camcontrol(8): add "smart" command
AbandonedPublic
Actions

Authored by trasz on Apr 26 2021, 1:11 PM.

Details

Reviewers

mav
imp

Group Reviewers

manpages

Summary

This implements "camcontrol smart", which can be used to check
status of ATA drives.

Sponsored by: NetApp, Inc.
Sponsored by: Klara, Inc.

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Passed

Unit

No Test Coverage

Build Status

Buildable 38867
Build 35756: arc lint + arc unit

Event Timeline

trasz created this revision.Apr 26 2021, 1:11 PM

Herald added a subscriber: imp. · View Herald TranscriptApr 26 2021, 1:11 PM

trasz requested review of this revision.Apr 26 2021, 1:11 PM

Harbormaster completed remote builds in B38865: Diff 88167.Apr 26 2021, 1:11 PM

trasz added reviewers: mav, imp.Apr 26 2021, 1:12 PM

trasz added a subscriber: NetApp.

I am not sure how useful it is to report only SMART status. I barely remember any cases when SMART reported actual error. Much more often we start replacing drives preventivelywhen they start reallocating and loose performance. But is you wish, I have no objections.

sbin/camcontrol/camcontrol.c
9884	It would be really good to support at very least SATL too. May be also SCSI and NVMe, even though SMART is very different there.
9926	Why shutdown these days?

Add man page.

Herald added a reviewer: manpages. · View Herald TranscriptApr 26 2021, 2:36 PM

Harbormaster completed remote builds in B38867: Diff 88174.Apr 26 2021, 2:37 PM

In D29992#672920, @mav wrote:

I am not sure how useful it is to report only SMART status. I barely remember any cases when SMART reported actual error. Much more often we start replacing drives preventivelywhen they start reallocating and loose performance.

I concur, at least consumer devices only start reporting SMART errors, after they have actually failed and data is no longer recoverable - but even then, its possible to NOT get a smart error.

Getting the full statistics - to the extent the device is willing to provide them - does allow a proactive approach. Dumping the statistics in a verbose mode would be helpful there.

I have to echo the comments here. fatal error only is useless on the vast majority of devices. We run a lot of our devices well past the time smart alerts (because the alert turned out to be bogus). Without statistics, this functionality is really useless.

SATL devices are SATA devices, and should be supported.

We should consider adding libxo so that the verbose stats from this (and other) commands can be output in a structured way.

sbin/camcontrol/camcontrol.c
9884	As mav@ points out, SATL device work just fine too.

rpokala added a subscriber: rpokala.Apr 26 2021, 5:44 PM

rpokala added inline comments.

sbin/camcontrol/camcontrol.c
9928	I've been hacking on ATA-related code for ages, and this is the first time I've seen a reference to this value. Which version of the standard introduced it?