zpool import progress kstat

[ofaaland]

Motivation and Context

When an import requires a long MMP activity check, or when the user
requests pool recovery, the import make take a long time. The user may
not know why, or be able to tell whether the import is progressing or is
hung.

Description

Add a kstat which lists all imports currently being processed by the
kernel (currently only one at a time is possible, but the kstat allows
for more than one). The kstat is at
/proc/spl/kstat/zfs/import_progress.

The kstat contents are as follows:

pool_guid            load_state multihost_secs  max_txg pool_name
16667015954387398856 3           15             0       tank3

load_state: the value of spa_load_state
multihost_secs: number of seconds left in the multihost activity check, if any.
max_txg: current spa_load_max_txg, if rewind is occurring

This could be used by outside tools, such as a pacemaker resource agent,
to report import progress, or as a part of manual troubleshooting. The
zpool import subcommand could also be modified to report this
information.

How Has This Been Tested?

Manual testing to confirm contents, including imports both with and without rewind and multihost activity check, together and separately. Multiple imports issued in userspace at the same time. Ran zloop and MMP tests in ZTS.

Updated zpool import to use this procfile (See #8646).

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the ZFS on Linux code style requirements.
• I have updated the documentation accordingly.
• I have read the contributing document.
• I have added tests to cover my changes.
• All new and existing tests passed.
• All commit messages are properly formatted and contain Signed-off-by.

[ofaaland]

Hi @utopiabound is it possible for you to review this patch? Thanks.

[ofaaland]

@behlendorf , I've addressed your comments above. I also addressed the issues we discussed offline:
Instead of reporting the start and end times of the activity check, the kstat now reports seconds remaining. I believe the failure code paths are now properly handled, and the load state is now updated everywhere.

[codecov]

Codecov Report

Merging #8696 into master will increase coverage by 0.17%.
The diff coverage is 77.87%.

@@            Coverage Diff             @@
##           master    #8696      +/-   ##
==========================================
+ Coverage   78.73%    78.9%   +0.17%     
==========================================
  Files         381      381              
  Lines      117674   117744      +70     
==========================================
+ Hits        92649    92909     +260     
+ Misses      25025    24835     -190

Flag	Coverage Δ
#kernel	79.42% <80.18%> (+0.11%)	⬆️
#user	67.69% <54.86%> (+0.46%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 1f02ecc...930f70a. Read the comment docs.

[behlendorf]

Thanks for addressing the review feedback so quickly.

[ofaaland]

Instead of reporting the start and end times of the activity check, the kstat now reports seconds remaining

A note for future me: the reasoning is that the kernel knows with certainty how much wall clock time is left in the test. It uses a monotonic clock source. However translating that monotonic end time to system time is error-prone due to the possibility that something (e.g. NTP) changes the system time during the test. So we just give seconds remaining.