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

jewel: osd: Reset() snaptrimmer on shutdown and do not default-abort on leaked pg refs #15322

Merged
merged 2 commits into from Aug 24, 2017

Conversation

Projects
None yet
3 participants
@gregsfortytwo
Member

gregsfortytwo commented May 26, 2017

http://tracker.ceph.com/issues/20084

We no longer default-assert on shutdown when we have leaked PGRefs, and the snap trimmer doesn't leak them gratuitously either!

gregsfortytwo added some commits May 26, 2017

osd: Reset() the snaptrimmer on shutdown
We were failing to exit various wait states which held PGRefs. Error!

Fixes: http://tracker.ceph.com/issues/19931

Signed-off-by: Greg Farnum <gfarnum@redhat.com>
(cherry picked from commit b0e9dee)

Conflicts:
	src/osd/ReplicatedPG.cc

Signed-off-by: Greg Farnum <gfarnum@redhat.com>
osd: do not default-abort on leaked pg refs
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
(cherry picked from commit 4caf2df)

Conflicts:
	qa/clusters/fixed-4.yaml
	src/osd/OSD.cc

Fixes: http://tracker.ceph.com/issues/20084

Signed-off-by: Greg Farnum <gfarnum@redhat.com>

@gregsfortytwo gregsfortytwo added this to the jewel milestone May 26, 2017

@jdurgin

lgtm assuming it passes tests

@smithfarm

This comment has been minimized.

Show comment
Hide comment
@smithfarm

smithfarm Jun 19, 2017

Contributor

Jenkins re-test this please

Contributor

smithfarm commented Jun 19, 2017

Jenkins re-test this please

@smithfarm

This comment has been minimized.

Show comment
Hide comment
@smithfarm

smithfarm Jun 19, 2017

Contributor

@gregsfortytwo Can you look at the Jenkins failure?

Unexpected die of job

The job died or timeout on unittest_pglog.log

Contributor

smithfarm commented Jun 19, 2017

@gregsfortytwo Can you look at the Jenkins failure?

Unexpected die of job

The job died or timeout on unittest_pglog.log

@smithfarm smithfarm changed the title from Jewel 19931 snaptrim shutdown to [DNM] Jewel 19931 snaptrim shutdown Jun 19, 2017

@gregsfortytwo

This comment has been minimized.

Show comment
Hide comment
@gregsfortytwo

gregsfortytwo Jun 19, 2017

Member

retest this please

This shouldn't have any impact on the PGLog and there doesn't appear to be any output from the failed test...

Member

gregsfortytwo commented Jun 19, 2017

retest this please

This shouldn't have any impact on the PGLog and there doesn't appear to be any output from the failed test...

@smithfarm smithfarm changed the title from [DNM] Jewel 19931 snaptrim shutdown to jewel: osd: Reset() snaptrimmer on shutdown and do not default-abort on leaked pg refs Jun 20, 2017

@smithfarm

This comment has been minimized.

Show comment
Hide comment
@smithfarm

smithfarm Aug 23, 2017

Contributor

@gregsfortytwo @jdurgin This passed a rados suite with 1 failure ( a segfault - see http://tracker.ceph.com/issues/21063#note-2 ). @liewegas says that KStore is not supported, so I'm tentatively ruling the suite a "pass".

Contributor

smithfarm commented Aug 23, 2017

@gregsfortytwo @jdurgin This passed a rados suite with 1 failure ( a segfault - see http://tracker.ceph.com/issues/21063#note-2 ). @liewegas says that KStore is not supported, so I'm tentatively ruling the suite a "pass".

@smithfarm smithfarm merged commit d79618b into ceph:jewel Aug 24, 2017

6 checks passed

Docs: build check OK - docs built
Details
Signed-off-by all commits in this PR are signed
Details
Unmodifed Submodules submodules for project are unmodified
Details
Unmodified Submodules submodules for project are unmodified
Details
default Build finished.
Details
make check make check succeeded
Details
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment