https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=261855
Bug ID: 261855
Summary: /etc/periodic/daily/800.scrub-zfs is using `zpool
history`
Product: Base System
Version: 13.0-RELEASE
Hardware: Any
OS: Any
Status: New
Severity: Affects Some People
Priority: ---
Component: bin
Assignee: [email protected]
Reporter: [email protected]
Daily scrub periodic task is currently using `zpool history` to check the last
time a `zpool scrub` was started (note: not completed, or executed, just
started).
On my system, which has a lots of filesystems and hourly
snapshot/send/recv/destroy jobs in order to replicate it to/from another system
my `zpool history` only covers 48h of data and thus periodic scrub (configured
to 35 days as per default) gets executed every 3 days (this is the reason I
looked into it in the first place).
Right now the logic is more or less:
1. check last scrub start in `zpool history`
2. calculate diff, if < 35 just stop
3. check "scan:" line from `zpool status` and do different stuff
I would suggest to remove the usage of `zpool history` and replace it with a
parsing of the `zpool status` "scan:" line, which can be done with `LANG=C date
-j -f '%a %b %d %T %Y'`.
% zpool status | fgrep 'scan:'
scan: scrub canceled on Thu Feb 10 08:40:04 2022
scan: scrub repaired 0B in 00:00:26 with 0 errors on Sun Jan 23 19:09:50 2022
This can be easily integrated in the existing `case` by adding a new match like
`*"scrub repaired"*|*"scrub canceled"*)` and calculate the end of last scrub
based on that.
Caveat: `zpool status` gets reset each time a (manual) scrub or pause or cancel
is done. But the current method has caveats too (using start time instead of
end time, using a hisotry command which can not be very trustworthy in some
situations) so I think this would be an overall better solution.
What do you think?
I can prepare a diff for this.
--
You are receiving this mail because:
You are the assignee for the bug.
