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: PG::all_unfound_are_queried_or_lost for non-existent osds #4416

Merged
1 commit merged into from May 15, 2015

Conversation

Projects
None yet
3 participants
@xinxinsh
Member

xinxinsh commented Apr 21, 2015

@ghost ghost added bug fix core labels Apr 21, 2015

@ghost ghost assigned xinxinsh Apr 21, 2015

@ghost ghost added this to the firefly milestone Apr 21, 2015

@ghost

This comment has been minimized.

ghost commented Apr 21, 2015

@xinxinsh could you please explain how the conflict was resolved for e111c96 ? There are explanations about that at http://tracker.ceph.com/projects/ceph-releases/wiki/HOWTO_backport_commits, including a oneliner that can conveniently be used to display the difference between the original commit and the backported commit that required resolution

@xinxinsh

This comment has been minimized.

Member

xinxinsh commented Apr 21, 2015

updated

@ghost

This comment has been minimized.

ghost commented Apr 21, 2015

@xinxinsh great, will be helpful to @trociny when reviewing :-)

@trociny

This comment has been minimized.

Contributor

trociny commented Apr 21, 2015

The conflict resolution (replace $TEST_POOL with data pool) looks ok to me. Other looks good too.

You might consider though if 7294e8c is going to be backported too in near future, and if it is, then it may be better to do it the first, to avoid double conflict.

Also, actually, taking into the consideration the current state of test_lost.sh and that is not integrated to 'make check', I hardly expect somebody to run these tests on firefly. So we may consider backporting only the fix (883f87f), without the supplementary tests.

Mykola Golub xinxin shu
osd: fix PG::all_unfound_are_queried_or_lost for non-existent osds
A common mistake upon osd loss is to remove the osd from the crush map
before marking the osd lost. This tends to make it so that the user
can no longer mark the osd lost to satisfy all_unfound_are_queried_or_lost.

The simple solution is for all_unfound_are_queried_or_lost to ignore
the osd if it does not exist.

Fixes: #10976
Backports: firefly,giant

Signed-off-by: Mykola Golub <mgolub@mirantis.com>
(cherry picked from commit 5bb5132)
@xinxinsh

This comment has been minimized.

Member

xinxinsh commented Apr 24, 2015

@trociny @dachary updated

@trociny

This comment has been minimized.

Contributor

trociny commented Apr 24, 2015

It looks ok to me.

@ghost

This comment has been minimized.

ghost commented May 15, 2015

@athanatos does this backport look good to merge ? It passed a run of the firefly rados suite ( see http://tracker.ceph.com/issues/11090#rados for details ). There was only one failure described at http://tracker.ceph.com/issues/11142#note-5

ghost pushed a commit that referenced this pull request May 15, 2015

Loic Dachary
Merge pull request #4416 from xinxinsh/wip-10976-firefly
fix PG::all_unfound_are_queried_or_lost for non-existent osds

Reviewed-by: Samuel Just <sjust@redhat.com>

@ghost ghost merged commit ac7d28a into ceph:firefly May 15, 2015

@ghost ghost changed the title from fix PG::all_unfound_are_queried_or_lost for non-existent osds to osd: PG::all_unfound_are_queried_or_lost for non-existent osds Jul 14, 2015

This issue was closed.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment