Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

osd/PeeringState: separate history's pruub from pg's #44015

Merged
merged 1 commit into from Dec 13, 2021

Conversation

liewegas
Copy link
Member

@liewegas liewegas commented Nov 18, 2021

(pruub = prior_readable_until_ub [upper bound])

During peering, a primary may conclude that it does not need to wait for
the prior interval(s)' read lease because it will query all such osds.
However, it is dangerous to reflect that local inference about future
peering effects in the info.history, which is freely shared with other
OSDs. For example, if the primary cleared the history pruub, shared it,
and then failed, the next primary may conclude that it does not need to
wait for the lease to expire.

Instead, track the pruub in the conventional way. Only at the end of
peering do we clear it (if there are no prior_interval_down_osds) before
ending our final activate infos out, when we have already talked to the
peer osds and they know the prior interval has finished.

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)
  • 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)
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

Copy link
Member

@neha-ojha neha-ojha left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@liewegas could you please link the tracker ticket in your commit? Also, you mentioned that you were able to reproduce the problem very easily, would it make sense to add a test case to replicate that?

(pruub = prior_readable_until_ub [upper bound])

During peering, a primary may conclude that it does not need to wait for
the prior interval(s)' read lease because it will query all such osds.
However, it is dangerous to reflect that local inference about future
peering effects in the info.history, which is freely shared with other
OSDs.  For example, if the primary cleared the history pruub, shared it,
and then failed, the next primary may conclude that it does not need to
wait for the lease to expire.

Instead, track the pruub in the conventional way.  Only at the end of
peering do we clear it (if there are no prior_interval_down_osds) before
ending our final activate infos out, when we have already talked to the
peer osds and they know the prior interval has finished.

Fixes: https://tracker.ceph.com/issues/53326
Signed-off-by: Sage Weil <sage@newdream.net>
Copy link
Contributor

@athanatos athanatos left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yep, that looks right.

@yuriw
Copy link
Contributor

yuriw commented Dec 13, 2021

@yuriw yuriw merged commit c3ee11e into ceph:master Dec 13, 2021
8 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
4 participants