diff --git a/en_US.ISO8859-1/books/handbook/cutting-edge/chapter.sgml b/en_US.ISO8859-1/books/handbook/cutting-edge/chapter.sgml index db61b13068..0a9d5d2c14 100644 --- a/en_US.ISO8859-1/books/handbook/cutting-edge/chapter.sgml +++ b/en_US.ISO8859-1/books/handbook/cutting-edge/chapter.sgml @@ -1,1844 +1,1844 @@ Jim Mock Restructured, reorganized, and parts updated by Jordan Hubbard Original work by Poul-Henning Kamp John Polstra Nik Clayton The Cutting Edge Synopsis &os; is under constant development between releases. For people who want to be on the cutting edge, there are several easy mechanisms for keeping your system in sync with the latest developments. Be warned—the cutting edge is not for everyone! This chapter will help you decide if you want to track the development system, or stick with one of the released versions. After reading this chapter, you will know: The difference between the two development branches: &os.stable; and &os.current;. How to keep your system up to date with CVSup, CVS, or CTM. How to rebuild and reinstall the entire base system with make buildworld (etc). Before reading this chapter, you should: Properly set up your network connection (). Know how to install additional third-party software (). &os.current; vs. &os.stable; -CURRENT -STABLE There are two development branches to FreeBSD: &os.current; and &os.stable;. This section will explain a bit about each and describe how to keep your system up-to-date with each respective tree. &os.current; will be discussed first, then &os.stable;. Staying Current with &os; As you read this, keep in mind that &os.current; is the bleeding edge of &os; development. &os.current; users are expected to have a high degree of technical skill, and should be capable of solving difficult system problems on their own. If you are new to &os;, think twice before installing it. What Is &os.current;? snapshot &os.current; is the latest working sources for &os;. This includes work in progress, experimental changes, and transitional mechanisms that might or might not be present in the next official release of the software. While many &os; developers compile the &os.current; source code daily, there are periods of time when the sources are not buildable. These problems are resolved as expeditiously as possible, but whether or not &os.current; brings disaster or greatly desired functionality can be a matter of which exact moment you grabbed the source code in! Who Needs &os.current;? &os.current; is made available for 3 primary interest groups: Members of the &os; community who are actively working on some part of the source tree and for whom keeping current is an absolute requirement. Members of the &os; community who are active testers, willing to spend time solving problems in order to ensure that &os.current; remains as sane as possible. These are also people who wish to make topical suggestions on changes and the general direction of &os;, and submit patches to implement them. Those who merely wish to keep an eye on things, or to use the current sources for reference purposes (e.g. for reading, not running). These people also make the occasional comment or contribute code. What Is &os.current; <emphasis>Not</emphasis>? A fast-track to getting pre-release bits because you heard there is some cool new feature in there and you want to be the first on your block to have it. Being the first on the block to get the new feature means that you're the first on the block to get the new bugs. A quick way of getting bug fixes. Any given version of &os.current; is just as likely to introduce new bugs as to fix existing ones. In any way officially supported. We do our best to help people genuinely in one of the 3 legitimate &os.current; groups, but we simply do not have the time to provide tech support. This is not because we are mean and nasty people who do not like helping people out (we would not even be doing &os; if we were). We simply cannot answer hundreds messages a day and work on FreeBSD! Given the choice between improving &os; and answering lots of questions on experimental code, the developers opt for the former. Using &os.current; -CURRENT using Join the &a.current.name; and the &a.cvsall.name; lists. This is not just a good idea, it is essential. If you are not on the &a.current.name; list, you will not see the comments that people are making about the current state of the system and thus will probably end up stumbling over a lot of problems that others have already found and solved. Even more importantly, you will miss out on important bulletins which may be critical to your system's continued health. The &a.cvsall.name; list will allow you to see the commit log entry for each change as it is made along with any pertinent information on possible side-effects. To join these lists, or one of the others available go to &a.mailman.lists.link; and click on the list that you wish to subscribe to. Instructions on the rest of the procedure are available there. Grab the sources from a &os; mirror site. You can do this in one of two ways: cvsup cron -CURRENT Syncing with CVSup Use the cvsup program with the supfile named standard-supfile available from /usr/share/examples/cvsup. This is the most recommended method, since it allows you to grab the entire collection once and then only what has changed from then on. Many people run cvsup from cron and keep their sources up-to-date automatically. You have to customize the sample supfile above, and configure cvsup for your environment. -CURRENT Syncing with CTM Use the CTM facility. If you have very bad connectivity (high price connections or only email access) CTM is an option. However, it is a lot of hassle and can give you broken files. This leads to it being rarely used, which again increases the chance of it not working for fairly long periods of time. We recommend using CVSup for anybody with a 9600 bps modem or faster connection. If you are grabbing the sources to run, and not just look at, then grab all of &os.current;, not just selected portions. The reason for this is that various parts of the source depend on updates elsewhere, and trying to compile just a subset is almost guaranteed to get you into trouble. -CURRENT compiling Before compiling &os.current;, read the Makefile in /usr/src carefully. You should at least install a new kernel and rebuild the world the first time through as part of the upgrading