (longlong_t)scn->scn_segs_this_txg, (longlong_t)NSEC2MSEC(gethrtime() - scn->scn_sync_start_time), (longlong_t)scn->scn_avg_zio_size_this_txg, (longlong_t)scn->scn_avg_seg_size_this_txg); } else if (scn->scn_done_txg != 0 && scn->scn_done_txg <= tx->tx_txg) { /* Finished with everything. Mark the scrub as complete */ zfs_dbgmsg("scan issuing complete txg %llu", (longlong_t)tx->tx_txg); ASSERT3U(scn->scn_done_txg, !=, 0); ASSERT0(spa->spa_scrub_inflight); ASSERT0(scn->scn_queues_pending); dsl_scan_done(scn, B_TRUE, tx); sync_type = SYNC_MANDATORY; } dsl_scan_sync_state(scn, tx, sync_type); } static void count_block_issued(spa_t *spa, const blkptr_t *bp, boolean_t all) { /* * Don't count embedded bp's, since we already did the work of * scanning these when we scanned the containing block. */ if (BP_IS_EMBEDDED(bp)) return; /* * Update the spa's stats on how many bytes we have issued. * Sequential scrubs create a zio for each DVA of the bp. Each * of these will include all DVAs for repair purposes, but the * zio code will only try the first one unless there is an issue. * Therefore, we should only count the first DVA for these IOs. */ atomic_add_64(&spa->spa_scan_pass_issued, all ? BP_GET_ASIZE(bp) : DVA_GET_ASIZE(&bp->blk_dva[0])); } static void count_block(zfs_all_blkstats_t *zab, const blkptr_t *bp) { /* * If we resume after a reboot, zab will be NULL; don't record * incomplete stats in that case. */ if (zab == NULL) return; for (int i = 0; i < 4; i++) { int l = (i < 2) ? BP_GET_LEVEL(bp) : DN_MAX_LEVELS; int t = (i & 1) ? BP_GET_TYPE(bp) : DMU_OT_TOTAL; if (t & DMU_OT_NEWTYPE) t = DMU_OT_OTHER; zfs_blkstat_t *zb = &zab->zab_type[l][t]; int equal; zb->zb_count++; zb->zb_asize += BP_GET_ASIZE(bp); zb->zb_lsize += BP_GET_LSIZE(bp); zb->zb_psize += BP_GET_PSIZE(bp); zb->zb_gangs += BP_COUNT_GANG(bp); switch (BP_GET_NDVAS(bp)) { case 2: if (DVA_GET_VDEV(&bp->blk_dva[0]) == DVA_GET_VDEV(&bp->blk_dva[1])) zb->zb_ditto_2_of_2_samevdev++; break; case 3: equal = (DVA_GET_VDEV(&bp->blk_dva[0]) == DVA_GET_VDEV(&bp->blk_dva[1])) + (DVA_GET_VDEV(&bp->blk_dva[0]) == DVA_GET_VDEV(&bp->blk_dva[2])) + (DVA_GET_VDEV(&bp->blk_dva[1]) == DVA_GET_VDEV(&bp->blk_dva[2])); if (equal == 1) zb->zb_ditto_2_of_3_samevdev++; else if (equal == 3) zb->zb_ditto_3_of_3_samevdev++; break; } } } static void scan_io_queue_insert_impl(dsl_scan_io_queue_t *queue, scan_io_t *sio) { avl_index_t idx; dsl_scan_t *scn = queue->q_scn; ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); if (unlikely(avl_is_empty(&queue->q_sios_by_addr))) atomic_add_64(&scn->scn_queues_pending, 1); if (avl_find(&queue->q_sios_by_addr, sio, &idx) != NULL) { /* block is already scheduled for reading */ sio_free(sio); return; } avl_insert(&queue->q_sios_by_addr, sio, idx); queue->q_sio_memused += SIO_GET_MUSED(sio); range_tree_add(queue->q_exts_by_addr, SIO_GET_OFFSET(sio), SIO_GET_ASIZE(sio)); } /* * Given all the info we got from our metadata scanning process, we * construct a scan_io_t and insert it into the scan sorting queue. The * I/O must already be suitable for us to process. This is controlled * by dsl_scan_enqueue(). */ static void scan_io_queue_insert(dsl_scan_io_queue_t *queue, const blkptr_t *bp, int dva_i, int zio_flags, const zbookmark_phys_t *zb) { scan_io_t *sio = sio_alloc(BP_GET_NDVAS(bp)); ASSERT0(BP_IS_GANG(bp)); ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); bp2sio(bp, sio, dva_i); sio->sio_flags = zio_flags; sio->sio_zb = *zb; queue->q_last_ext_addr = -1; scan_io_queue_insert_impl(queue, sio); } /* * Given a set of I/O parameters as discovered by the metadata traversal * process, attempts to place the I/O into the sorted queues (if allowed), * or immediately executes the I/O. */ static void dsl_scan_enqueue(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags, const zbookmark_phys_t *zb) { spa_t *spa = dp->dp_spa; ASSERT(!BP_IS_EMBEDDED(bp)); /* * Gang blocks are hard to issue sequentially, so we just issue them * here immediately instead of queuing them. */ if (!dp->dp_scan->scn_is_sorted || BP_IS_GANG(bp)) { scan_exec_io(dp, bp, zio_flags, zb, NULL); return; } for (int i = 0; i < BP_GET_NDVAS(bp); i++) { dva_t dva; vdev_t *vdev; dva = bp->blk_dva[i]; vdev = vdev_lookup_top(spa, DVA_GET_VDEV(&dva)); ASSERT(vdev != NULL); mutex_enter(&vdev->vdev_scan_io_queue_lock); if (vdev->vdev_scan_io_queue == NULL) vdev->vdev_scan_io_queue = scan_io_queue_create(vdev); ASSERT(dp->dp_scan != NULL); scan_io_queue_insert(vdev->vdev_scan_io_queue, bp, i, zio_flags, zb); mutex_exit(&vdev->vdev_scan_io_queue_lock); } } static int dsl_scan_scrub_cb(dsl_pool_t *dp, const blkptr_t *bp, const zbookmark_phys_t *zb) { dsl_scan_t *scn = dp->dp_scan; spa_t *spa = dp->dp_spa; uint64_t phys_birth = BP_PHYSICAL_BIRTH(bp); size_t psize = BP_GET_PSIZE(bp); boolean_t needs_io = B_FALSE; int zio_flags = ZIO_FLAG_SCAN_THREAD | ZIO_FLAG_RAW | ZIO_FLAG_CANFAIL; count_block(dp->dp_blkstats, bp); if (phys_birth <= scn->scn_phys.scn_min_txg || phys_birth >= scn->scn_phys.scn_max_txg) { count_block_issued(spa, bp, B_TRUE); return (0); } /* Embedded BP's have phys_birth==0, so we reject them above. */ ASSERT(!BP_IS_EMBEDDED(bp)); ASSERT(DSL_SCAN_IS_SCRUB_RESILVER(scn)); if (scn->scn_phys.scn_func == POOL_SCAN_SCRUB) { zio_flags |= ZIO_FLAG_SCRUB; needs_io = B_TRUE; } else { ASSERT3U(scn->scn_phys.scn_func, ==, POOL_SCAN_RESILVER); zio_flags |= ZIO_FLAG_RESILVER; needs_io = B_FALSE; } /* If it's an intent log block, failure is expected. */ if (zb->zb_level == ZB_ZIL_LEVEL) zio_flags |= ZIO_FLAG_SPECULATIVE; for (int d = 0; d < BP_GET_NDVAS(bp); d++) { const dva_t *dva = &bp->blk_dva[d]; /* * Keep track of how much data we've examined so that * zpool(8) status can make useful progress reports. */ uint64_t asize = DVA_GET_ASIZE(dva); scn->scn_phys.scn_examined += asize; spa->spa_scan_pass_exam += asize; /* if it's a resilver, this may not be in the target range */ if (!needs_io) needs_io = dsl_scan_need_resilver(spa, dva, psize, phys_birth); } if (needs_io && !zfs_no_scrub_io) { dsl_scan_enqueue(dp, bp, zio_flags, zb); } else { count_block_issued(spa, bp, B_TRUE); } /* do not relocate this block */ return (0); } static void dsl_scan_scrub_done(zio_t *zio) { spa_t *spa = zio->io_spa; blkptr_t *bp = zio->io_bp; dsl_scan_io_queue_t *queue = zio->io_private; abd_free(zio->io_abd); if (queue == NULL) { mutex_enter(&spa->spa_scrub_lock); ASSERT3U(spa->spa_scrub_inflight, >=, BP_GET_PSIZE(bp)); spa->spa_scrub_inflight -= BP_GET_PSIZE(bp); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); } else { mutex_enter(&queue->q_vd->vdev_scan_io_queue_lock); ASSERT3U(queue->q_inflight_bytes, >=, BP_GET_PSIZE(bp)); queue->q_inflight_bytes -= BP_GET_PSIZE(bp); cv_broadcast(&queue->q_zio_cv); mutex_exit(&queue->q_vd->vdev_scan_io_queue_lock); } if (zio->io_error && (zio->io_error != ECKSUM || !(zio->io_flags & ZIO_FLAG_SPECULATIVE))) { atomic_inc_64(&spa->spa_dsl_pool->dp_scan->scn_phys.scn_errors); } } /* * Given a scanning zio's information, executes the zio. The zio need * not necessarily be only sortable, this function simply executes the * zio, no matter what it is. The optional queue argument allows the * caller to specify that they want per top level vdev IO rate limiting * instead of the legacy global limiting. */ static void scan_exec_io(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags, const zbookmark_phys_t *zb, dsl_scan_io_queue_t *queue) { spa_t *spa = dp->dp_spa; dsl_scan_t *scn = dp->dp_scan; size_t size = BP_GET_PSIZE(bp); abd_t *data = abd_alloc_for_io(size, B_FALSE); zio_t *pio; if (queue == NULL) { ASSERT3U(scn->scn_maxinflight_bytes, >, 0); mutex_enter(&spa->spa_scrub_lock); while (spa->spa_scrub_inflight >= scn->scn_maxinflight_bytes) cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock); spa->spa_scrub_inflight += BP_GET_PSIZE(bp); mutex_exit(&spa->spa_scrub_lock); pio = scn->scn_zio_root; } else { kmutex_t *q_lock = &queue->q_vd->vdev_scan_io_queue_lock; ASSERT3U(queue->q_maxinflight_bytes, >, 0); mutex_enter(q_lock); while (queue->q_inflight_bytes >= queue->q_maxinflight_bytes) cv_wait(&queue->q_zio_cv, q_lock); queue->q_inflight_bytes += BP_GET_PSIZE(bp); pio = queue->q_zio; mutex_exit(q_lock); } ASSERT(pio != NULL); count_block_issued(spa, bp, queue == NULL); zio_nowait(zio_read(pio, spa, bp, data, size, dsl_scan_scrub_done, queue, ZIO_PRIORITY_SCRUB, zio_flags, zb)); } /* * This is the primary extent sorting algorithm. We balance two parameters: * 1) how many bytes of I/O are in an extent * 2) how well the extent is filled with I/O (as a fraction of its total size) * Since we allow extents to have gaps between their constituent I/Os, it's * possible to have a fairly large extent that contains the same amount of * I/O bytes than a much smaller extent, which just packs the I/O more tightly. * The algorithm sorts based on a score calculated from the extent's size, * the relative fill volume (in %) and a "fill weight" parameter that controls * the split between whether we prefer larger extents or more well populated * extents: * * SCORE = FILL_IN_BYTES + (FILL_IN_PERCENT * FILL_IN_BYTES * FILL_WEIGHT) * * Example: * 1) assume extsz = 64 MiB * 2) assume fill = 32 MiB (extent is half full) * 3) assume fill_weight = 3 * 4) SCORE = 32M + (((32M * 100) / 64M) * 3 * 32M) / 100 * SCORE = 32M + (50 * 3 * 32M) / 100 * SCORE = 32M + (4800M / 100) * SCORE = 32M + 48M * ^ ^ * | +--- final total relative fill-based score * +--------- final total fill-based score * SCORE = 80M * * As can be seen, at fill_ratio=3, the algorithm is slightly biased towards * extents that are more completely filled (in a 3:2 ratio) vs just larger. * Note that as an optimization, we replace multiplication and division by * 100 with bitshifting by 7 (which effectively multiplies and divides by 128). * * Since we do not care if one extent is only few percent better than another, * compress the score into 6 bits via binary logarithm AKA highbit64() and * put into otherwise unused due to ashift high bits of offset. This allows * to reduce q_exts_by_size B-tree elements to only 64 bits and compare them * with single operation. Plus it makes scrubs more sequential and reduces * chances that minor extent change move it within the B-tree. */ static int ext_size_compare(const void *x, const void *y) { const uint64_t *a = x, *b = y; return (TREE_CMP(*a, *b)); } static void ext_size_create(range_tree_t *rt, void *arg) { (void) rt; zfs_btree_t *size_tree = arg; zfs_btree_create(size_tree, ext_size_compare, sizeof (uint64_t)); } static void ext_size_destroy(range_tree_t *rt, void *arg) { (void) rt; zfs_btree_t *size_tree = arg; ASSERT0(zfs_btree_numnodes(size_tree)); zfs_btree_destroy(size_tree); } static uint64_t ext_size_value(range_tree_t *rt, range_seg_gap_t *rsg) { (void) rt; uint64_t size = rsg->rs_end - rsg->rs_start; uint64_t score = rsg->rs_fill + ((((rsg->rs_fill << 7) / size) * fill_weight * rsg->rs_fill) >> 7); ASSERT3U(rt->rt_shift, >=, 8); return (((uint64_t)(64 - highbit64(score)) << 56) | rsg->rs_start); } static void ext_size_add(range_tree_t *rt, range_seg_t *rs, void *arg) { zfs_btree_t *size_tree = arg; ASSERT3U(rt->rt_type, ==, RANGE_SEG_GAP); uint64_t v = ext_size_value(rt, (range_seg_gap_t *)rs); zfs_btree_add(size_tree, &v); } static void ext_size_remove(range_tree_t *rt, range_seg_t *rs, void *arg) { zfs_btree_t *size_tree = arg; ASSERT3U(rt->rt_type, ==, RANGE_SEG_GAP); uint64_t v = ext_size_value(rt, (range_seg_gap_t *)rs); zfs_btree_remove(size_tree, &v); } static void ext_size_vacate(range_tree_t *rt, void *arg) { zfs_btree_t *size_tree = arg; zfs_btree_clear(size_tree); zfs_btree_destroy(size_tree); ext_size_create(rt, arg); } static const range_tree_ops_t ext_size_ops = { .rtop_create = ext_size_create, .rtop_destroy = ext_size_destroy, .rtop_add = ext_size_add, .rtop_remove = ext_size_remove, .rtop_vacate = ext_size_vacate }; /* * Comparator for the q_sios_by_addr tree. Sorting is simply performed * based on LBA-order (from lowest to highest). */ static int sio_addr_compare(const void *x, const void *y) { const scan_io_t *a = x, *b = y; return (TREE_CMP(SIO_GET_OFFSET(a), SIO_GET_OFFSET(b))); } /* IO queues are created on demand when they are needed. */ static dsl_scan_io_queue_t * scan_io_queue_create(vdev_t *vd) { dsl_scan_t *scn = vd->vdev_spa->spa_dsl_pool->dp_scan; dsl_scan_io_queue_t *q = kmem_zalloc(sizeof (*q), KM_SLEEP); q->q_scn = scn; q->q_vd = vd; q->q_sio_memused = 0; q->q_last_ext_addr = -1; cv_init(&q->q_zio_cv, NULL, CV_DEFAULT, NULL); q->q_exts_by_addr = range_tree_create_gap(&ext_size_ops, RANGE_SEG_GAP, &q->q_exts_by_size, 0, vd->vdev_ashift, zfs_scan_max_ext_gap); avl_create(&q->q_sios_by_addr, sio_addr_compare, sizeof (scan_io_t), offsetof(scan_io_t, sio_nodes.sio_addr_node)); return (q); } /* * Destroys a scan queue and all segments and scan_io_t's contained in it. * No further execution of I/O occurs, anything pending in the queue is * simply freed without being executed. */ void dsl_scan_io_queue_destroy(dsl_scan_io_queue_t *queue) { dsl_scan_t *scn = queue->q_scn; scan_io_t *sio; void *cookie = NULL; ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); if (!avl_is_empty(&queue->q_sios_by_addr)) atomic_add_64(&scn->scn_queues_pending, -1); while ((sio = avl_destroy_nodes(&queue->q_sios_by_addr, &cookie)) != NULL) { ASSERT(range_tree_contains(queue->q_exts_by_addr, SIO_GET_OFFSET(sio), SIO_GET_ASIZE(sio))); queue->q_sio_memused -= SIO_GET_MUSED(sio); sio_free(sio); } ASSERT0(queue->q_sio_memused); range_tree_vacate(queue->q_exts_by_addr, NULL, queue); range_tree_destroy(queue->q_exts_by_addr); avl_destroy(&queue->q_sios_by_addr); cv_destroy(&queue->q_zio_cv); kmem_free(queue, sizeof (*queue)); } /* * Properly transfers a dsl_scan_queue_t from `svd' to `tvd'. This is * called on behalf of vdev_top_transfer when creating or destroying * a mirror vdev due to zpool attach/detach. */ void dsl_scan_io_queue_vdev_xfer(vdev_t *svd, vdev_t *tvd) { mutex_enter(&svd->vdev_scan_io_queue_lock); mutex_enter(&tvd->vdev_scan_io_queue_lock); VERIFY3P(tvd->vdev_scan_io_queue, ==, NULL); tvd->vdev_scan_io_queue = svd->vdev_scan_io_queue; svd->vdev_scan_io_queue = NULL; if (tvd->vdev_scan_io_queue != NULL) tvd->vdev_scan_io_queue->q_vd = tvd; mutex_exit(&tvd->vdev_scan_io_queue_lock); mutex_exit(&svd->vdev_scan_io_queue_lock); } static void scan_io_queues_destroy(dsl_scan_t *scn) { vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev; for (uint64_t i = 0; i < rvd->vdev_children; i++) { vdev_t *tvd = rvd->vdev_child[i]; mutex_enter(&tvd->vdev_scan_io_queue_lock); if (tvd->vdev_scan_io_queue != NULL) dsl_scan_io_queue_destroy(tvd->vdev_scan_io_queue); tvd->vdev_scan_io_queue = NULL; mutex_exit(&tvd->vdev_scan_io_queue_lock); } } static void dsl_scan_freed_dva(spa_t *spa, const blkptr_t *bp, int dva_i) { dsl_pool_t *dp = spa->spa_dsl_pool; dsl_scan_t *scn = dp->dp_scan; vdev_t *vdev; kmutex_t *q_lock; dsl_scan_io_queue_t *queue; scan_io_t *srch_sio, *sio; avl_index_t idx; uint64_t start, size; vdev = vdev_lookup_top(spa, DVA_GET_VDEV(&bp->blk_dva[dva_i])); ASSERT(vdev != NULL); q_lock = &vdev->vdev_scan_io_queue_lock; queue = vdev->vdev_scan_io_queue; mutex_enter(q_lock); if (queue == NULL) { mutex_exit(q_lock); return; } srch_sio = sio_alloc(BP_GET_NDVAS(bp)); bp2sio(bp, srch_sio, dva_i); start = SIO_GET_OFFSET(srch_sio); size = SIO_GET_ASIZE(srch_sio); /* * We can find the zio in two states: * 1) Cold, just sitting in the queue of zio's to be issued at * some point in the future. In this case, all we do is * remove the zio from the q_sios_by_addr tree, decrement * its data volume from the containing range_seg_t and * resort the q_exts_by_size tree to reflect that the * range_seg_t has lost some of its 'fill'. We don't shorten * the range_seg_t - this is usually rare enough not to be * worth the extra hassle of trying keep track of precise * extent boundaries. * 2) Hot, where the zio is currently in-flight in * dsl_scan_issue_ios. In this case, we can't simply * reach in and stop the in-flight zio's, so we instead * block the caller. Eventually, dsl_scan_issue_ios will * be done with issuing the zio's it gathered and will * signal us. */ sio = avl_find(&queue->q_sios_by_addr, srch_sio, &idx); sio_free(srch_sio); if (sio != NULL) { blkptr_t tmpbp; /* Got it while it was cold in the queue */ ASSERT3U(start, ==, SIO_GET_OFFSET(sio)); ASSERT3U(size, ==, SIO_GET_ASIZE(sio)); avl_remove(&queue->q_sios_by_addr, sio); if (avl_is_empty(&queue->q_sios_by_addr)) atomic_add_64(&scn->scn_queues_pending, -1); queue->q_sio_memused -= SIO_GET_MUSED(sio); ASSERT(range_tree_contains(queue->q_exts_by_addr, start, size)); range_tree_remove_fill(queue->q_exts_by_addr, start, size); /* count the block as though we issued it */ sio2bp(sio, &tmpbp); count_block_issued(spa, &tmpbp, B_FALSE); sio_free(sio); } mutex_exit(q_lock); } /* * Callback invoked when a zio_free() zio is executing. This needs to be * intercepted to prevent the zio from deallocating a particular portion * of disk space and it then getting reallocated and written to, while we * still have it queued up for processing. */ void dsl_scan_freed(spa_t *spa, const blkptr_t *bp) { dsl_pool_t *dp = spa->spa_dsl_pool; dsl_scan_t *scn = dp->dp_scan; ASSERT(!BP_IS_EMBEDDED(bp)); ASSERT(scn != NULL); if (!dsl_scan_is_running(scn)) return; for (int i = 0; i < BP_GET_NDVAS(bp); i++) dsl_scan_freed_dva(spa, bp, i); } /* * Check if a vdev needs resilvering (non-empty DTL), if so, and resilver has * not started, start it. Otherwise, only restart if max txg in DTL range is * greater than the max txg in the current scan. If the DTL max is less than * the scan max, then the vdev has not missed any new data since the resilver * started, so a restart is not needed. */ void dsl_scan_assess_vdev(dsl_pool_t *dp, vdev_t *vd) { uint64_t min, max; if (!vdev_resilver_needed(vd, &min, &max)) return; if (!dsl_scan_resilvering(dp)) { spa_async_request(dp->dp_spa, SPA_ASYNC_RESILVER); return; } if (max <= dp->dp_scan->scn_phys.scn_max_txg) return; /* restart is needed, check if it can be deferred */ if (spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER)) vdev_defer_resilver(vd); else spa_async_request(dp->dp_spa, SPA_ASYNC_RESILVER); } /* BEGIN CSTYLED */ ZFS_MODULE_PARAM(zfs, zfs_, scan_vdev_limit, ULONG, ZMOD_RW, "Max bytes in flight per leaf vdev for scrubs and resilvers"); ZFS_MODULE_PARAM(zfs, zfs_, scrub_min_time_ms, INT, ZMOD_RW, "Min millisecs to scrub per txg"); ZFS_MODULE_PARAM(zfs, zfs_, obsolete_min_time_ms, INT, ZMOD_RW, "Min millisecs to obsolete per txg"); ZFS_MODULE_PARAM(zfs, zfs_, free_min_time_ms, INT, ZMOD_RW, "Min millisecs to free per txg"); ZFS_MODULE_PARAM(zfs, zfs_, resilver_min_time_ms, INT, ZMOD_RW, "Min millisecs to resilver per txg"); ZFS_MODULE_PARAM(zfs, zfs_, scan_suspend_progress, INT, ZMOD_RW, "Set to prevent scans from progressing"); ZFS_MODULE_PARAM(zfs, zfs_, no_scrub_io, INT, ZMOD_RW, "Set to disable scrub I/O"); ZFS_MODULE_PARAM(zfs, zfs_, no_scrub_prefetch, INT, ZMOD_RW, "Set to disable scrub prefetching"); ZFS_MODULE_PARAM(zfs, zfs_, async_block_max_blocks, ULONG, ZMOD_RW, "Max number of blocks freed in one txg"); ZFS_MODULE_PARAM(zfs, zfs_, max_async_dedup_frees, ULONG, ZMOD_RW, "Max number of dedup blocks freed in one txg"); ZFS_MODULE_PARAM(zfs, zfs_, free_bpobj_enabled, INT, ZMOD_RW, "Enable processing of the free_bpobj"); ZFS_MODULE_PARAM(zfs, zfs_, scan_blkstats, INT, ZMOD_RW, "Enable block statistics calculation during scrub"); ZFS_MODULE_PARAM(zfs, zfs_, scan_mem_lim_fact, INT, ZMOD_RW, "Fraction of RAM for scan hard limit"); ZFS_MODULE_PARAM(zfs, zfs_, scan_issue_strategy, INT, ZMOD_RW, "IO issuing strategy during scrubbing. " "0 = default, 1 = LBA, 2 = size"); ZFS_MODULE_PARAM(zfs, zfs_, scan_legacy, INT, ZMOD_RW, "Scrub using legacy non-sequential method"); ZFS_MODULE_PARAM(zfs, zfs_, scan_checkpoint_intval, INT, ZMOD_RW, "Scan progress on-disk checkpointing interval"); ZFS_MODULE_PARAM(zfs, zfs_, scan_max_ext_gap, ULONG, ZMOD_RW, "Max gap in bytes between sequential scrub / resilver I/Os"); ZFS_MODULE_PARAM(zfs, zfs_, scan_mem_lim_soft_fact, INT, ZMOD_RW, "Fraction of hard limit used as soft limit"); ZFS_MODULE_PARAM(zfs, zfs_, scan_strict_mem_lim, INT, ZMOD_RW, "Tunable to attempt to reduce lock contention"); ZFS_MODULE_PARAM(zfs, zfs_, scan_fill_weight, INT, ZMOD_RW, "Tunable to adjust bias towards more filled segments during scans"); +ZFS_MODULE_PARAM(zfs, zfs_, scan_report_txgs, UINT, ZMOD_RW, + "Tunable to report resilver performance over the last N txgs"); + ZFS_MODULE_PARAM(zfs, zfs_, resilver_disable_defer, INT, ZMOD_RW, "Process all resilvers immediately"); /* END CSTYLED */ diff --git a/module/zfs/spa_misc.c b/module/zfs/spa_misc.c index a57f0727db31..10837241b52c 100644 --- a/module/zfs/spa_misc.c +++ b/module/zfs/spa_misc.c @@ -1,2967 +1,2966 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2019 by Delphix. All rights reserved. * Copyright 2015 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. * Copyright 2013 Saso Kiselkov. All rights reserved. * Copyright (c) 2017 Datto Inc. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2019, loli10K . All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zfs_prop.h" #include #include #include #include /* * SPA locking * * There are three basic locks for managing spa_t structures: * * spa_namespace_lock (global mutex) * * This lock must be acquired to do any of the following: * * - Lookup a spa_t by name * - Add or remove a spa_t from the namespace * - Increase spa_refcount from non-zero * - Check if spa_refcount is zero * - Rename a spa_t * - add/remove/attach/detach devices * - Held for the duration of create/destroy/import/export * * It does not need to handle recursion. A create or destroy may * reference objects (files or zvols) in other pools, but by * definition they must have an existing reference, and will never need * to lookup a spa_t by name. * * spa_refcount (per-spa zfs_refcount_t protected by mutex) * * This reference count keep track of any active users of the spa_t. The * spa_t cannot be destroyed or freed while this is non-zero. Internally, * the refcount is never really 'zero' - opening a pool implicitly keeps * some references in the DMU. Internally we check against spa_minref, but * present the image of a zero/non-zero value to consumers. * * spa_config_lock[] (per-spa array of rwlocks) * * This protects the spa_t from config changes, and must be held in * the following circumstances: * * - RW_READER to perform I/O to the spa * - RW_WRITER to change the vdev config * * The locking order is fairly straightforward: * * spa_namespace_lock -> spa_refcount * * The namespace lock must be acquired to increase the refcount from 0 * or to check if it is zero. * * spa_refcount -> spa_config_lock[] * * There must be at least one valid reference on the spa_t to acquire * the config lock. * * spa_namespace_lock -> spa_config_lock[] * * The namespace lock must always be taken before the config lock. * * * The spa_namespace_lock can be acquired directly and is globally visible. * * The namespace is manipulated using the following functions, all of which * require the spa_namespace_lock to be held. * * spa_lookup() Lookup a spa_t by name. * * spa_add() Create a new spa_t in the namespace. * * spa_remove() Remove a spa_t from the namespace. This also * frees up any memory associated with the spa_t. * * spa_next() Returns the next spa_t in the system, or the * first if NULL is passed. * * spa_evict_all() Shutdown and remove all spa_t structures in * the system. * * spa_guid_exists() Determine whether a pool/device guid exists. * * The spa_refcount is manipulated using the following functions: * * spa_open_ref() Adds a reference to the given spa_t. Must be * called with spa_namespace_lock held if the * refcount is currently zero. * * spa_close() Remove a reference from the spa_t. This will * not free the spa_t or remove it from the * namespace. No locking is required. * * spa_refcount_zero() Returns true if the refcount is currently * zero. Must be called with spa_namespace_lock * held. * * The spa_config_lock[] is an array of rwlocks, ordered as follows: * SCL_CONFIG > SCL_STATE > SCL_ALLOC > SCL_ZIO > SCL_FREE > SCL_VDEV. * spa_config_lock[] is manipulated with spa_config_{enter,exit,held}(). * * To read the configuration, it suffices to hold one of these locks as reader. * To modify the configuration, you must hold all locks as writer. To modify * vdev state without altering the vdev tree's topology (e.g. online/offline), * you must hold SCL_STATE and SCL_ZIO as writer. * * We use these distinct config locks to avoid recursive lock entry. * For example, spa_sync() (which holds SCL_CONFIG as reader) induces * block allocations (SCL_ALLOC), which may require reading space maps * from disk (dmu_read() -> zio_read() -> SCL_ZIO). * * The spa config locks cannot be normal rwlocks because we need the * ability to hand off ownership. For example, SCL_ZIO is acquired * by the issuing thread and later released by an interrupt thread. * They do, however, obey the usual write-wanted semantics to prevent * writer (i.e. system administrator) starvation. * * The lock acquisition rules are as follows: * * SCL_CONFIG * Protects changes to the vdev tree topology, such as vdev * add/remove/attach/detach. Protects the dirty config list * (spa_config_dirty_list) and the set of spares and l2arc devices. * * SCL_STATE * Protects changes to pool state and vdev state, such as vdev * online/offline/fault/degrade/clear. Protects the dirty state list * (spa_state_dirty_list) and global pool state (spa_state). * * SCL_ALLOC * Protects changes to metaslab groups and classes. * Held as reader by metaslab_alloc() and metaslab_claim(). * * SCL_ZIO * Held by bp-level zios (those which have no io_vd upon entry) * to prevent changes to the vdev tree. The bp-level zio implicitly * protects all of its vdev child zios, which do not hold SCL_ZIO. * * SCL_FREE * Protects changes to metaslab groups and classes. * Held as reader by metaslab_free(). SCL_FREE is distinct from * SCL_ALLOC, and lower than SCL_ZIO, so that we can safely free * blocks in zio_done() while another i/o that holds either * SCL_ALLOC or SCL_ZIO is waiting for this i/o to complete. * * SCL_VDEV * Held as reader to prevent changes to the vdev tree during trivial * inquiries such as bp_get_dsize(). SCL_VDEV is distinct from the * other locks, and lower than all of them, to ensure that it's safe * to acquire regardless of caller context. * * In addition, the following rules apply: * * (a) spa_props_lock protects pool properties, spa_config and spa_config_list. * The lock ordering is SCL_CONFIG > spa_props_lock. * * (b) I/O operations on leaf vdevs. For any zio operation that takes * an explicit vdev_t argument -- such as zio_ioctl(), zio_read_phys(), * or zio_write_phys() -- the caller must ensure that the config cannot * cannot change in the interim, and that the vdev cannot be reopened. * SCL_STATE as reader suffices for both. * * The vdev configuration is protected by spa_vdev_enter() / spa_vdev_exit(). * * spa_vdev_enter() Acquire the namespace lock and the config lock * for writing. * * spa_vdev_exit() Release the config lock, wait for all I/O * to complete, sync the updated configs to the * cache, and release the namespace lock. * * vdev state is protected by spa_vdev_state_enter() / spa_vdev_state_exit(). * Like spa_vdev_enter/exit, these are convenience wrappers -- the actual * locking is, always, based on spa_namespace_lock and spa_config_lock[]. */ static avl_tree_t spa_namespace_avl; kmutex_t spa_namespace_lock; static kcondvar_t spa_namespace_cv; int spa_max_replication_override = SPA_DVAS_PER_BP; static kmutex_t spa_spare_lock; static avl_tree_t spa_spare_avl; static kmutex_t spa_l2cache_lock; static avl_tree_t spa_l2cache_avl; kmem_cache_t *spa_buffer_pool; spa_mode_t spa_mode_global = SPA_MODE_UNINIT; #ifdef ZFS_DEBUG /* * Everything except dprintf, set_error, spa, and indirect_remap is on * by default in debug builds. */ int zfs_flags = ~(ZFS_DEBUG_DPRINTF | ZFS_DEBUG_SET_ERROR | ZFS_DEBUG_INDIRECT_REMAP); #else int zfs_flags = 0; #endif /* * zfs_recover can be set to nonzero to attempt to recover from * otherwise-fatal errors, typically caused by on-disk corruption. When * set, calls to zfs_panic_recover() will turn into warning messages. * This should only be used as a last resort, as it typically results * in leaked space, or worse. */ int zfs_recover = B_FALSE; /* * If destroy encounters an EIO while reading metadata (e.g. indirect * blocks), space referenced by the missing metadata can not be freed. * Normally this causes the background destroy to become "stalled", as * it is unable to make forward progress. While in this stalled state, * all remaining space to free from the error-encountering filesystem is * "temporarily leaked". Set this flag to cause it to ignore the EIO, * permanently leak the space from indirect blocks that can not be read, * and continue to free everything else that it can. * * The default, "stalling" behavior is useful if the storage partially * fails (i.e. some but not all i/os fail), and then later recovers. In * this case, we will be able to continue pool operations while it is * partially failed, and when it recovers, we can continue to free the * space, with no leaks. However, note that this case is actually * fairly rare. * * Typically pools either (a) fail completely (but perhaps temporarily, * e.g. a top-level vdev going offline), or (b) have localized, * permanent errors (e.g. disk returns the wrong data due to bit flip or * firmware bug). In case (a), this setting does not matter because the * pool will be suspended and the sync thread will not be able to make * forward progress regardless. In case (b), because the error is * permanent, the best we can do is leak the minimum amount of space, * which is what setting this flag will do. Therefore, it is reasonable * for this flag to normally be set, but we chose the more conservative * approach of not setting it, so that there is no possibility of * leaking space in the "partial temporary" failure case. */ int zfs_free_leak_on_eio = B_FALSE; /* * Expiration time in milliseconds. This value has two meanings. First it is * used to determine when the spa_deadman() logic should fire. By default the * spa_deadman() will fire if spa_sync() has not completed in 600 seconds. * Secondly, the value determines if an I/O is considered "hung". Any I/O that * has not completed in zfs_deadman_synctime_ms is considered "hung" resulting * in one of three behaviors controlled by zfs_deadman_failmode. */ unsigned long zfs_deadman_synctime_ms = 600000UL; /* * This value controls the maximum amount of time zio_wait() will block for an * outstanding IO. By default this is 300 seconds at which point the "hung" * behavior will be applied as described for zfs_deadman_synctime_ms. */ unsigned long zfs_deadman_ziotime_ms = 300000UL; /* * Check time in milliseconds. This defines the frequency at which we check * for hung I/O. */ unsigned long zfs_deadman_checktime_ms = 60000UL; /* * By default the deadman is enabled. */ int zfs_deadman_enabled = 1; /* * Controls the behavior of the deadman when it detects a "hung" I/O. * Valid values are zfs_deadman_failmode=. * * wait - Wait for the "hung" I/O (default) * continue - Attempt to recover from a "hung" I/O * panic - Panic the system */ char *zfs_deadman_failmode = "wait"; /* * The worst case is single-sector max-parity RAID-Z blocks, in which * case the space requirement is exactly (VDEV_RAIDZ_MAXPARITY + 1) * times the size; so just assume that. Add to this the fact that * we can have up to 3 DVAs per bp, and one more factor of 2 because * the block may be dittoed with up to 3 DVAs by ddt_sync(). All together, * the worst case is: * (VDEV_RAIDZ_MAXPARITY + 1) * SPA_DVAS_PER_BP * 2 == 24 */ int spa_asize_inflation = 24; /* * Normally, we don't allow the last 3.2% (1/(2^spa_slop_shift)) of space in * the pool to be consumed (bounded by spa_max_slop). This ensures that we * don't run the pool completely out of space, due to unaccounted changes (e.g. * to the MOS). It also limits the worst-case time to allocate space. If we * have less than this amount of free space, most ZPL operations (e.g. write, * create) will return ENOSPC. The ZIL metaslabs (spa_embedded_log_class) are * also part of this 3.2% of space which can't be consumed by normal writes; * the slop space "proper" (spa_get_slop_space()) is decreased by the embedded * log space. * * Certain operations (e.g. file removal, most administrative actions) can * use half the slop space. They will only return ENOSPC if less than half * the slop space is free. Typically, once the pool has less than the slop * space free, the user will use these operations to free up space in the pool. * These are the operations that call dsl_pool_adjustedsize() with the netfree * argument set to TRUE. * * Operations that are almost guaranteed to free up space in the absence of * a pool checkpoint can use up to three quarters of the slop space * (e.g zfs destroy). * * A very restricted set of operations are always permitted, regardless of * the amount of free space. These are the operations that call * dsl_sync_task(ZFS_SPACE_CHECK_NONE). If these operations result in a net * increase in the amount of space used, it is possible to run the pool * completely out of space, causing it to be permanently read-only. * * Note that on very small pools, the slop space will be larger than * 3.2%, in an effort to have it be at least spa_min_slop (128MB), * but we never allow it to be more than half the pool size. * * Further, on very large pools, the slop space will be smaller than * 3.2%, to avoid reserving much more space than we actually need; bounded * by spa_max_slop (128GB). * * See also the comments in zfs_space_check_t. */ int spa_slop_shift = 5; uint64_t spa_min_slop = 128ULL * 1024 * 1024; uint64_t spa_max_slop = 128ULL * 1024 * 1024 * 1024; int spa_allocators = 4; /*PRINTFLIKE2*/ void spa_load_failed(spa_t *spa, const char *fmt, ...) { va_list adx; char buf[256]; va_start(adx, fmt); (void) vsnprintf(buf, sizeof (buf), fmt, adx); va_end(adx); zfs_dbgmsg("spa_load(%s, config %s): FAILED: %s", spa->spa_name, spa->spa_trust_config ? "trusted" : "untrusted", buf); } /*PRINTFLIKE2*/ void spa_load_note(spa_t *spa, const char *fmt, ...) { va_list adx; char buf[256]; va_start(adx, fmt); (void) vsnprintf(buf, sizeof (buf), fmt, adx); va_end(adx); zfs_dbgmsg("spa_load(%s, config %s): %s", spa->spa_name, spa->spa_trust_config ? "trusted" : "untrusted", buf); } /* * By default dedup and user data indirects land in the special class */ int zfs_ddt_data_is_special = B_TRUE; int zfs_user_indirect_is_special = B_TRUE; /* * The percentage of special class final space reserved for metadata only. * Once we allocate 100 - zfs_special_class_metadata_reserve_pct we only * let metadata into the class. */ int zfs_special_class_metadata_reserve_pct = 25; /* * ========================================================================== * SPA config locking * ========================================================================== */ static void spa_config_lock_init(spa_t *spa) { for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; mutex_init(&scl->scl_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&scl->scl_cv, NULL, CV_DEFAULT, NULL); scl->scl_writer = NULL; scl->scl_write_wanted = 0; scl->scl_count = 0; } } static void spa_config_lock_destroy(spa_t *spa) { for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; mutex_destroy(&scl->scl_lock); cv_destroy(&scl->scl_cv); ASSERT(scl->scl_writer == NULL); ASSERT(scl->scl_write_wanted == 0); ASSERT(scl->scl_count == 0); } } int spa_config_tryenter(spa_t *spa, int locks, void *tag, krw_t rw) { for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; if (!