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

Jump to bottom

[DNM] jewel: osd ops (sent and?) arrive at osd out of order #13885

Closed
wants to merge 3 commits into from
Closed

[DNM] jewel: osd ops (sent and?) arrive at osd out of order #13885

wants to merge 3 commits into from

Conversation

shinobu-x
Copy link
Contributor

@shinobu-x shinobu-x commented Mar 8, 2017 •
edited by smithfarm

@liewegas @shinobu-x
osdc/Objecter: resend RWORDERED ops on full
04ef7c9 
Our condition for respecting the FULL flag is complex, and involves
the WRITE | RWORDERED flags vs the FULL_FORCE | FULL_TRY flags.  Previously,
we could block a read bc of RWORDRED but not resend it later.

Fix by capturing the complex condition in a respects_full() bool and using
it both for the blocking-on-send and resending-on-possibly-notfull-later
checks.

Fixes: http://tracker.ceph.com/issues/19133
Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit c4b73f1)

Conflicts:
	src/osdc/Objecter.cc
@liewegas @shinobu-x
qa/tasks/thrashosds: chance_thrash_cluster_full
d71baee 
Induce a momentarily full cluster.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit a202b68)

Conflicts:
	qa/tasks/ceph_manager.py
	qa/tasks/thrashosds.py
@liewegas @shinobu-x
qa/tasks/ceph_manager: use new luminous set-full-ratio etc
7fb9245 
Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 2967080)
@smithfarm smithfarm added this to the jewel milestone Apr 9, 2017
@smithfarm smithfarm added the core label Apr 9, 2017
@smithfarm smithfarm mentioned this pull request Apr 9, 2017
Closed
@smithfarm
Copy link
Contributor

@tchaikov There was a massive failure of the upgrade/hammer-x suite when I added this PR to the integration branch.

Before: http://pulpito.ceph.com/smithfarm-2017-04-07_08:06:45-upgrade:hammer-x-wip-jewel-backports-distro-basic-vps/

After: http://pulpito.ceph.com/smithfarm-2017-04-09_21:01:21-upgrade:hammer-x-wip-jewel-backports-distro-basic-vps/

Is this PR is dependent on #13615 ?

@smithfarm smithfarm changed the title jewel: osd ops (sent and?) arrive at osd out of order [DNM] jewel: osd ops (sent and?) arrive at osd out of order Apr 10, 2017
@tchaikov
Copy link
Contributor

we should only backport c4b73f1. the other two commits are using the command only available in master: "ceph osd set-[near]full-ratio", see 6422e0a

@smithfarm
Copy link
Contributor

closing in favor of #14427

@smithfarm smithfarm closed this Apr 13, 2017
tobias-urdin pushed a commit to tobias-urdin/ceph that referenced this pull request Aug 2, 2023
@zhouyuan
gevent: bump to >=1.0
22edd83 
There are some DNS resolving issue on 0.13.6.
Bumping to >=1.0 fixes the issues for me.

Fixes: ceph#13885

Signed-off-by: Yuan Zhou <yuan.zhou@intel.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
bug-fix core
Projects
None yet
Milestone
jewel
4 participants
@shinobu-x @smithfarm @tchaikov @liewegas