process. Reading the &a.current; and /usr/src/UPDATING will keep you up-to-date on other bootstrapping procedures that sometimes become necessary as we move toward the next release. Be active! If you are running &os.current;, we want to know what you have to say about it, especially if you have suggestions for enhancements or bug fixes. Suggestions with accompanying code are received most enthusiastically! Staying Stable with &os; What Is &os.stable;? -STABLE &os.stable; is our development branch from which major releases are made. Changes go into this branch at a different pace, and with the general assumption that they have first gone into &os.current; for testing. This is still a development branch, however, and this means that at any given time, the sources for &os.stable; may or may not be suitable for any particular purpose. It is simply another engineering development track, not a resource for end-users. Who Needs &os.stable;? If you are interested in tracking or contributing to the FreeBSD development process, especially as it relates to the next point release of FreeBSD, then you should consider following &os.stable;. While it is true that security fixes also go into the &os.stable; branch, you do not need to track &os.stable; to do this. Every security advisory for FreeBSD explains how to fix the problem for the releases it affects That is not quite true. We can not continue to support old releases of FreeBSD forever, although we do support them for many years. For a complete description of the current security policy for old releases of FreeBSD, please see http://www.FreeBSD.org/security/. , and tracking an entire development branch just for security reasons is likely to bring in a lot of unwanted changes as well. Although we endeavor to ensure that the &os.stable; branch compiles and runs at all times, this cannot be guaranteed. In addition, while code is developed in &os.current; before including it in &os.stable;, more people run &os.stable; than &os.current;, so it is inevitable that bugs and corner cases will sometimes be found in &os.stable; that were not apparent in &os.current;. For these reasons, we do not recommend that you blindly track &os.stable;, and it is particularly important that you do not update any production servers to &os.stable; without first thoroughly testing the code in your development environment. If you do not have the resources to do this then we recommend that you run the most recent release of FreeBSD, and use the binary update mechanism to move from release to release. Using &os.stable; -STABLE using Join the &a.stable.name; list. This will keep you informed of build-dependencies that may appear in &os.stable; or any other issues requiring special attention. Developers will also make announcements in this mailing list when they are contemplating some controversial fix or update, giving the users a chance to respond if they have any issues to raise concerning the proposed change. The &a.cvsall.name; list will allow you to see the commit log entry for each change as it is made along with any pertinent information on possible side-effects. To join these lists, or one of the others available go to &a.mailman.lists.link; and click on the list that you wish to subscribe to. Instructions on the rest of the procedure are available there. If you are installing a new system and want it to be as stable as possible, you can simply grab the latest dated branch snapshot from and install it like any other release. Or you can install the most recent &os.stable; release from the mirror sites and follow the instructions below to upgrade your system to the most up to date &os.stable; source code. If you are already running a previous release of &os; and wish to upgrade via sources then you can easily do so from &os; mirror site. This can be done in one of two ways: cvsup cron -STABLE syncing with CVSup Use the cvsup program with the supfile named stable-supfile from the directory /usr/share/examples/cvsup. This is the most recommended method, since it allows you to grab the entire collection once and then only what has changed from then on. Many people run cvsup from cron to keep their sources up-to-date automatically. You have to customize the sample supfile above, and configure cvsup for your environment. -STABLE syncing with CTM Use the CTM facility. If you do not have a fast and inexpensive connection to the Internet, this is the method you should consider using. Essentially, if you need rapid on-demand access to the source and communications bandwidth is not a consideration, use cvsup or ftp. Otherwise, use CTM. -STABLE compiling Before compiling &os.stable;, read the Makefile in /usr/src carefully. You should at least install a new kernel and rebuild the world the first time through as part of the upgrading process. Reading the &a.stable; and /usr/src/UPDATING will keep you up-to-date on other bootstrapping procedures that sometimes become necessary as we move toward the next release. Synchronizing Your Source There are various ways of using an Internet (or email) connection to stay up-to-date with any given area of the &os; project sources, or all areas, depending on what interests you. The primary services we offer are Anonymous CVS, CVSup, and CTM. While it is possible to update only parts of your source tree, the only supported update procedure is to update the entire tree and recompile both userland (i.e., all the programs that run in user space, such as those in /bin and /sbin) and kernel sources. Updating only part of your source tree, only the kernel, or only userland will often result in problems. These problems may range from compile errors to kernel panics or data corruption. anonymous CVS Anonymous CVS and CVSup use the pull model of updating sources. In the case of CVSup the user (or a cron script) invokes the cvsup program, and it interacts with a cvsupd server somewhere to bring your files up-to-date. The updates you receive are up-to-the-minute and you get them when, and only when, you want them. You can easily restrict your updates to the specific files or directories that are of interest to you. Updates are generated on the fly by the server, according to what you have and what you want to have. Anonymous CVS is quite a bit more simplistic than CVSup in that it is just an extension to CVS which allows it to pull changes directly from a remote CVS repository. CVSup can do