(locks & (1 << i))) continue; mutex_enter(&scl->scl_lock); if (rw == RW_READER) { if (scl->scl_writer || scl->scl_write_wanted) { mutex_exit(&scl->scl_lock); spa_config_exit(spa, locks & ((1 << i) - 1), tag); return (0); } } else { ASSERT(scl->scl_writer != curthread); if (scl->scl_count != 0) { mutex_exit(&scl->scl_lock); spa_config_exit(spa, locks & ((1 << i) - 1), tag); return (0); } scl->scl_writer = curthread; } scl->scl_count++; mutex_exit(&scl->scl_lock); } return (1); } void spa_config_enter(spa_t *spa, int locks, const void *tag, krw_t rw) { (void) tag; int wlocks_held = 0; ASSERT3U(SCL_LOCKS, <, sizeof (wlocks_held) * NBBY); for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; if (scl->scl_writer == curthread) wlocks_held |= (1 << i); if (!(locks & (1 << i))) continue; mutex_enter(&scl->scl_lock); if (rw == RW_READER) { while (scl->scl_writer || scl->scl_write_wanted) { cv_wait(&scl->scl_cv, &scl->scl_lock); } } else { ASSERT(scl->scl_writer != curthread); while (scl->scl_count != 0) { scl->scl_write_wanted++; cv_wait(&scl->scl_cv, &scl->scl_lock); scl->scl_write_wanted--; } scl->scl_writer = curthread; } scl->scl_count++; mutex_exit(&scl->scl_lock); } ASSERT3U(wlocks_held, <=, locks); } void spa_config_exit(spa_t *spa, int locks, const void *tag) { (void) tag; for (int i = SCL_LOCKS - 1; i >= 0; i--) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; if (!(locks & (1 << i))) continue; mutex_enter(&scl->scl_lock); ASSERT(scl->scl_count > 0); if (--scl->scl_count == 0) { ASSERT(scl->scl_writer == NULL || scl->scl_writer == curthread); scl->scl_writer = NULL; /* OK in either case */ cv_broadcast(&scl->scl_cv); } mutex_exit(&scl->scl_lock); } } int spa_config_held(spa_t *spa, int locks, krw_t rw) { int locks_held = 0; for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; if (!(locks & (1 << i))) continue; if ((rw == RW_READER && scl->scl_count != 0) || (rw == RW_WRITER && scl->scl_writer == curthread)) locks_held |= 1 << i; } return (locks_held); } /* * ========================================================================== * SPA namespace functions * ========================================================================== */ /* * Lookup the named spa_t in the AVL tree. The spa_namespace_lock must be held. * Returns NULL if no matching spa_t is found. */ spa_t * spa_lookup(const char *name) { static spa_t search; /* spa_t is large; don't allocate on stack */ spa_t *spa; avl_index_t where; char *cp; ASSERT(MUTEX_HELD(&spa_namespace_lock)); (void) strlcpy(search.spa_name, name, sizeof (search.spa_name)); /* * If it's a full dataset name, figure out the pool name and * just use that. */ cp = strpbrk(search.spa_name, "/@#"); if (cp != NULL) *cp = '\0'; spa = avl_find(&spa_namespace_avl, &search, &where); return (spa); } /* * Fires when spa_sync has not completed within zfs_deadman_synctime_ms. * If the zfs_deadman_enabled flag is set then it inspects all vdev queues * looking for potentially hung I/Os. */ void spa_deadman(void *arg) { spa_t *spa = arg; /* Disable the deadman if the pool is suspended. */ if (spa_suspended(spa)) return; zfs_dbgmsg("slow spa_sync: started %llu seconds ago, calls %llu", (gethrtime() - spa->spa_sync_starttime) / NANOSEC, (u_longlong_t)++spa->spa_deadman_calls); if (zfs_deadman_enabled) vdev_deadman(spa->spa_root_vdev, FTAG); spa->spa_deadman_tqid = taskq_dispatch_delay(system_delay_taskq, spa_deadman, spa, TQ_SLEEP, ddi_get_lbolt() + MSEC_TO_TICK(zfs_deadman_checktime_ms)); } static int spa_log_sm_sort_by_txg(const void *va, const void *vb) { const spa_log_sm_t *a = va; const spa_log_sm_t *b = vb; return (TREE_CMP(a->sls_txg, b->sls_txg)); } /* * Create an uninitialized spa_t with the given name. Requires * spa_namespace_lock. The caller must ensure that the spa_t doesn't already * exist by calling spa_lookup() first. */ spa_t * spa_add(const char *name, nvlist_t *config, const char *altroot) { spa_t *spa; spa_config_dirent_t *dp; ASSERT(MUTEX_HELD(&spa_namespace_lock)); spa = kmem_zalloc(sizeof (spa_t), KM_SLEEP); mutex_init(&spa->spa_async_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_errlist_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_errlog_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_evicting_os_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_history_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_proc_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_props_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_cksum_tmpls_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_scrub_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_suspend_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_vdev_top_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_feat_stats_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_flushed_ms_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_activities_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&spa->spa_async_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_evicting_os_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_proc_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_scrub_io_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_suspend_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_activities_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_waiters_cv, NULL, CV_DEFAULT, NULL); for (int t = 0; t < TXG_SIZE; t++) bplist_create(&spa->spa_free_bplist[t]); (void) strlcpy(spa->spa_name, name, sizeof (spa->spa_name)); spa->spa_state = POOL_STATE_UNINITIALIZED; spa->spa_freeze_txg = UINT64_MAX; spa->spa_final_txg = UINT64_MAX; spa->spa_load_max_txg = UINT64_MAX; spa->spa_proc = &p0; spa->spa_proc_state = SPA_PROC_NONE; spa->spa_trust_config = B_TRUE; spa->spa_hostid = zone_get_hostid(NULL); spa->spa_deadman_synctime = MSEC2NSEC(zfs_deadman_synctime_ms); spa->spa_deadman_ziotime = MSEC2NSEC(zfs_deadman_ziotime_ms); spa_set_deadman_failmode(spa, zfs_deadman_failmode); zfs_refcount_create(&spa->spa_refcount); spa_config_lock_init(spa); spa_stats_init(spa); avl_add(&spa_namespace_avl, spa); /* * Set the alternate root, if there is one. */ if (altroot) spa->spa_root = spa_strdup(altroot); spa->spa_alloc_count = spa_allocators; spa->spa_allocs = kmem_zalloc(spa->spa_alloc_count * sizeof (spa_alloc_t), KM_SLEEP); for (int i = 0; i < spa->spa_alloc_count; i++) { mutex_init(&spa->spa_allocs[i].spaa_lock, NULL, MUTEX_DEFAULT, NULL); avl_create(&spa->spa_allocs[i].spaa_tree, zio_bookmark_compare, sizeof (zio_t), offsetof(zio_t, io_alloc_node)); } avl_create(&spa->spa_metaslabs_by_flushed, metaslab_sort_by_flushed, sizeof (metaslab_t), offsetof(metaslab_t, ms_spa_txg_node)); avl_create(&spa->spa_sm_logs_by_txg, spa_log_sm_sort_by_txg, sizeof (spa_log_sm_t), offsetof(spa_log_sm_t, sls_node)); list_create(&spa->spa_log_summary, sizeof (log_summary_entry_t), offsetof(log_summary_entry_t, lse_node)); /* * Every pool starts with the default cachefile */ list_create(&spa->spa_config_list, sizeof (spa_config_dirent_t), offsetof(spa_config_dirent_t, scd_link)); dp = kmem_zalloc(sizeof (spa_config_dirent_t), KM_SLEEP); dp->scd_path = altroot ? NULL : spa_strdup(spa_config_path); list_insert_head(&spa->spa_config_list, dp); VERIFY(nvlist_alloc(&spa->spa_load_info, NV_UNIQUE_NAME, KM_SLEEP) == 0); if (config != NULL) { nvlist_t *features; if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_FEATURES_FOR_READ, &features) == 0) { VERIFY(nvlist_dup(features, &spa->spa_label_features, 0) == 0); } VERIFY(nvlist_dup(config, &spa->spa_config, 0) == 0); } if (spa->spa_label_features == NULL) { VERIFY(nvlist_alloc(&spa->spa_label_features, NV_UNIQUE_NAME, KM_SLEEP) == 0); } spa->spa_min_ashift = INT_MAX; spa->spa_max_ashift = 0; spa->spa_min_alloc = INT_MAX; /* Reset cached value */ spa->spa_dedup_dspace = ~0ULL; /* * As a pool is being created, treat all features as disabled by * setting SPA_FEATURE_DISABLED for all entries in the feature * refcount cache. */ for (int i = 0; i < SPA_FEATURES; i++) { spa->spa_feat_refcount_cache[i] = SPA_FEATURE_DISABLED; } list_create(&spa->spa_leaf_list, sizeof (vdev_t), offsetof(vdev_t, vdev_leaf_node)); return (spa); } /* * Removes a spa_t from the namespace, freeing up any memory used. Requires * spa_namespace_lock. This is called only after the spa_t has been closed and * deactivated. */ void spa_remove(spa_t *spa) { spa_config_dirent_t *dp; ASSERT(MUTEX_HELD(&spa_namespace_lock)); ASSERT(spa_state(spa) == POOL_STATE_UNINITIALIZED); ASSERT3U(zfs_refcount_count(&spa->spa_refcount), ==, 0); ASSERT0(spa->spa_waiters); nvlist_free(spa->spa_config_splitting); avl_remove(&spa_namespace_avl, spa); cv_broadcast(&spa_namespace_cv); if (spa->spa_root) spa_strfree(spa->spa_root); while ((dp = list_head(&spa->spa_config_list)) != NULL) { list_remove(&spa->spa_config_list, dp); if (dp->scd_path != NULL) spa_strfree(dp->scd_path); kmem_free(dp, sizeof (spa_config_dirent_t)); } for (int i = 0; i < spa->spa_alloc_count; i++) { avl_destroy(&spa->spa_allocs[i].spaa_tree); mutex_destroy(&spa->spa_allocs[i].spaa_lock); } kmem_free(spa->spa_allocs, spa->spa_alloc_count * sizeof (spa_alloc_t)); avl_destroy(&spa->spa_metaslabs_by_flushed); avl_destroy(&spa->spa_sm_logs_by_txg); list_destroy(&spa->spa_log_summary); list_destroy(&spa->spa_config_list); list_destroy(&spa->spa_leaf_list); nvlist_free(spa->spa_label_features); nvlist_free(spa->spa_load_info); nvlist_free(spa->spa_feat_stats); spa_config_set(spa, NULL); zfs_refcount_destroy(&spa->spa_refcount); spa_stats_destroy(spa); spa_config_lock_destroy(spa); for (int t = 0; t < TXG_SIZE; t++) bplist_destroy(&spa->spa_free_bplist[t]); zio_checksum_templates_free(spa); cv_destroy(&spa->spa_async_cv); cv_destroy(&spa->spa_evicting_os_cv); cv_destroy(&spa->spa_proc_cv); cv_destroy(&spa->spa_scrub_io_cv); cv_destroy(&spa->spa_suspend_cv); cv_destroy(&spa->spa_activities_cv); cv_destroy(&spa->spa_waiters_cv); mutex_destroy(&spa->spa_flushed_ms_lock); mutex_destroy(&spa->spa_async_lock); mutex_destroy(&spa->spa_errlist_lock); mutex_destroy(&spa->spa_errlog_lock); mutex_destroy(&spa->spa_evicting_os_lock); mutex_destroy(&spa->spa_history_lock); mutex_destroy(&spa->spa_proc_lock); mutex_destroy(&spa->spa_props_lock); mutex_destroy(&spa->spa_cksum_tmpls_lock); mutex_destroy(&spa->spa_scrub_lock); mutex_destroy(&spa->spa_suspend_lock); mutex_destroy(&spa->spa_vdev_top_lock); mutex_destroy(&spa->spa_feat_stats_lock); mutex_destroy(&spa->spa_activities_lock); kmem_free(spa, sizeof (spa_t)); } /* * Given a pool, return the next pool in the namespace, or NULL if there is * none. If 'prev' is NULL, return the first pool. */ spa_t * spa_next(spa_t *prev) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); if (prev) return (AVL_NEXT(&spa_namespace_avl, prev)); else return (avl_first(&spa_namespace_avl)); } /* * ========================================================================== * SPA refcount functions * ========================================================================== */ /* * Add a reference to the given spa_t. Must have at least one reference, or * have the namespace lock held. */ void spa_open_ref(spa_t *spa, void *tag) { ASSERT(zfs_refcount_count(&spa->spa_refcount) >= spa->spa_minref || MUTEX_HELD(&spa_namespace_lock)); (void) zfs_refcount_add(&spa->spa_refcount, tag); } /* * Remove a reference to the given spa_t. Must have at least one reference, or * have the namespace lock held. */ void spa_close(spa_t *spa, void *tag) { ASSERT(zfs_refcount_count(&spa->spa_refcount) > spa->spa_minref || MUTEX_HELD(&spa_namespace_lock)); (void) zfs_refcount_remove(&spa->spa_refcount, tag); } /* * Remove a reference to the given spa_t held by a dsl dir that is * being asynchronously released. Async releases occur from a taskq * performing eviction of dsl datasets and dirs. The namespace lock * isn't held and the hold by the object being evicted may contribute to * spa_minref (e.g. dataset or directory released during pool export), * so the asserts in spa_close() do not apply. */ void spa_async_close(spa_t *spa, void *tag) { (void) zfs_refcount_remove(&spa->spa_refcount, tag); } /* * Check to see if the spa refcount is zero. Must be called with * spa_namespace_lock held. We really compare against spa_minref, which is the * number of references acquired when opening a pool */ boolean_t spa_refcount_zero(spa_t *spa) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); return (zfs_refcount_count(&spa->spa_refcount) == spa->spa_minref); } /* * ========================================================================== * SPA spare and l2cache tracking * ========================================================================== */ /* * Hot spares and cache devices are tracked using the same code below, * for 'auxiliary' devices. */ typedef struct spa_aux { uint64_t aux_guid; uint64_t aux_pool; avl_node_t aux_avl; int aux_count; } spa_aux_t; static inline int spa_aux_compare(const void *a, const void *b) { const spa_aux_t *sa = (const spa_aux_t *)a; const spa_aux_t *sb = (const spa_aux_t *)b; return (TREE_CMP(sa->aux_guid, sb->aux_guid)); } static void spa_aux_add(vdev_t *vd, avl_tree_t *avl) { avl_index_t where; spa_aux_t search; spa_aux_t *aux; search.aux_guid = vd->vdev_guid; if ((aux = avl_find(avl, &search, &where)) != NULL) { aux->aux_count++; } else { aux = kmem_zalloc(sizeof (spa_aux_t), KM_SLEEP); aux->aux_guid = vd->vdev_guid; aux->aux_count = 1; avl_insert(avl, aux, where); } } static void spa_aux_remove(vdev_t *vd, avl_tree_t *avl) { spa_aux_t search; spa_aux_t *aux; avl_index_t where; search.aux_guid = vd->vdev_guid; aux = avl_find(avl, &search, &where); ASSERT(aux != NULL); if (--aux->aux_count == 0) { avl_remove(avl, aux); kmem_free(aux, sizeof (spa_aux_t)); } else if (aux->aux_pool == spa_guid(vd->vdev_spa)) { aux->aux_pool = 0ULL; } } static boolean_t spa_aux_exists(uint64_t guid, uint64_t *pool, int *refcnt, avl_tree_t *avl) { spa_aux_t search, *found; search.aux_guid = guid; found = avl_find(avl, &search, NULL); if (pool) { if (found) *pool = found->aux_pool; else *pool = 0ULL; } if (refcnt) { if (found) *refcnt = found->aux_count; else *refcnt = 0; } return (found != NULL); } static void spa_aux_activate(vdev_t *vd, avl_tree_t *avl) { spa_aux_t search, *found; avl_index_t where; search.aux_guid = vd->vdev_guid; found = avl_find(avl, &search, &where); ASSERT(found != NULL); ASSERT(found->aux_pool == 0ULL); found->aux_pool = spa_guid(vd->vdev_spa); } /* * Spares are tracked globally due to the following constraints: * * - A spare may be part of multiple pools. * - A spare may be added to a pool even if it's actively in use within * another pool. * - A spare in use in any pool can only be the source of a replacement if * the target is a spare in the same pool. * * We keep track of all spares on the system through the use of a reference * counted AVL tree. When a vdev is added as a spare, or used as a replacement * spare, then we bump the reference count in the AVL tree. In addition, we set * the 'vdev_isspare' member to indicate that the device is a spare (active or * inactive). When a spare is made active (used to replace a device in the * pool), we also keep track of which pool its been made a part of. * * The 'spa_spare_lock' protects the AVL tree. These functions are normally * called under the spa_namespace lock as part of vdev reconfiguration. The * separate spare lock exists for the status query path, which does not need to * be completely consistent with respect to other vdev configuration changes. */ static int spa_spare_compare(const void *a, const void *b) { return (spa_aux_compare(a, b)); } void spa_spare_add(vdev_t *vd) { mutex_enter(&spa_spare_lock); ASSERT(!vd->vdev_isspare); spa_aux_add(vd, &spa_spare_avl); vd->vdev_isspare = B_TRUE; mutex_exit(&spa_spare_lock); } void spa_spare_remove(vdev_t *vd) { mutex_enter(&spa_spare_lock); ASSERT(vd->vdev_isspare); spa_aux_remove(vd, &spa_spare_avl); vd->vdev_isspare = B_FALSE; mutex_exit(&spa_spare_lock); } boolean_t spa_spare_exists(uint64_t guid, uint64_t *pool, int *refcnt) { boolean_t found; mutex_enter(&spa_spare_lock); found = spa_aux_exists(guid, pool, refcnt, &spa_spare_avl); mutex_exit(&spa_spare_lock); return (found); } void spa_spare_activate(vdev_t *vd) { mutex_enter(&spa_spare_lock); ASSERT(vd->vdev_isspare); spa_aux_activate(vd, &spa_spare_avl); mutex_exit(&spa_spare_lock); } /* * Level 2 ARC devices are tracked globally for the same reasons as spares. * Cache devices currently only support one pool per cache device, and so * for these devices the aux reference count is currently unused beyond 1. */ static int spa_l2cache_compare(const void *a, const void *b) { return (spa_aux_compare(a, b)); } void spa_l2cache_add(vdev_t *vd) { mutex_enter(&spa_l2cache_lock); ASSERT(!vd->vdev_isl2cache); spa_aux_add(vd, &spa_l2cache_avl); vd->vdev_isl2cache = B_TRUE; mutex_exit(&spa_l2cache_lock); } void spa_l2cache_remove(vdev_t *vd) { mutex_enter(&spa_l2cache_lock); ASSERT(vd->vdev_isl2cache); spa_aux_remove(vd, &spa_l2cache_avl); vd->vdev_isl2cache = B_FALSE; mutex_exit(&spa_l2cache_lock); } boolean_t spa_l2cache_exists(uint64_t guid, uint64_t *pool) { boolean_t found; mutex_enter(&spa_l2cache_lock); found = spa_aux_exists(guid, pool, NULL, &spa_l2cache_avl); mutex_exit(&spa_l2cache_lock); return (found); } void spa_l2cache_activate(vdev_t *vd) { mutex_enter(&spa_l2cache_lock); ASSERT(vd->vdev_isl2cache); spa_aux_activate(vd, &spa_l2cache_avl); mutex_exit(&spa_l2cache_lock); } /* * ========================================================================== * SPA vdev locking * ========================================================================== */ /* * Lock the given spa_t for the purpose of adding or removing a vdev. * Grabs the global spa_namespace_lock plus the spa config lock for writing. * It returns the next transaction group for the spa_t. */ uint64_t spa_vdev_enter(spa_t *spa) { mutex_enter(&spa->spa_vdev_top_lock); mutex_enter(&spa_namespace_lock); vdev_autotrim_stop_all(spa); return (spa_vdev_config_enter(spa)); } /* * The same as spa_vdev_enter() above but additionally takes the guid of * the vdev being detached. When there is a rebuild in process it will be * suspended while the vdev tree is modified then resumed by spa_vdev_exit(). * The rebuild is canceled if only a single child remains after the detach. */ uint64_t spa_vdev_detach_enter(spa_t *spa, uint64_t guid) { mutex_enter(&spa->spa_vdev_top_lock); mutex_enter(&spa_namespace_lock); vdev_autotrim_stop_all(spa); if (guid != 0) { vdev_t *vd = spa_lookup_by_guid(spa, guid, B_FALSE); if (vd) { vdev_rebuild_stop_wait(vd->vdev_top); } } return (spa_vdev_config_enter(spa)); } /* * Internal implementation for spa_vdev_enter(). Used when a vdev * operation requires multiple syncs (i.e. removing a device) while * keeping the spa_namespace_lock held. */ uint64_t spa_vdev_config_enter(spa_t *spa) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); spa_config_enter(spa, SCL_ALL, spa, RW_WRITER); return (spa_last_synced_txg(spa) + 1); } /* * Used in combination with spa_vdev_config_enter() to allow the syncing * of multiple transactions without releasing the spa_namespace_lock. */ void spa_vdev_config_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error, char *tag) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); int config_changed = B_FALSE; ASSERT(txg > spa_last_synced_txg(spa)); spa->spa_pending_vdev = NULL; /* * Reassess the DTLs. */ vdev_dtl_reassess(spa->spa_root_vdev, 0, 0, B_FALSE, B_FALSE); if (error == 0 && !list_is_empty(&spa->spa_config_dirty_list)) { config_changed = B_TRUE; spa->spa_config_generation++; } /* * Verify the metaslab classes. */ ASSERT(metaslab_class_validate(spa_normal_class(spa)) == 0); ASSERT(metaslab_class_validate(spa_log_class(spa)) == 0); ASSERT(metaslab_class_validate(spa_embedded_log_class(spa)) == 0); ASSERT(metaslab_class_validate(spa_special_class(spa)) == 0); ASSERT(metaslab_class_validate(spa_dedup_class(spa)) == 0); spa_config_exit(spa, SCL_ALL, spa); /* * Panic the system if the specified tag requires it. This * is useful for ensuring that configurations are updated * transactionally. */ if (zio_injection_enabled) zio_handle_panic_injection(spa, tag, 0); /* * Note: this txg_wait_synced() is important because it ensures * that there won't be more than one config change per txg. * This allows us to use the txg as the generation number. */ if (error == 0) txg_wait_synced(spa->spa_dsl_pool, txg); if (vd != NULL) { ASSERT(!vd->vdev_detached || vd->vdev_dtl_sm == NULL); if (vd->vdev_ops->vdev_op_leaf) { mutex_enter(&vd->vdev_initialize_lock); vdev_initialize_stop(vd, VDEV_INITIALIZE_CANCELED, NULL); mutex_exit(&vd->vdev_initialize_lock); mutex_enter(&vd->vdev_trim_lock); vdev_trim_stop(vd, VDEV_TRIM_CANCELED, NULL); mutex_exit(&vd->vdev_trim_lock); } /* * The vdev may be both a leaf and top-level device. */ vdev_autotrim_stop_wait(vd); spa_config_enter(spa, SCL_STATE_ALL, spa, RW_WRITER); vdev_free(vd); spa_config_exit(spa, SCL_STATE_ALL, spa); } /* * If the config changed, update the config cache. */ if (config_changed) spa_write_cachefile(spa, B_FALSE, B_TRUE, B_TRUE); } /* * Unlock the spa_t after adding or removing a vdev. Besides undoing the * locking of spa_vdev_enter(), we also want make sure the transactions have * synced to disk, and then update the global configuration cache with the new * information. */ int spa_vdev_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error) { vdev_autotrim_restart(spa); vdev_rebuild_restart(spa); spa_vdev_config_exit(spa, vd, txg, error, FTAG); mutex_exit(&spa_namespace_lock); mutex_exit(&spa->spa_vdev_top_lock); return (error); } /* * Lock the given spa_t for the purpose of changing vdev state. */ void spa_vdev_state_enter(spa_t *spa, int oplocks) { int locks = SCL_STATE_ALL | oplocks; /* * Root pools may need to read of the underlying devfs filesystem * when opening up a vdev. Unfortunately if we're holding the * SCL_ZIO lock it will result in a deadlock when we try to issue * the read from the root filesystem. Instead we "prefetch" * the associated vnodes that we need prior to opening the * underlying devices and cache them so that we can prevent * any I/O when we are doing the actual open. */ if (spa_is_root(spa)) { int low = locks & ~(SCL_ZIO - 1); int high = locks & ~low; spa_config_enter(spa, high, spa, RW_WRITER); vdev_hold(spa->spa_root_vdev); spa_config_enter(spa, low, spa, RW_WRITER); } else { spa_config_enter(spa, locks, spa, RW_WRITER); } spa->spa_vdev_locks = locks; } int spa_vdev_state_exit(spa_t *spa, vdev_t *vd, int error) { boolean_t config_changed = B_FALSE; vdev_t *vdev_top; if (vd == NULL || vd == spa->spa_root_vdev) { vdev_top = spa->spa_root_vdev; } else { vdev_top = vd->vdev_top; } if (vd != NULL || error == 0) vdev_dtl_reassess(vdev_top, 0, 0, B_FALSE, B_FALSE); if (vd != NULL) { if (vd != spa->spa_root_vdev) vdev_state_dirty(vdev_top); config_changed = B_TRUE; spa->spa_config_generation++; } if (spa_is_root(spa)) vdev_rele(spa->spa_root_vdev); ASSERT3U(spa->spa_vdev_locks, >=, SCL_STATE_ALL); spa_config_exit(spa, spa->spa_vdev_locks, spa); /* * If anything changed, wait for it to sync. This ensures that, * from the system administrator's perspective, zpool(8) commands * are synchronous. This is important for things like zpool offline: * when the command completes, you expect no further I/O from ZFS. */ if (vd != NULL) txg_wait_synced(spa->spa_dsl_pool, 0); /* * If the config changed, update the config cache. */ if (config_changed) { mutex_enter(&spa_namespace_lock); spa_write_cachefile(spa, B_FALSE, B_TRUE, B_FALSE); mutex_exit(&spa_namespace_lock); } return (error); } /* * ========================================================================== * Miscellaneous functions * ========================================================================== */ void spa_activate_mos_feature(spa_t *spa, const char *feature, dmu_tx_t *tx) { if (!nvlist_exists(spa->spa_label_features, feature)) { fnvlist_add_boolean(spa->spa_label_features, feature); /* * When we are creating the pool (tx_txg==TXG_INITIAL), we can't * dirty the vdev config because lock SCL_CONFIG is not held. * Thankfully, in this case we don't need to dirty the config * because it will be written out anyway when we finish * creating the pool. */ if (tx->tx_txg != TXG_INITIAL) vdev_config_dirty(spa->spa_root_vdev); } } void spa_deactivate_mos_feature(spa_t *spa, const char *feature) { if (nvlist_remove_all(spa->spa_label_features, feature) == 0) vdev_config_dirty(spa->spa_root_vdev); } /* * Return the spa_t associated with given pool_guid, if it exists. If * device_guid is non-zero, determine whether the pool exists *and* contains * a device with the specified device_guid. */ spa_t * spa_by_guid(uint64_t pool_guid, uint64_t device_guid) { spa_t *spa; avl_tree_t *t = &spa_namespace_avl; ASSERT(MUTEX_HELD(&spa_namespace_lock)); for (spa = avl_first(t); spa != NULL; spa = AVL_NEXT(t, spa)) { if (spa->spa_state == POOL_STATE_UNINITIALIZED) continue; if (spa->spa_root_vdev == NULL) continue; if (spa_guid(spa) == pool_guid) { if (device_guid == 0) break; if (vdev_lookup_by_guid(spa->spa_root_vdev, device_guid) != NULL) break; /* * Check any devices we may be in the process of adding. */ if (spa->spa_pending_vdev) { if (vdev_lookup_by_guid(spa->spa_pending_vdev, device_guid) != NULL) break; } } } return (spa); } /* * Determine whether a pool with the given pool_guid exists. */ boolean_t spa_guid_exists(uint64_t pool_guid, uint64_t device_guid) { return (spa_by_guid(pool_guid, device_guid) != NULL); } char * spa_strdup(const char *s) { size_t len; char *new; len = strlen(s); new = kmem_alloc(len + 1, KM_SLEEP); bcopy(s, new, len); new[len] = '\0'; return (new); } void spa_strfree(char *s) { kmem_free(s, strlen(s) + 1); } uint64_t spa_generate_guid(spa_t *spa) { uint64_t guid; if (spa != NULL) { do { (void) random_get_pseudo_bytes((void *)&guid, sizeof (guid)); } while (guid == 0 || spa_guid_exists(spa_guid(spa), guid)); } else { do { (void) random_get_pseudo_bytes((void *)&guid, sizeof (guid)); } while (guid == 0 || spa_guid_exists(guid, 0)); } return (guid); } void snprintf_blkptr(char *buf, size_t buflen, const blkptr_t *bp) { char type[256]; char *checksum = NULL; char *compress = NULL; if (bp != NULL) { if (BP_GET_TYPE(bp) & DMU_OT_NEWTYPE) { dmu_object_byteswap_t bswap = DMU_OT_BYTESWAP(BP_GET_TYPE(bp)); (void) snprintf(type, sizeof (type), "bswap %s %s", DMU_OT_IS_METADATA(BP_GET_TYPE(bp)) ? "metadata" : "data", dmu_ot_byteswap[bswap].ob_name); } else { (void) strlcpy(type, dmu_ot[BP_GET_TYPE(bp)].ot_name, sizeof (type)); } if (!BP_IS_EMBEDDED(bp)) { checksum = zio_checksum_table[BP_GET_CHECKSUM(bp)].ci_name; } compress = zio_compress_table[BP_GET_COMPRESS(bp)].ci_name; } SNPRINTF_BLKPTR(snprintf, ' ', buf, buflen, bp, type, checksum, compress); } void spa_freeze(spa_t *spa) { uint64_t freeze_txg = 0; spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); if (spa->spa_freeze_txg == UINT64_MAX) { freeze_txg = spa_last_synced_txg(spa) + TXG_SIZE; spa->spa_freeze_txg = freeze_txg; } spa_config_exit(spa, SCL_ALL, FTAG); if (freeze_txg != 0) txg_wait_synced(spa_get_dsl(spa), freeze_txg); } void zfs_panic_recover(const char *fmt, ...) { va_list adx; va_start(adx, fmt); vcmn_err(zfs_recover ? CE_WARN : CE_PANIC, fmt, adx); va_end(adx); } /* * This is a stripped-down version of strtoull, suitable only for converting * lowercase hexadecimal numbers that don't overflow. */ uint64_t zfs_strtonum(const char *str, char **nptr) { uint64_t val = 0; char c; int digit; while ((c = *str) != '\0') { if (c >= '0' && c <= '9') digit = c - '0'; else if (c >= 'a' && c <= 'f') digit = 10 + c - 'a'; else break; val *= 16; val += digit; str++; } if (nptr) *nptr = (char *)str; return (val); } void spa_activate_allocation_classes(spa_t *spa, dmu_tx_t *tx) { /* * We bump the feature refcount for each special vdev added to the pool */ ASSERT(spa_feature_is_enabled(spa, SPA_FEATURE_ALLOCATION_CLASSES)); spa_feature_incr(spa, SPA_FEATURE_ALLOCATION_CLASSES, tx); } /* * ========================================================================== * Accessor functions * ========================================================================== */ boolean_t spa_shutting_down(spa_t *spa) { return (spa->spa_async_suspended); } dsl_pool_t * spa_get_dsl(spa_t *spa) { return (spa->spa_dsl_pool); } boolean_t spa_is_initializing(spa_t *spa) { return (spa->spa_is_initializing); } boolean_t spa_indirect_vdevs_loaded(spa_t *spa) { return (spa->spa_indirect_vdevs_loaded); } blkptr_t * spa_get_rootblkptr(spa_t *spa) { return (&spa->spa_ubsync.ub_rootbp); } void spa_set_rootblkptr(spa_t *spa, const blkptr_t *bp) { spa->spa_uberblock.ub_rootbp = *bp; } void spa_altroot(spa_t *spa, char *buf, size_t buflen) { if (spa->spa_root == NULL) buf[0] = '\0'; else (void) strncpy(buf, spa->spa_root, buflen); } int spa_sync_pass(spa_t *spa) { return (spa->spa_sync_pass); } char * spa_name(spa_t *spa) { return (spa->spa_name); } uint64_t spa_guid(spa_t *spa) { dsl_pool_t *dp = spa_get_dsl(spa); uint64_t guid; /* * If we fail to parse the config during spa_load(), we can go through * the error path (which posts an ereport) and end up here with no root * vdev. We stash the original pool guid in 'spa_config_guid' to handle * this case. */ if (spa->spa_root_vdev == NULL) return (spa->spa_config_guid); guid = spa->spa_last_synced_guid != 0 ? spa->spa_last_synced_guid : spa->spa_root_vdev->vdev_guid; /* * Return the most recently synced out guid unless we're * in syncing context. */ if (dp && dsl_pool_sync_context(dp)) return (spa->spa_root_vdev->vdev_guid); else return (guid); } uint64_t spa_load_guid(spa_t *spa) { /* * This is a GUID that exists solely as a reference for the * purposes of the arc. It is generated at load time, and * is never written to persistent storage. */ return (spa->spa_load_guid); } uint64_t spa_last_synced_txg(spa_t *spa) { return (spa->spa_ubsync.ub_txg); } uint64_t spa_first_txg(spa_t *spa) { return (spa->spa_first_txg); } uint64_t spa_syncing_txg(spa_t *spa) { return (spa->spa_syncing_txg); } /* * Return the last txg where data can be dirtied. The final txgs * will be used to just clear out any deferred frees that remain. */ uint64_t spa_final_dirty_txg(spa_t *spa) { return (spa->spa_final_txg - TXG_DEFER_SIZE); } pool_state_t spa_state(spa_t *spa) { return (spa->spa_state); } spa_load_state_t spa_load_state(spa_t *spa) { return (spa->spa_load_state); } uint64_t spa_freeze_txg(spa_t *spa) { return (spa->spa_freeze_txg); } /* * Return the inflated asize for a logical write in bytes. This is used by the * DMU to calculate the space a logical write will require on disk. * If lsize is smaller than the largest physical block size allocatable on this * pool we use its value instead, since the write will end up using the whole * block anyway. */ uint64_t spa_get_worst_case_asize(spa_t *spa, uint64_t lsize) { if (lsize == 0) return (0); /* No inflation needed */ return (MAX(lsize, 1 << spa->spa_max_ashift) * spa_asize_inflation); } /* * Return the amount of slop space in bytes. It is typically 1/32 of the pool * (3.2%), minus the embedded log space. On very small pools, it may be * slightly larger than this. On very large pools, it will be capped to * the value of spa_max_slop. The embedded log space is not included in * spa_dspace. By subtracting it, the usable space (per "zfs list") is a * constant 97% of the total space, regardless of metaslab size (assuming the * default spa_slop_shift=5 and a non-tiny pool). * * See the comment above spa_slop_shift for more details. */ uint64_t spa_get_slop_space(spa_t *spa) { uint64_t space = 0; uint64_t slop = 0; /* * Make sure spa_dedup_dspace has been set. */ if (spa->spa_dedup_dspace == ~0ULL) spa_update_dspace(spa); /* * spa_get_dspace() includes the space only logically "used" by * deduplicated data, so since it's not useful to reserve more * space with more deduplicated data, we subtract that out here. */ space = spa_get_dspace(spa) - spa->spa_dedup_dspace; slop = MIN(space >> spa_slop_shift, spa_max_slop); /* * Subtract the embedded log space, but no more than half the (3.2%) * unusable space. Note, the "no more than half" is only relevant if * zfs_embedded_slog_min_ms >> spa_slop_shift < 2, which is not true by * default. */ uint64_t embedded_log = metaslab_class_get_dspace(spa_embedded_log_class(spa)); slop -= MIN(embedded_log, slop >> 1); /* * Slop space should be at least spa_min_slop, but no more than half * the entire pool. */ slop = MAX(slop, MIN(space >> 1, spa_min_slop)); return (slop); } uint64_t spa_get_dspace(spa_t *spa) { return (spa->spa_dspace); } uint64_t spa_get_checkpoint_space(spa_t *spa) { return (spa->spa_checkpoint_info.sci_dspace); } void spa_update_dspace(spa_t *spa) { spa->spa_dspace = metaslab_class_get_dspace(spa_normal_class(spa)) + ddt_get_dedup_dspace(spa); if (spa->spa_vdev_removal != NULL) { /* * We can't allocate from the removing device, so subtract * its size if it was included in dspace (i.e. if this is a * normal-class vdev, not special/dedup). This prevents the * DMU/DSL from filling up the (now smaller) pool while we * are in the middle of removing the device. * * Note that the DMU/DSL doesn't actually know or care * how much space is allocated (it does its own tracking * of how much space has been logically used). So it * doesn't matter that the data we are moving may be * allocated twice (on the old device and the new * device). */ spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); vdev_t *vd = vdev_lookup_top(spa, spa->spa_vdev_removal->svr_vdev_id); /* * If the stars align, we can wind up here after * vdev_remove_complete() has cleared vd->vdev_mg but before * spa->spa_vdev_removal gets cleared, so we must check before * we dereference. */ if (vd->vdev_mg && vd->vdev_mg->mg_class == spa_normal_class(spa)) { spa->spa_dspace -= spa_deflate(spa) ? vd->vdev_stat.vs_dspace : vd->vdev_stat.vs_space; } spa_config_exit(spa, SCL_VDEV, FTAG); } } /* * Return the failure mode that has been set to this pool. The default * behavior will be to block all I/Os when a complete failure occurs. */ uint64_t spa_get_failmode(spa_t *spa) { return (spa->spa_failmode); } boolean_t spa_suspended(spa_t *spa) { return (spa->spa_suspended != ZIO_SUSPEND_NONE); } uint64_t spa_version(spa_t *spa) { return (spa->spa_ubsync.ub_version); } boolean_t spa_deflate(spa_t *spa) { return (spa->spa_deflate); } metaslab_class_t * spa_normal_class(spa_t *spa) { return (spa->spa_normal_class); } metaslab_class_t * spa_log_class(spa_t *spa) { return (spa->spa_log_class); } metaslab_class_t * spa_embedded_log_class(spa_t *spa) { return (spa->spa_embedded_log_class); } metaslab_class_t * spa_special_class(spa_t *spa) { return (spa->spa_special_class); } metaslab_class_t * spa_dedup_class(spa_t *spa) { return (spa->spa_dedup_class); } /* * Locate an appropriate allocation class */ metaslab_class_t * spa_preferred_class(spa_t *spa, uint64_t size, dmu_object_type_t objtype, uint_t level, uint_t special_smallblk) { /* * ZIL allocations determine their class in zio_alloc_zil(). */ ASSERT(objtype != DMU_OT_INTENT_LOG); boolean_t has_special_class = spa->spa_special_class->mc_groups != 0; if (DMU_OT_IS_DDT(objtype)) { if (spa->spa_dedup_class->mc_groups != 0) return (spa_dedup_class(spa)); else if (has_special_class && zfs_ddt_data_is_special) return (spa_special_class(spa)); else return (spa_normal_class(spa)); } /* Indirect blocks for user data can land in special if allowed */ if (level > 0 && (DMU_OT_IS_FILE(objtype) || objtype == DMU_OT_ZVOL)) { if (has_special_class && zfs_user_indirect_is_special) return (spa_special_class(spa)); else return (spa_normal_class(spa)); } if (DMU_OT_IS_METADATA(objtype) || level > 0) { if (has_special_class) return (spa_special_class(spa)); else return (spa_normal_class(spa)); } /* * Allow small file blocks in special class in some cases (like * for the dRAID vdev feature). But always leave a reserve of * zfs_special_class_metadata_reserve_pct exclusively for metadata. */ if (DMU_OT_IS_FILE(objtype) && has_special_class && size <= special_smallblk) { metaslab_class_t *special = spa_special_class(spa); uint64_t alloc = metaslab_class_get_alloc(special); uint64_t space = metaslab_class_get_space(special); uint64_t limit = (space * (100 - zfs_special_class_metadata_reserve_pct)) / 100; if (alloc < limit) return (special); } return (spa_normal_class(spa)); } void spa_evicting_os_register(spa_t *spa, objset_t *os) { mutex_enter(&spa->spa_evicting_os_lock); list_insert_head(&spa->spa_evicting_os_list, os); mutex_exit(&spa->spa_evicting_os_lock); } void spa_evicting_os_deregister(spa_t *spa, objset_t *os) { mutex_enter(&spa->spa_evicting_os_lock); list_remove(&spa->spa_evicting_os_list, os); cv_broadcast(&spa->spa_evicting_os_cv); mutex_exit(&spa->spa_evicting_os_lock); } void spa_evicting_os_wait(spa_t *spa) { mutex_enter(&spa->spa_evicting_os_lock); while (!list_is_empty(&spa->spa_evicting_os_list)) cv_wait(&spa->spa_evicting_os_cv, &spa->spa_evicting_os_lock); mutex_exit(&spa->spa_evicting_os_lock); dmu_buf_user_evict_wait(); } int spa_max_replication(spa_t *spa) { /* * As of SPA_VERSION == SPA_VERSION_DITTO_BLOCKS, we are able to * handle BPs with more than one DVA allocated. Set our max * replication level accordingly. */ if (spa_version(spa) < SPA_VERSION_DITTO_BLOCKS) return (1); return (MIN(SPA_DVAS_PER_BP, spa_max_replication_override)); } int spa_prev_software_version(spa_t *spa) { return (spa->spa_prev_software_version); } uint64_t spa_deadman_synctime(spa_t *spa) { return (spa->spa_deadman_synctime); } spa_autotrim_t spa_get_autotrim(spa_t *spa) { return (spa->spa_autotrim); } uint64_t spa_deadman_ziotime(spa_t *spa) { return (spa->spa_deadman_ziotime); } uint64_t spa_get_deadman_failmode(spa_t *spa) { return (spa->spa_deadman_failmode); } void spa_set_deadman_failmode(spa_t *spa, const char *failmode) { if (strcmp(failmode, "wait") == 0) spa->spa_deadman_failmode = ZIO_FAILURE_MODE_WAIT; else if (strcmp(failmode, "continue") == 0) spa->spa_deadman_failmode = ZIO_FAILURE_MODE_CONTINUE; else if (strcmp(failmode, "panic") == 0) spa->spa_deadman_failmode = ZIO_FAILURE_MODE_PANIC; else spa->spa_deadman_failmode = ZIO_FAILURE_MODE_WAIT; } void spa_set_deadman_ziotime(hrtime_t ns) { spa_t *spa = NULL; if (spa_mode_global != SPA_MODE_UNINIT) { mutex_enter(&spa_namespace_lock); while ((spa = spa_next(spa)) != NULL) spa->spa_deadman_ziotime = ns; mutex_exit(&spa_namespace_lock); } } void spa_set_deadman_synctime(hrtime_t ns) { spa_t *spa = NULL; if (spa_mode_global != SPA_MODE_UNINIT) { mutex_enter(&spa_namespace_lock); while ((spa = spa_next(spa)) != NULL) spa->spa_deadman_synctime = ns; mutex_exit(&spa_namespace_lock); } } uint64_t dva_get_dsize_sync(spa_t *spa, const dva_t *dva) { uint64_t asize = DVA_GET_ASIZE(dva); uint64_t dsize = asize; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); if (asize != 0 && spa->spa_deflate) { vdev_t *vd = vdev_lookup_top(spa, DVA_GET_VDEV(dva)); if (vd != NULL) dsize = (asize >> SPA_MINBLOCKSHIFT) * vd->vdev_deflate_ratio; } return (dsize); } uint64_t bp_get_dsize_sync(spa_t *spa, const blkptr_t *bp) { uint64_t dsize = 0; for (int d = 0; d < BP_GET_NDVAS(bp); d++) dsize += dva_get_dsize_sync(spa, &bp->blk_dva[d]); return (dsize); } uint64_t bp_get_dsize(spa_t *spa, const blkptr_t *bp) { uint64_t dsize = 0; spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); for (int d = 0; d < BP_GET_NDVAS(bp); d++) dsize += dva_get_dsize_sync(spa, &bp->blk_dva[d]); spa_config_exit(spa, SCL_VDEV, FTAG); return (dsize); } uint64_t spa_dirty_data(spa_t *spa) { return (spa->spa_dsl_pool->dp_dirty_total); } /* * ========================================================================== * SPA Import Progress Routines * ========================================================================== */ typedef struct spa_import_progress { uint64_t pool_guid; /* unique id for updates */ char *pool_name; spa_load_state_t spa_load_state; uint64_t mmp_sec_remaining; /* MMP activity check */ uint64_t spa_load_max_txg; /* rewind txg */ procfs_list_node_t smh_node; } spa_import_progress_t; spa_history_list_t *spa_import_progress_list = NULL; static int spa_import_progress_show_header(struct seq_file *f) { seq_printf(f, "%-20s %-14s %-14s %-12s %s\n", "pool_guid", "load_state", "multihost_secs", "max_txg", "pool_name"); return (0); } static int spa_import_progress_show(struct seq_file *f, void *data) { spa_import_progress_t *sip = (spa_import_progress_t *)data; seq_printf(f, "%-20llu %-14llu %-14llu %-12llu %s\n", (u_longlong_t)sip->pool_guid, (u_longlong_t)sip->spa_load_state, (u_longlong_t)sip->mmp_sec_remaining, (u_longlong_t)sip->spa_load_max_txg, (sip->pool_name ? sip->pool_name : "-")); return (0); } /* Remove oldest elements from list until there are no more than 'size' left */ static void spa_import_progress_truncate(spa_history_list_t *shl, unsigned int size) { spa_import_progress_t *sip; while (shl->size > size) { sip = list_remove_head(&shl->procfs_list.pl_list); if (sip->pool_name) spa_strfree(sip->pool_name); kmem_free(sip, sizeof (spa_import_progress_t)); shl->size--; } IMPLY(size == 0, list_is_empty(&shl->procfs_list.pl_list)); } static void spa_import_progress_init(void) { spa_import_progress_list = kmem_zalloc(sizeof (spa_history_list_t), KM_SLEEP); spa_import_progress_list->size = 0; spa_import_progress_list->procfs_list.pl_private = spa_import_progress_list; procfs_list_install("zfs", NULL, "import_progress", 0644, &spa_import_progress_list->procfs_list, spa_import_progress_show, spa_import_progress_show_header, NULL, offsetof(spa_import_progress_t, smh_node)); } static void spa_import_progress_destroy(void) { spa_history_list_t *shl = spa_import_progress_list; procfs_list_uninstall(&shl->procfs_list); spa_import_progress_truncate(shl, 0); procfs_list_destroy(&shl->procfs_list); kmem_free(shl, sizeof (spa_history_list_t)); } int spa_import_progress_set_state(uint64_t pool_guid, spa_load_state_t load_state) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; int error = ENOENT; if (shl->size == 0) return (0); mutex_enter(&shl->procfs_list.pl_lock); for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL; sip = list_prev(&shl->procfs_list.pl_list, sip)) { if (sip->pool_guid == pool_guid) { sip->spa_load_state = load_state; error = 0; break; } } mutex_exit(&shl->procfs_list.pl_lock); return (error); } int spa_import_progress_set_max_txg(uint64_t pool_guid, uint64_t load_max_txg) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; int error = ENOENT; if (shl->size == 0) return (0); mutex_enter(&shl->procfs_list.pl_lock); for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL; sip = list_prev(&shl->procfs_list.pl_list, sip)) { if (sip->pool_guid == pool_guid) { sip->spa_load_max_txg = load_max_txg; error = 0; break; } } mutex_exit(&shl->procfs_list.pl_lock); return (error); } int spa_import_progress_set_mmp_check(uint64_t pool_guid, uint64_t mmp_sec_remaining) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; int error = ENOENT; if (shl->size == 0) return (0); mutex_enter(&shl->procfs_list.pl_lock); for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL; sip = list_prev(&shl->procfs_list.pl_list, sip)) { if (sip->pool_guid == pool_guid) { sip->mmp_sec_remaining = mmp_sec_remaining; error = 0; break; } } mutex_exit(&shl->procfs_list.pl_lock); return (error); } /* * A new import is in progress, add an entry. */ void spa_import_progress_add(spa_t *spa) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; char *poolname = NULL; sip = kmem_zalloc(sizeof (spa_import_progress_t), KM_SLEEP); sip->pool_guid = spa_guid(spa); (void) nvlist_lookup_string(spa->spa_config, ZPOOL_CONFIG_POOL_NAME, &poolname); if (poolname == NULL) poolname = spa_name(spa); sip->pool_name = spa_strdup(poolname); sip->spa_load_state = spa_load_state(spa); mutex_enter(&shl->procfs_list.pl_lock); procfs_list_add(&shl->procfs_list, sip); shl->size++; mutex_exit(&shl->procfs_list.pl_lock); } void spa_import_progress_remove(uint64_t pool_guid) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; mutex_enter(&shl->procfs_list.pl_lock); for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL; sip = list_prev(&shl->procfs_list.pl_list, sip)) { if (sip->pool_guid == pool_guid) { if (sip->pool_name) spa_strfree(sip->pool_name); list_remove(&shl->procfs_list.pl_list, sip); shl->size--; kmem_free(sip, sizeof (spa_import_progress_t)); break; } } mutex_exit(&shl->procfs_list.pl_lock); } /* * ========================================================================== * Initialization and Termination * ========================================================================== */ static int spa_name_compare(const void *a1, const void *a2) { const spa_t *s1 = a1; const spa_t *s2 = a2; int s; s = strcmp(s1->spa_name, s2->spa_name); return (TREE_ISIGN(s)); } void spa_boot_init(void) { spa_config_load(); } void spa_init(spa_mode_t mode) { mutex_init(&spa_namespace_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa_spare_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa_l2cache_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&spa_namespace_cv, NULL, CV_DEFAULT, NULL); avl_create(&spa_namespace_avl, spa_name_compare, sizeof (spa_t), offsetof(spa_t, spa_avl)); avl_create(&spa_spare_avl, spa_spare_compare, sizeof (spa_aux_t), offsetof(spa_aux_t, aux_avl)); avl_create(&spa_l2cache_avl, spa_l2cache_compare, sizeof (spa_aux_t), offsetof(spa_aux_t, aux_avl)); spa_mode_global = mode; #ifndef _KERNEL if (spa_mode_global != SPA_MODE_READ && dprintf_find_string("watch")) { struct sigaction sa; sa.sa_flags = SA_SIGINFO; sigemptyset(&sa.sa_mask); sa.sa_sigaction = arc_buf_sigsegv; if (sigaction(SIGSEGV, &sa, NULL) == -1) { perror("could not enable watchpoints: " "sigaction(SIGSEGV, ...) = "); } else { arc_watch = B_TRUE; } } #endif fm_init(); zfs_refcount_init(); unique_init(); zfs_btree_init(); metaslab_stat_init(); ddt_init(); zio_init(); dmu_init(); zil_init(); vdev_cache_stat_init(); vdev_mirror_stat_init(); vdev_raidz_math_init(); vdev_file_init(); zfs_prop_init(); zpool_prop_init(); zpool_feature_init(); spa_config_load(); l2arc_start(); scan_init(); qat_init(); spa_import_progress_init(); } void spa_fini(void) { l2arc_stop(); spa_evict_all(); vdev_file_fini(); vdev_cache_stat_fini(); vdev_mirror_stat_fini(); vdev_raidz_math_fini(); zil_fini(); dmu_fini(); zio_fini(); ddt_fini(); metaslab_stat_fini(); zfs_btree_fini(); unique_fini(); zfs_refcount_fini(); fm_fini(); scan_fini(); qat_fini(); spa_import_progress_destroy(); avl_destroy(&spa_namespace_avl); avl_destroy(&spa_spare_avl); avl_destroy(&spa_l2cache_avl); cv_destroy(&spa_namespace_cv); mutex_destroy(&spa_namespace_lock); mutex_destroy(&spa_spare_lock); mutex_destroy(&spa_l2cache_lock); } /* * Return whether this pool has a dedicated slog device. No locking needed. * It's not a problem if the wrong answer is returned as it's only for * performance and not correctness. */ boolean_t spa_has_slogs(spa_t *spa) { return (spa->spa_log_class->mc_groups != 0); } spa_log_state_t spa_get_log_state(spa_t *spa) { return (spa->spa_log_state); } void spa_set_log_state(spa_t *spa, spa_log_state_t state) { spa->spa_log_state = state; } boolean_t spa_is_root(spa_t *spa) { return (spa->spa_is_root); } boolean_t spa_writeable(spa_t *spa) { return (!!