Wip nitzan pglog ec getattr error

[NitzanMordhai]

For 'copy object' in EC pool, when the target copy does not exist,
getattr_maybe_cache returns ENODATA. That will cause ref_count to wildcard tag (backward capability),
which might lead pglog to have refcounts for each copy, and to grow quickly.

Fixes: https://tracker.ceph.com/issues/56707
Signed-off-by: Nitzan Mordechai nmordec@redhat.com

Contribution Guidelines

• To sign and title your commits, please refer to Submitting Patches to Ceph.

• If you are submitting a fix for a stable branch (e.g. "pacific"), please refer to Submitting Patches to Ceph - Backports for the proper workflow.

Checklist

• Tracker (select at least one)
  • References tracker ticket
  • Very recent bug; references commit where it was introduced
  • New feature (ticket optional)
  • Doc update (no ticket needed)
  • Code cleanup (no ticket needed)
• Component impact
  • Affects Dashboard, opened tracker ticket
  • Affects Orchestrator, opened tracker ticket
  • No impact that needs to be tracked
• Documentation (select at least one)
  • Updates relevant documentation
  • No doc update is appropriate
• Tests (select at least one)
  • Includes unit test(s)
  • Includes integration test(s)
  • Includes bug reproducer
  • No tests

Show available Jenkins commands

• jenkins retest this please
• jenkins test classic perf
• jenkins test crimson perf
• jenkins test signed
• jenkins test make check
• jenkins test make check arm64
• jenkins test submodules
• jenkins test dashboard
• jenkins test dashboard cephadm
• jenkins test api
• jenkins test docs
• jenkins render docs
• jenkins test ceph-volume all
• jenkins test ceph-volume tox
• jenkins test windows

[NitzanMordhai]

jenkins test make check

[NitzanMordhai]

jenkins test make check

[ronen-fr]

I'd suggest to rephrase the PR/commit message. Maybe:
For 'copy object' in EC pool, when the target copy does not exist,
getattr_maybe_cache returns ENODATA. That will cause ref_count to wildcard tag (??),
which might lead pglog to have refcounts for each copy, and to grow quickly.

(hope I got the meaning correctly)

[NitzanMordhai]

I'd suggest to rephrase the PR/commit message. Maybe:
For 'copy object' in EC pool, when the target copy does not exist,
getattr_maybe_cache returns ENODATA. That will cause ref_count to wildcard tag (??),
which might lead pglog to have refcounts for each copy, and to grow quickly.

Yes, that sounds better, will update it

[athanatos]

I also don't understand the "ref_count to wildcard tag" portion of the commit message.

[athanatos]

Fix LGTM though.

[NitzanMordhai]

jenkins test make check

[NitzanMordhai]

jenkins test make check arm64

[NitzanMordhai]

jenkins test windows

[athanatos]

Is there a better explanation somewhere of how rgw uses this xattr? Why exactly does returning the wrong error code here cause another xattr to increase in size without bound?

[NitzanMordhai]

Is there a better explanation somewhere of how rgw uses this xattr? Why exactly does returning the wrong error code here cause another xattr to increase in size without bound?

read_refcount is checking the return code from cls_cxx_getxattr, in case that the copy is not existing we will get ENODATA (before the change) and return 0, cls_rc_refcount_read\get\put will check that return code and reset the refcount.
After the fix, we will get return code of -ENOENT and will not call set_refcount (just like replica pools).