this far more efficiently, but Anonymous CVS is easier to use. CTM CTM, on the other hand, does not interactively compare the sources you have with those on the master archive or otherwise pull them across. Instead, a script which identifies changes in files since its previous run is executed several times a day on the master CTM machine, any detected changes being compressed, stamped with a sequence-number and encoded for transmission over email (in printable ASCII only). Once received, these CTM deltas can then be handed to the &man.ctm.rmail.1; utility which will automatically decode, verify and apply the changes to the user's copy of the sources. This process is far more efficient than CVSup, and places less strain on our server resources since it is a push rather than a pull model. There are other trade-offs, of course. If you inadvertently wipe out portions of your archive, CVSup will detect and rebuild the damaged portions for you. CTM will not do this, and if you wipe some portion of your source tree out (and do not have it backed up) then you will have to start from scratch (from the most recent CVS base delta) and rebuild it all with CTM or, with Anonymous CVS, simply delete the bad bits and resync. Rebuilding <quote>world</quote> Rebuilding world Once you have synchronized your local source tree against a particular version of &os; (&os.stable;, &os.current;, and so on) you can then use the source tree to rebuild the system. Take a Backup It cannot be stressed enough how important it is to take a backup of your system before you do this. While rebuilding the world is (as long as you follow these instructions) an easy task to do, there will inevitably be times when you make mistakes, or when mistakes made by others in the source tree render your system unbootable. Make sure you have taken a backup. And have a fixit floppy or bootable CD at hand. You will probably never have to use it, but it is better to be safe than sorry! Subscribe to the Right Mailing List mailing list The &os.stable; and &os.current; branches are, by their nature, in development. People that contribute to &os; are human, and mistakes occasionally happen. Sometimes these mistakes can be quite harmless, just causing your system to print a new diagnostic warning. Or the change may be catastrophic, and render your system unbootable or destroy your file systems (or worse). If problems like these occur, a heads up is posted to the appropriate mailing list, explaining the nature of the problem and which systems it affects. And an all clear announcement is posted when the problem has been solved. If you try to track &os.stable; or &os.current; and do not read the &a.stable; or the &a.current; respectively, then you are asking for trouble. Do not use <command>make world</command> A lot of older documentation recommends using make world for this. Doing that skips some important steps and should only be used if you are sure of what you are doing. For almost all circumstances make world is the wrong thing to do, and the procedure described here should be used instead. The Canonical Way to Update Your System To update your system, you should use the following procedure: &prompt.root; make buildworld &prompt.root; make buildkernel &prompt.root; make installkernel &prompt.root; reboot You should boot in single user mode (using boot -s from loader prompt for example). Then run: &prompt.root; mergemaster -p &prompt.root; make installworld &prompt.root; mergemaster &prompt.root; reboot Read Further Explanations The sequence described above is only a short resume to help you getting started. You should however read the following sections to clearly understand each step, especially if you want to use a custom kernel configuration. Read <filename>/usr/src/UPDATING</filename> Before you do anything else, read /usr/src/UPDATING (or the equivalent file wherever you have a copy of the source code). This file should contain important information about problems you might encounter, or specify the order in which you might have to run certain commands. If UPDATING contradicts something you read here, UPDATING takes precedence. Reading UPDATING is not an acceptable substitute for subscribing to the correct mailing list, as described previously. The two requirements are complementary, not exclusive. Check <filename>/etc/make.conf</filename> make.conf Examine the files /usr/share/examples/etc/make.conf (called /etc/defaults/make.conf in &os; 4.X) and /etc/make.conf. The first contains some default defines – most of which are commented out. To make use of them when you rebuild your system from source, add them to /etc/make.conf. Keep in mind that anything you add to /etc/make.conf is also used every time you run make, so it is a good idea to set them to something sensible for your system. A typical user will probably want to copy the CFLAGS and NOPROFILE lines found in /usr/share/examples/etc/make.conf (or in /etc/defaults/make.conf on &os; 4.X) to /etc/make.conf and uncomment them. Examine the other definitions (COPTFLAGS, NOPORTDOCS and so on) and decide if they are relevant to you. Update the Files in <filename>/etc</filename> The /etc directory contains a large part of your system's configuration information, as well as scripts that are run at system startup. Some of these scripts change from version to version of FreeBSD. Some of the configuration files are also used in the day to day running of the system. In particular, /etc/group. There have been occasions when the installation part of - make installworld has expected certain usernames or groups + make installworld has expected certain usernames or groups to exist. When performing an upgrade it is likely that these users or groups did not exist. This caused problems when upgrading. - In some cases make buildworld will check to see if + In some cases make buildworld will check to see if these users or groups exist. A recent example of this is when the smmsp user was added. Users had the installation process fail for them when &man.mtree.8; was trying to create /var/spool/clientmqueue. The solution is to examine /usr/src/etc/group and compare its list of groups with your own. If there are any groups in the new file that are not in your file then copy them over. Similarly, you should rename any groups in /etc/group which have the same GID but a different name to those in /usr/src/etc/group. Since 4.6-RELEASE you can run &man.mergemaster.8; in pre-buildworld mode by providing the option. This will compare only those files that are essential for the success of buildworld or installworld. If your old version of mergemaster does not support , use the new version in the source tree when running for the first time: &prompt.root; cd /usr/src/usr.sbin/mergemaster &prompt.root; ./mergemaster.sh -p If you are feeling particularly paranoid, you can check your system to see which files are owned by the group you are renaming or deleting: &prompt.root; find / -group GID -print will show all files owned by group GID (which can be either a group name or a numeric group ID). Drop to Single User Mode single-user mode You may want to compile the system in single user mode. Apart from the obvious benefit of making things go slightly faster, reinstalling the system will touch a lot of important system files, all the standard system binaries, libraries, include files and so on. Changing these on a running system (particularly if you have active users on the system at the time) is asking for trouble. multi-user mode Another method is to compile the system in multi-user mode, and then drop into single user mode for the installation. If you would like to do it this way, simply hold off on the following steps until the build has completed. You can postpone dropping to single user mode until you have to installkernel or installworld. As the superuser, you can execute: &prompt.root; shutdown now from a running system, which will drop it to single user mode. Alternatively, reboot the system, and at the boot prompt, enter the flag. The system will then boot single user. At the shell prompt you should then run: &prompt.root; fsck -p &prompt.root; mount -u / &prompt.root; mount -a -t ufs &prompt.root; swapon -a This checks the file systems, remounts / read/write, mounts all the other UFS file systems referenced in /etc/fstab and then turns swapping on. If your CMOS clock is set to local time and not to GMT (this is true if the output of the &man.date.1; command does not show the correct time and zone), you may also need to run the following command: &prompt.root; adjkerntz -i This will make sure that your local time-zone settings get set up correctly — without this, you may later run into some problems. Remove <filename>/usr/obj</filename> As parts of the system are rebuilt they are placed in directories which (by default) go under /usr/obj. The directories shadow those under /usr/src. - You can speed up the make buildworld process, and + You can speed up the make buildworld process, and possibly save yourself some dependency headaches by removing this directory as well. Some files below /usr/obj may have the immutable flag set (see &man.chflags.1; for more information) which must be removed first. &prompt.root; cd /usr/obj &prompt.root; chflags -R noschg * &prompt.root; rm -rf * Recompile the Source Saving the Output It is a good idea to save the output you get from running &man.make.1; to another file. If something goes wrong you will have a copy of the error message. While this might not help you in diagnosing what has gone wrong, it can help others if you post your problem to one of the &os; mailing lists. The easiest way to do this is to use the &man.script.1; command, with a parameter that specifies the name of the file to save all output to. You would do this immediately before rebuilding the world, and then type exit when the process has finished. &prompt.root; script /var/tmp/mw.out Script started, output file is /var/tmp/mw.out &prompt.root; make TARGET … compile, compile, compile … &prompt.root; exit Script done, … If you do this, do not save the output in /tmp. This directory may be cleared next time you reboot. A better place to store it is in /var/tmp (as in the previous example) or in root's home directory. Compile the Base System You must be in the /usr/src directory: &prompt.root; cd /usr/src (unless, of course, your source code is elsewhere, in which case change to that directory instead). make To rebuild the world you use the &man.make.1; command. This command reads instructions from the Makefile, which describes how the programs that comprise &os; should be rebuilt, the order in which they should be built, and so on. The general format of the command line you will type is as follows: &prompt.root; make -x -DVARIABLE target In this example, is an option that you would pass to &man.make.1;. See the &man.make.1; manual page for an example of the options you can pass. passes a variable to the Makefile. The behavior of the Makefile is controlled by these variables. These are the same variables as are set in /etc/make.conf, and this provides another way of setting them. &prompt.root; make -DNOPROFILE target is another way of specifying that profiled libraries should not be built, and corresponds with the NOPROFILE= true # Avoid compiling profiled libraries line in /etc/make.conf. target tells &man.make.1; what you want to do. Each Makefile defines a number of different targets, and your choice of target determines what happens. Some targets are listed in the Makefile, but are not meant for you to run. Instead, they are used by the build process to break out the steps necessary to rebuild the system into a number of sub-steps. Most of the time you will not need to pass any parameters to &man.make.1;, and so your command like will look like this: &prompt.root; make target Beginning with version 2.2.5 of &os; (actually, it was first created on the &os.current; branch, and then retrofitted to &os.stable; midway between 2.2.2 and 2.2.5) the world target has been split in two: buildworld and installworld. Beginning with version 5.3 of &os; the world target will be changed so it will not work at all by default because it is actually dangerous for most users. As the names imply, buildworld builds a complete new tree under /usr/obj, and installworld installs this tree on the current machine. This is very useful for 2 reasons. First, it allows you to do the build safe in the knowledge that no components of your running system will be affected. The build is self hosted. Because of this, you can safely run buildworld on a machine running in multi-user mode with no fear of ill-effects. It is still recommended that you run the installworld part in single user mode, though. Secondly, it allows you to use NFS mounts to upgrade multiple machines on your