(spa->spa_mode & SPA_MODE_WRITE) && spa->spa_trust_config); } /* * Returns true if there is a pending sync task in any of the current * syncing txg, the current quiescing txg, or the current open txg. */ boolean_t spa_has_pending_synctask(spa_t *spa) { return (!txg_all_lists_empty(&spa->spa_dsl_pool->dp_sync_tasks) || !txg_all_lists_empty(&spa->spa_dsl_pool->dp_early_sync_tasks)); } spa_mode_t spa_mode(spa_t *spa) { return (spa->spa_mode); } uint64_t spa_bootfs(spa_t *spa) { return (spa->spa_bootfs); } uint64_t spa_delegation(spa_t *spa) { return (spa->spa_delegation); } objset_t * spa_meta_objset(spa_t *spa) { return (spa->spa_meta_objset); } enum zio_checksum spa_dedup_checksum(spa_t *spa) { return (spa->spa_dedup_checksum); } /* * Reset pool scan stat per scan pass (or reboot). */ void spa_scan_stat_init(spa_t *spa) { /* data not stored on disk */ spa->spa_scan_pass_start = gethrestime_sec(); if (dsl_scan_is_paused_scrub(spa->spa_dsl_pool->dp_scan)) spa->spa_scan_pass_scrub_pause = spa->spa_scan_pass_start; else spa->spa_scan_pass_scrub_pause = 0; spa->spa_scan_pass_scrub_spent_paused = 0; spa->spa_scan_pass_exam = 0; spa->spa_scan_pass_issued = 0; - vdev_scan_stat_init(spa->spa_root_vdev); } /* * Get scan stats for zpool status reports */ int spa_scan_get_stats(spa_t *spa, pool_scan_stat_t *ps) { dsl_scan_t *scn = spa->spa_dsl_pool ? spa->spa_dsl_pool->dp_scan : NULL; if (scn == NULL || scn->scn_phys.scn_func == POOL_SCAN_NONE) return (SET_ERROR(ENOENT)); bzero(ps, sizeof (pool_scan_stat_t)); /* data stored on disk */ ps->pss_func = scn->scn_phys.scn_func; ps->pss_state = scn->scn_phys.scn_state; ps->pss_start_time = scn->scn_phys.scn_start_time; ps->pss_end_time = scn->scn_phys.scn_end_time; ps->pss_to_examine = scn->scn_phys.scn_to_examine; ps->pss_examined = scn->scn_phys.scn_examined; ps->pss_to_process = scn->scn_phys.scn_to_process; ps->pss_processed = scn->scn_phys.scn_processed; ps->pss_errors = scn->scn_phys.scn_errors; /* data not stored on disk */ ps->pss_pass_exam = spa->spa_scan_pass_exam; ps->pss_pass_start = spa->spa_scan_pass_start; ps->pss_pass_scrub_pause = spa->spa_scan_pass_scrub_pause; ps->pss_pass_scrub_spent_paused = spa->spa_scan_pass_scrub_spent_paused; ps->pss_pass_issued = spa->spa_scan_pass_issued; ps->pss_issued = scn->scn_issued_before_pass + spa->spa_scan_pass_issued; return (0); } int spa_maxblocksize(spa_t *spa) { if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_BLOCKS)) return (SPA_MAXBLOCKSIZE); else return (SPA_OLD_MAXBLOCKSIZE); } /* * Returns the txg that the last device removal completed. No indirect mappings * have been added since this txg. */ uint64_t spa_get_last_removal_txg(spa_t *spa) { uint64_t vdevid; uint64_t ret = -1ULL; spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); /* * sr_prev_indirect_vdev is only modified while holding all the * config locks, so it is sufficient to hold SCL_VDEV as reader when * examining it. */ vdevid = spa->spa_removing_phys.sr_prev_indirect_vdev; while (vdevid != -1ULL) { vdev_t *vd = vdev_lookup_top(spa, vdevid); vdev_indirect_births_t *vib = vd->vdev_indirect_births; ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops); /* * If the removal did not remap any data, we don't care. */ if (vdev_indirect_births_count(vib) != 0) { ret = vdev_indirect_births_last_entry_txg(vib); break; } vdevid = vd->vdev_indirect_config.vic_prev_indirect_vdev; } spa_config_exit(spa, SCL_VDEV, FTAG); IMPLY(ret != -1ULL, spa_feature_is_active(spa, SPA_FEATURE_DEVICE_REMOVAL)); return (ret); } int spa_maxdnodesize(spa_t *spa) { if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_DNODE)) return (DNODE_MAX_SIZE); else return (DNODE_MIN_SIZE); } boolean_t spa_multihost(spa_t *spa) { return (spa->spa_multihost ? B_TRUE : B_FALSE); } uint32_t spa_get_hostid(spa_t *spa) { return (spa->spa_hostid); } boolean_t spa_trust_config(spa_t *spa) { return (spa->spa_trust_config); } uint64_t spa_missing_tvds_allowed(spa_t *spa) { return (spa->spa_missing_tvds_allowed); } space_map_t * spa_syncing_log_sm(spa_t *spa) { return (spa->spa_syncing_log_sm); } void spa_set_missing_tvds(spa_t *spa, uint64_t missing) { spa->spa_missing_tvds = missing; } /* * Return the pool state string ("ONLINE", "DEGRADED", "SUSPENDED", etc). */ const char * spa_state_to_name(spa_t *spa) { ASSERT3P(spa, !=, NULL); /* * it is possible for the spa to exist, without root vdev * as the spa transitions during import/export */ vdev_t *rvd = spa->spa_root_vdev; if (rvd == NULL) { return ("TRANSITIONING"); } vdev_state_t state = rvd->vdev_state; vdev_aux_t aux = rvd->vdev_stat.vs_aux; if (spa_suspended(spa) && (spa_get_failmode(spa) != ZIO_FAILURE_MODE_CONTINUE)) return ("SUSPENDED"); switch (state) { case VDEV_STATE_CLOSED: case VDEV_STATE_OFFLINE: return ("OFFLINE"); case VDEV_STATE_REMOVED: return ("REMOVED"); case VDEV_STATE_CANT_OPEN: if (aux == VDEV_AUX_CORRUPT_DATA || aux == VDEV_AUX_BAD_LOG) return ("FAULTED"); else if (aux == VDEV_AUX_SPLIT_POOL) return ("SPLIT"); else return ("UNAVAIL"); case VDEV_STATE_FAULTED: return ("FAULTED"); case VDEV_STATE_DEGRADED: return ("DEGRADED"); case VDEV_STATE_HEALTHY: return ("ONLINE"); default: break; } return ("UNKNOWN"); } boolean_t spa_top_vdevs_spacemap_addressable(spa_t *spa) { vdev_t *rvd = spa->spa_root_vdev; for (uint64_t c = 0; c < rvd->vdev_children; c++) { if (!vdev_is_spacemap_addressable(rvd->vdev_child[c])) return (B_FALSE); } return (B_TRUE); } boolean_t spa_has_checkpoint(spa_t *spa) { return (spa->spa_checkpoint_txg != 0); } boolean_t spa_importing_readonly_checkpoint(spa_t *spa) { return ((spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT) && spa->spa_mode == SPA_MODE_READ); } uint64_t spa_min_claim_txg(spa_t *spa) { uint64_t checkpoint_txg = spa->spa_uberblock.ub_checkpoint_txg; if (checkpoint_txg != 0) return (checkpoint_txg + 1); return (spa->spa_first_txg); } /* * If there is a checkpoint, async destroys may consume more space from * the pool instead of freeing it. In an attempt to save the pool from * getting suspended when it is about to run out of space, we stop * processing async destroys. */ boolean_t spa_suspend_async_destroy(spa_t *spa) { dsl_pool_t *dp = spa_get_dsl(spa); uint64_t unreserved = dsl_pool_unreserved_space(dp, ZFS_SPACE_CHECK_EXTRA_RESERVED); uint64_t used = dsl_dir_phys(dp->dp_root_dir)->dd_used_bytes; uint64_t avail = (unreserved > used) ? (unreserved - used) : 0; if (spa_has_checkpoint(spa) && avail == 0) return (B_TRUE); return (B_FALSE); } #if defined(_KERNEL) int param_set_deadman_failmode_common(const char *val) { spa_t *spa = NULL; char *p; if (val == NULL) return (SET_ERROR(EINVAL)); if ((p = strchr(val, '\n')) != NULL) *p = '\0'; if (strcmp(val, "wait") != 0 && strcmp(val, "continue") != 0 && strcmp(val, "panic")) return (SET_ERROR(EINVAL)); if (spa_mode_global != SPA_MODE_UNINIT) { mutex_enter(&spa_namespace_lock); while ((spa = spa_next(spa)) != NULL) spa_set_deadman_failmode(spa, val); mutex_exit(&spa_namespace_lock); } return (0); } #endif /* Namespace manipulation */ EXPORT_SYMBOL(spa_lookup); EXPORT_SYMBOL(spa_add); EXPORT_SYMBOL(spa_remove); EXPORT_SYMBOL(spa_next); /* Refcount functions */ EXPORT_SYMBOL(spa_open_ref); EXPORT_SYMBOL(spa_close); EXPORT_SYMBOL(spa_refcount_zero); /* Pool configuration lock */ EXPORT_SYMBOL(spa_config_tryenter); EXPORT_SYMBOL(spa_config_enter); EXPORT_SYMBOL(spa_config_exit); EXPORT_SYMBOL(spa_config_held); /* Pool vdev add/remove lock */ EXPORT_SYMBOL(spa_vdev_enter); EXPORT_SYMBOL(spa_vdev_exit); /* Pool vdev state change lock */ EXPORT_SYMBOL(spa_vdev_state_enter); EXPORT_SYMBOL(spa_vdev_state_exit); /* Accessor functions */ EXPORT_SYMBOL(spa_shutting_down); EXPORT_SYMBOL(spa_get_dsl); EXPORT_SYMBOL(spa_get_rootblkptr); EXPORT_SYMBOL(spa_set_rootblkptr); EXPORT_SYMBOL(spa_altroot); EXPORT_SYMBOL(spa_sync_pass); EXPORT_SYMBOL(spa_name); EXPORT_SYMBOL(spa_guid); EXPORT_SYMBOL(spa_last_synced_txg); EXPORT_SYMBOL(spa_first_txg); EXPORT_SYMBOL(spa_syncing_txg); EXPORT_SYMBOL(spa_version); EXPORT_SYMBOL(spa_state); EXPORT_SYMBOL(spa_load_state); EXPORT_SYMBOL(spa_freeze_txg); EXPORT_SYMBOL(spa_get_dspace); EXPORT_SYMBOL(spa_update_dspace); EXPORT_SYMBOL(spa_deflate); EXPORT_SYMBOL(spa_normal_class); EXPORT_SYMBOL(spa_log_class); EXPORT_SYMBOL(spa_special_class); EXPORT_SYMBOL(spa_preferred_class); EXPORT_SYMBOL(spa_max_replication); EXPORT_SYMBOL(spa_prev_software_version); EXPORT_SYMBOL(spa_get_failmode); EXPORT_SYMBOL(spa_suspended); EXPORT_SYMBOL(spa_bootfs); EXPORT_SYMBOL(spa_delegation); EXPORT_SYMBOL(spa_meta_objset); EXPORT_SYMBOL(spa_maxblocksize); EXPORT_SYMBOL(spa_maxdnodesize); /* Miscellaneous support routines */ EXPORT_SYMBOL(spa_guid_exists); EXPORT_SYMBOL(spa_strdup); EXPORT_SYMBOL(spa_strfree); EXPORT_SYMBOL(spa_generate_guid); EXPORT_SYMBOL(snprintf_blkptr); EXPORT_SYMBOL(spa_freeze); EXPORT_SYMBOL(spa_upgrade); EXPORT_SYMBOL(spa_evict_all); EXPORT_SYMBOL(spa_lookup_by_guid); EXPORT_SYMBOL(spa_has_spare); EXPORT_SYMBOL(dva_get_dsize_sync); EXPORT_SYMBOL(bp_get_dsize_sync); EXPORT_SYMBOL(bp_get_dsize); EXPORT_SYMBOL(spa_has_slogs); EXPORT_SYMBOL(spa_is_root); EXPORT_SYMBOL(spa_writeable); EXPORT_SYMBOL(spa_mode); EXPORT_SYMBOL(spa_namespace_lock); EXPORT_SYMBOL(spa_trust_config); EXPORT_SYMBOL(spa_missing_tvds_allowed); EXPORT_SYMBOL(spa_set_missing_tvds); EXPORT_SYMBOL(spa_state_to_name); EXPORT_SYMBOL(spa_importing_readonly_checkpoint); EXPORT_SYMBOL(spa_min_claim_txg); EXPORT_SYMBOL(spa_suspend_async_destroy); EXPORT_SYMBOL(spa_has_checkpoint); EXPORT_SYMBOL(spa_top_vdevs_spacemap_addressable); ZFS_MODULE_PARAM(zfs, zfs_, flags, UINT, ZMOD_RW, "Set additional debugging flags"); ZFS_MODULE_PARAM(zfs, zfs_, recover, INT, ZMOD_RW, "Set to attempt to recover from fatal errors"); ZFS_MODULE_PARAM(zfs, zfs_, free_leak_on_eio, INT, ZMOD_RW, "Set to ignore IO errors during free and permanently leak the space"); ZFS_MODULE_PARAM(zfs_deadman, zfs_deadman_, checktime_ms, ULONG, ZMOD_RW, "Dead I/O check interval in milliseconds"); ZFS_MODULE_PARAM(zfs_deadman, zfs_deadman_, enabled, INT, ZMOD_RW, "Enable deadman timer"); ZFS_MODULE_PARAM(zfs_spa, spa_, asize_inflation, INT, ZMOD_RW, "SPA size estimate multiplication factor"); ZFS_MODULE_PARAM(zfs, zfs_, ddt_data_is_special, INT, ZMOD_RW, "Place DDT data into the special class"); ZFS_MODULE_PARAM(zfs, zfs_, user_indirect_is_special, INT, ZMOD_RW, "Place user data indirect blocks into the special class"); /* BEGIN CSTYLED */ ZFS_MODULE_PARAM_CALL(zfs_deadman, zfs_deadman_, failmode, param_set_deadman_failmode, param_get_charp, ZMOD_RW, "Failmode for deadman timer"); ZFS_MODULE_PARAM_CALL(zfs_deadman, zfs_deadman_, synctime_ms, param_set_deadman_synctime, param_get_ulong, ZMOD_RW, "Pool sync expiration time in milliseconds"); ZFS_MODULE_PARAM_CALL(zfs_deadman, zfs_deadman_, ziotime_ms, param_set_deadman_ziotime, param_get_ulong, ZMOD_RW, "IO expiration time in milliseconds"); ZFS_MODULE_PARAM(zfs, zfs_, special_class_metadata_reserve_pct, INT, ZMOD_RW, "Small file blocks in special vdevs depends on this much " "free space available"); /* END CSTYLED */ ZFS_MODULE_PARAM_CALL(zfs_spa, spa_, slop_shift, param_set_slop_shift, param_get_int, ZMOD_RW, "Reserved free space in pool");