Bug 248641 - Openzfs 2.0 integration issue - autotrim not enabled - POLA
Summary: Openzfs 2.0 integration issue - autotrim not enabled - POLA
Status: Closed FIXED
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: CURRENT
Hardware: Any Any
: --- Affects Many People
Assignee: freebsd-bugs (Nobody)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2020-08-13 14:44 UTC by Jeremy Faulkner
Modified: 2023-03-04 00:02 UTC (History)
6 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Jeremy Faulkner 2020-08-13 14:44:29 UTC
Autotrim defaults to off even on SSDs, in violation to of POLA as no user action was previously required for trim in ZFS.
Comment 1 Matt Macy freebsd_committer freebsd_triage 2020-08-13 23:37:33 UTC
The autotrim property now defaults to on when built within FreeBSD
Comment 2 Mark Johnston freebsd_committer freebsd_triage 2020-08-19 14:24:18 UTC
https://github.com/openzfs/zfs/pull/10719
Comment 3 Adam McDougall 2023-03-04 00:01:52 UTC
I noticed recent FreeBSD 13 systems consuming more space than expected and I traced it down to the new autotrim zpool property on my 13 installs. I can watch gstat -d output during/after data deletions and BIO_DELETE IO does not happen. If I manually trim the pool or enable zpool set autotrim=on rpool then deletions happen. I am not relying on what state the autotrim property says, I am basing my conclusions on measurements. Unless something has changed within the last few months, I believe this is still the case.
Comment 4 Adam McDougall 2023-03-04 00:02:03 UTC
I noticed recent FreeBSD 13 systems consuming more space than expected and I traced it down to the new autotrim zpool property on my 13 installs. I can watch gstat -d output during/after data deletions and BIO_DELETE IO does not happen. If I manually trim the pool or enable zpool set autotrim=on rpool then deletions happen. I am not relying on what state the autotrim property says, I am basing my conclusions on measurements. Unless something has changed within the last few months, I believe this is still the case.