network. If you have three machines, A, B and C that you want to upgrade, run make buildworld and make installworld on A. B and C should then NFS mount /usr/src and /usr/obj from A, and you can then run make installworld to install the results of the build on B and C. Although the world target still exists, you are strongly encouraged not to use it. Run &prompt.root; make buildworld It is now possible to specify a option to make which will cause it to spawn several simultaneous processes. This is most useful on multi-CPU machines. However, since much of the compiling process is IO bound rather than CPU bound it is also useful on single CPU machines. On a typical single-CPU machine you would run: &prompt.root; make -j4 buildworld &man.make.1; will then have up to 4 processes running at any one time. Empirical evidence posted to the mailing lists shows this generally gives the best performance benefit. If you have a multi-CPU machine and you are using an SMP configured kernel try values between 6 and 10 and see how they speed things up. Be aware that this is still somewhat experimental, and commits to the source tree may occasionally break this feature. If the world fails to compile using this parameter try again without it before you report any problems. Timings Rebuilding world timings Many factors influence the build time, but currently a 500 MHz &pentium; III with 128 MB of RAM takes about 2 hours to build the &os.stable; tree, with no tricks or shortcuts used during the process. A &os.current; tree will take somewhat longer. Compile and Install a New Kernel kernel compiling To take full advantage of your new system you should recompile the kernel. This is practically a necessity, as certain memory structures may have changed, and programs like &man.ps.1; and &man.top.1; will fail to work until the kernel and source code versions are the same. The simplest, safest way to do this is to build and install a kernel based on GENERIC. While GENERIC may not have all the necessary devices for your system, it should contain everything necessary to boot your system back to single user mode. This is a good test that the new system works properly. After booting from GENERIC and verifying that your system works you can then build a new kernel based on your normal kernel configuration file. On modern versions of FreeBSD it is important to build world before building a new kernel. If you want to build a custom kernel, and already have a configuration file, just use KERNCONF=MYKERNEL like this: &prompt.root; cd /usr/src &prompt.root; make buildkernel KERNCONF=MYKERNEL &prompt.root; make installkernel KERNCONF=MYKERNEL In FreeBSD 4.2 and older you must replace KERNCONF= with KERNEL=. 4.2-STABLE that was fetched before Feb 2nd, 2001 does not recognize KERNCONF=. Note that if you have raised kern.securelevel above 1 and you have set either the noschg or similar flags to your kernel binary, you might find it necessary to drop into single user mode to use installkernel. Otherwise you should be able to run both these commands from multi user mode without problems. See &man.init.8; for details about kern.securelevel and &man.chflags.1; for details about the various file flags. If you are upgrading to a version of &os; below 4.0 you should use the old kernel build procedure. However, it is recommended that you use the new version of &man.config.8;, using a command line like this. &prompt.root; /usr/obj/usr/src/usr.sbin/config/config KERNELNAME Reboot into Single User Mode single-user mode You should reboot into single user mode to test the new kernel works. Do this by following the instructions in . Install the New System Binaries If you were building a version of &os; recent enough to have used make buildworld then you should now use installworld to install the new system binaries. Run &prompt.root; cd /usr/src &prompt.root; make installworld If you specified variables on the make buildworld command line, you must specify the same variables in the make installworld command line. This does not necessarily hold true for other options; for example, must never be used with installworld. For example, if you ran: &prompt.root; make -DNOPROFILE buildworld you must install the results with: &prompt.root; make -DNOPROFILE installworld otherwise it would try to install profiled libraries that had not been built during the make buildworld phase. Update Files Not Updated by <command>make installworld</command> Remaking the world will not update certain directories (in particular, /etc, /var and /usr) with new or changed configuration files. The simplest way to update these files is to use &man.mergemaster.8;, though it is possible to do it manually if you would prefer to do that. Regardless of which way you choose, be sure to make a backup of /etc in case anything goes wrong. Tom Rhodes Contributed by <command>mergemaster</command> mergemaster The &man.mergemaster.8; utility is a Bourne script that will aid you in determining the differences between your configuration files in /etc, and the configuration files in the source tree /usr/src/etc. This is the recommended solution for keeping the system configuration files up to date with those located in the source tree. mergemaster was integrated into the FreeBSD base system between 3.3-RELEASE and 3.4-RELEASE, which means it is present in all -STABLE and -CURRENT systems since 3.3. To begin simply type mergemaster at your prompt, and watch it start going. mergemaster will then build a temporary root environment, from / down, and populate it with various system configuration files. Those files are then compared to the ones currently installed in your system. At this point, files that differ will be shown in &man.diff.1; format, with the sign representing added or modified lines, and representing lines that will be either removed completely, or replaced with a new line. See the &man.diff.1; manual page for more information about the &man.diff.1; syntax and how file differences are shown. &man.mergemaster.8; will then show you each file that displays variances, and at this point you will have the option of either deleting the new file (referred to as the temporary file), installing the temporary file in its unmodified state, merging the temporary file with the currently installed file, or viewing the &man.diff.1; results again. Choosing to delete the temporary file will tell &man.mergemaster.8; that we wish to keep our current file unchanged, and to delete the new version. This option is not recommended, unless you see no reason to change the current file. You can get help at any time by typing ? at the &man.mergemaster.8; prompt. If the user chooses to skip a file, it will be presented again after all other files have been dealt with. Choosing to install the unmodified temporary file will replace the current file with the new one. For most unmodified files, this is the best option. Choosing to merge the file will present you with a text editor, and the contents of both files. You can now merge them by reviewing both files side by side on the screen, and choosing parts from both to create a finished product. When the files are compared side by side, the l key will select the left contents and the r key will select contents from your right. The final output will be a file consisting of both parts, which can then be installed. This option is customarily used for files where settings have been modified by the user. Choosing to view the &man.diff.1; results again will show you the file differences just like &man.mergemaster.8; did before prompting you for an option. After &man.mergemaster.8; is done with the system files you will be prompted for other options. &man.mergemaster.8; may ask if you want to rebuild the password file and/or run &man.MAKEDEV.8; if you run a FreeBSD version prior to 5.0, and will finish up with an option to remove left-over temporary files. Manual Update If you wish to do the update manually, however, you cannot just copy over the files from /usr/src/etc to /etc and have it work. Some of these files must be installed first. This is because the /usr/src/etc directory is not a copy of what your /etc directory should look like. In addition, there are files that should be in /etc that are not in /usr/src/etc. If you are using &man.mergemaster.8; (as recommended), you can skip forward to the next section. The simplest way to do this by hand is to install the files into a new directory, and then work through them looking for differences. Backup Your Existing <filename>/etc</filename> Although, in theory, nothing is going to touch this directory automatically, it is always better to be sure. So copy your existing /etc directory somewhere safe. Something like: &prompt.root; cp -Rp /etc /etc.old does a recursive copy, preserves times, ownerships on files and suchlike. You need to build a dummy set of directories to install the new /etc and other files into. /var/tmp/root is a reasonable choice, and there are a number of subdirectories required under this as well. &prompt.root; mkdir /var/tmp/root &prompt.root; cd /usr/src/etc &prompt.root; make DESTDIR=/var/tmp/root distrib-dirs distribution This will build the necessary directory structure and install the files. A lot of the subdirectories that have been created under /var/tmp/root are empty and should be deleted. The simplest way to do this is to: &prompt.root; cd /var/tmp/root &prompt.root; find -d . -type d | xargs rmdir 2>/dev/null This will remove all empty directories. (Standard error is redirected to /dev/null to prevent the warnings about the directories that are not empty.) /var/tmp/root now contains all the files that should be placed in appropriate locations below /. You now have to go through each of these files, determining how they differ with your existing files. Note that some of the files that will have been installed in /var/tmp/root have a leading .. At the time of writing the only files like this are shell startup files in /var/tmp/root/ and /var/tmp/root/root/, although there may be others (depending on when you are reading this). Make sure you use ls -a to catch them. The simplest way to do this is to use &man.diff.1; to compare the two files: &prompt.root; diff /etc/shells /var/tmp/root/etc/shells This will show you the differences between your /etc/shells file and the new /var/tmp/root/etc/shells file. Use these to decide whether to merge in changes that you have made or whether to copy over your old file. Name the New Root Directory (<filename>/var/tmp/root</filename>) with a Time Stamp, so You Can Easily Compare Differences Between Versions Frequently rebuilding the world means that you have to update /etc frequently as well, which can be a bit of a chore. You can speed this process up by keeping a copy of the last set of changed files that you merged into /etc. The following procedure gives one idea of how to do this. Make the world as normal. When you want to update /etc and the other directories, give the target directory a name based on the current date. If you were doing this on the 14th of February 1998 you could do the following: &prompt.root; mkdir /var/tmp/root-19980214 &prompt.root; cd /usr/src/etc &prompt.root; make DESTDIR=/var/tmp/root-19980214 \ distrib-dirs distribution Merge in the changes from this directory as outlined above. Do not remove the /var/tmp/root-19980214 directory when you have finished. When you have downloaded the latest version of the source and remade it, follow step 1. This will give you a new directory, which might be called /var/tmp/root-19980221 (if you wait a week between doing updates). You can now see the differences that have been made in the intervening week using &man.diff.1; to create a recursive diff between the two directories: &prompt.root; cd /var/tmp &prompt.root; diff -r root-19980214 root-19980221 Typically, this will be a much smaller set of differences than those between /var/tmp/root-19980221/etc and /etc. Because the set of differences is smaller, it is easier to migrate those changes across into your /etc directory. You can now remove the older of the two /var/tmp/root-* directories: &prompt.root; rm -rf /var/tmp/root-19980214 Repeat this process every time you need to merge in changes to /etc. You can use &man.date.1; to automate the generation of the directory names: &prompt.root; mkdir /var/tmp/root-`date "+%Y%m%d"` Update <filename>/dev</filename> DEVFS If you are running FreeBSD 5.0 or later you can safely skip this section. These versions use &man.devfs.5; to allocate device nodes transparently for the user. In most cases, the &man.mergemaster.8; tool will realize when it is necessary to update the device nodes, and offer to complete it automatically. These instructions tell how to update the device nodes manually. For safety's sake, this is a multi-step process. Copy /var/tmp/root/dev/MAKEDEV to /dev: &prompt.root; cp /var/tmp/root/dev/MAKEDEV /dev MAKEDEV If you used &man.mergemaster.8; to update /etc, then your MAKEDEV script should have been updated already, though it cannot hurt to check (with &man.diff.1;) and copy it manually if necessary. Now, take a snapshot of your current /dev. This snapshot needs to contain the permissions, ownerships, major and minor numbers of each filename, but it should not contain the time stamps. The easiest way to do this is to use &man.awk.1; to strip out some of the information: &prompt.root; cd /dev &prompt.root; ls -l | awk '{print $1, $2, $3, $4, $5, $6, $NF}' > /var/tmp/dev.out Remake all the device nodes: &prompt.root; sh MAKEDEV all Write another snapshot of the directory, this time to /var/tmp/dev2.out. Now look through these two files for any device node that you missed creating. There should not be any, but it is better to be safe than sorry. &prompt.root; diff /var/tmp/dev.out /var/tmp/dev2.out You are most likely to notice disk slice discrepancies which will involve commands such as: &prompt.root; sh MAKEDEV sd0s1 to recreate the slice entries. Your precise circumstances may vary. Update <filename>/stand</filename> This step is included only for completeness. It can safely be omitted. If you are using FreeBSD 5.2 or later, the /rescue directory is automatically updated for the user with current, statically compiled binaries during make installworld, thus obsoleting the need to update /stand. For the sake of completeness, you may want to update the files in /stand as well. These files consist of hard links to the /stand/sysinstall binary. This binary should be statically linked, so that it can work when no other file systems (and in particular /usr) have been mounted. &prompt.root; cd /usr/src/release/sysinstall &prompt.root; make all install Rebooting You are now done. After you have verified that everything appears to be in the right place you can reboot the system. A simple &man.shutdown.8; should do it: &prompt.root; shutdown -r now Finished You should now have successfully upgraded your &os; system. Congratulations. If things went slightly wrong, it is easy to rebuild a particular piece of the system. For example, if you accidentally deleted /etc/magic as part of the upgrade or merge of /etc, the &man.file.1; command will stop working. In this case, the fix would be to run: &prompt.root; cd /usr/src/usr.bin/file &prompt.root; make all install Questions Do I need to re-make the world for every change? There is no easy answer to this one, as it depends on the nature of the change. For example, if you just ran CVSup, and it has shown the following files as being updated: src/games/cribbage/instr.c src/games/sail/pl_main.c src/release/sysinstall/config.c src/release/sysinstall/media.c src/share/mk/bsd.port.mk it probably is not worth rebuilding the entire world. You could just go to the appropriate sub-directories and make all install, and that's about it. But if something major changed, for example src/lib/libc/stdlib then you should either re-make the world, or at least those parts of it that are statically linked (as well as anything else you might have added that is statically linked). At the end of the day, it is your call. You might be happy re-making the world every fortnight say, and let changes accumulate over that fortnight. Or you might want to re-make just those things that have changed, and be confident you can spot all the dependencies. And, of course, this all depends on how often you want to upgrade, and whether you are tracking &os.stable; or &os.current;. My compile failed with lots of signal 11 (or other signal number) errors. What has happened? signal 11 This is normally indicative of hardware problems. (Re)making the world is an effective way to stress test your hardware, and will frequently throw up memory problems. These normally manifest themselves as the compiler mysteriously dying on receipt of strange signals. A sure indicator of this is if you can restart the make and it dies at a different point in the process. In this instance there is little you can do except start swapping around the components in your machine to determine which one is failing. Can I remove /usr/obj when I have finished? The short answer is yes. /usr/obj contains all the object files that were produced during the compilation phase. Normally, one - of the first steps in the make buildworld process is to + of the first steps in the make buildworld process is to remove this directory and start afresh. In this case, keeping /usr/obj around after you have finished makes little sense, and will free up a large chunk of disk space (currently about 340 MB). However, if you know what you are doing you can have - make buildworld skip this step. This will make subsequent + make buildworld skip this step. This will make subsequent builds run much faster, since most of sources will not need to be recompiled. The flip side of this is that subtle dependency problems can creep in, causing your build to fail in odd ways. This frequently generates noise on the &os; mailing lists, when one person complains that their build has failed, not realizing that it is because they have tried to cut corners. Can interrupted builds be resumed? This depends on how far through the process you got before you found a problem. In general (and this is not a hard and - fast rule) the make buildworld process builds new + fast rule) the make buildworld process builds new copies of essential tools (such as &man.gcc.1;, and &man.make.1;) and the system libraries. These tools and libraries are then installed. The new tools and libraries are then used to rebuild themselves, and are installed again. The entire system (now including regular user programs, such as &man.ls.1; or &man.grep.1;) is then rebuilt with the new system files. If you are at the last stage, and you know it (because you have looked through the output that you were storing) then you can (fairly safely) do: … fix the problem … &prompt.root; cd /usr/src &prompt.root; make -DNOCLEAN all This will not undo the work of the previous - make buildworld. + make buildworld. If you see the message: -------------------------------------------------------------- Building everything.. -------------------------------------------------------------- - in the make buildworld output then it is + in the make buildworld output then it is probably fairly safe to do so. If you do not see that message, or you are not sure, then it is always better to be safe than sorry, and restart the build from scratch. How can I speed up making the world? Run in single user mode. Put the /usr/src and /usr/obj directories on separate file systems held on separate disks. If possible, put these disks on separate disk controllers. Better still, put these file systems across multiple disks using the &man.ccd.4; (concatenated disk driver) device. Turn off profiling (set NOPROFILE=true in /etc/make.conf). You almost certainly do not need it. Also in /etc/make.conf, set CFLAGS to something like . The optimization is much slower, and the optimization difference between and is normally negligible. lets the compiler use pipes rather than temporary files for communication, which saves disk access (at the expense of memory). Pass the option to &man.make.1; to run multiple processes in parallel. This usually helps regardless of whether you have a single or a multi processor machine. The file system holding /usr/src can be mounted (or remounted) with the option. This prevents the file system from recording the file access time. You probably do not need this information anyway. &prompt.root; mount -u -o noatime /usr/src The example assumes /usr/src is on its own file system. If it is not (if it is a part of /usr for example) then you will need to use that file system mount point, and not /usr/src. The file system holding /usr/obj can be mounted (or remounted) with the option. This causes disk writes to happen asynchronously. In other words, the write completes immediately, and the data is written to the disk a few seconds later. This allows writes to be clustered together, and can be a dramatic performance boost. Keep in mind that this option makes your file system more fragile. With this option there is an increased chance that, should power fail, the file system will be in an unrecoverable state when the machine restarts. If /usr/obj is the only thing on this file system then it is not a problem. If you have other, valuable data on the same file system then ensure your backups are fresh before you enable this option. &prompt.root; mount -u -o async /usr/obj As above, if /usr/obj is not on its own file system, replace it in the example with the name of the appropriate mount point. What do I do if something goes wrong? Make absolutely sure your environment has no extraneous cruft from earlier builds. This is simple enough. &prompt.root; chflags -R noschg /usr/obj/usr &prompt.root; rm -rf /usr/obj/usr &prompt.root; cd /usr/src &prompt.root; make cleandir &prompt.root; make cleandir Yes, make cleandir really should be run twice. Then restart the whole process, starting with make buildworld. If you still have problems, send the error and the output of uname -a to &a.questions;. Be prepared to answer other questions about your setup! Mike Meyer Contributed by Tracking for Multiple Machines NFS installing multiple machines If you have multiple machines that you want to track the same source tree, then having all of them download sources and rebuild everything seems like a waste of resources: disk space, network bandwidth, and CPU cycles. It is, and the solution is to have one machine do most of the work, while the rest of the machines mount that work via NFS. This section outlines a method of doing so. Preliminaries First, identify a set of machines that is going to run the same set of binaries, which we will call a build set. Each machine can have a custom kernel, but they will be running the same userland binaries. From that set, choose a machine to be the build machine. It is going to be the machine that the world and kernel are built on. Ideally, it should be a fast machine that has sufficient spare CPU to run make buildworld and make buildkernel. You will also want to choose a machine to be the test machine, which will test software updates before they are put into production. This must be a machine that you can afford to have down for an extended period of time. It can be the build machine, but need not be. All the machines in this build set need to mount /usr/obj and /usr/src from the same machine, and at the same point. Ideally, those are on two different drives on the build machine, but they can be NFS mounted on that machine as well. If you have multiple build sets, /usr/src should be on one build machine, and NFS mounted on the rest. Finally make sure that /etc/make.conf on all the machines in the build set agrees with the build machine. That means that the build machine must build all the parts of the base system that any machine in the build set is going to install. Also, each build machine should have its kernel name set with KERNCONF in /etc/make.conf, and the build machine should list them all in KERNCONF, listing its own kernel first. The build machine must have the kernel configuration files for each machine in /usr/src/sys/arch/conf if it is going to build their kernels. The Base System Now that all that is done, you are ready to build everything. Build the kernel and world as described in on the build machine, but do not install anything. After the build has finished, go to the test machine, and install the kernel you just built. If this machine mounts /usr/src and /usr/obj via NFS, when you reboot to single user you will need to enable the network and mount them. The easiest way to do this is to boot to multi-user, then run shutdown now to go to single user mode. Once there, you can install the new kernel and world and run mergemaster just as you normally would. When done, reboot to return to normal multi-user operations for this machine. After you are certain that everything on the test machine is working properly, use the same procedure to install the new software on each of the other machines in the build set. Ports The same ideas can be used for the ports tree. The first critical step is mounting /usr/ports from the same machine to all the machines in the build set. You can then set up /etc/make.conf properly to share distfiles. You should set DISTDIR to a common shared directory that is writable by whichever user root is mapped to by your NFS mounts. Each machine should set WRKDIRPREFIX to a local build directory. Finally, if you are going to be building and distributing packages, you should set PACKAGES to a directory similar to DISTDIR.