Bug #6179: ceph_test_rados user_version checks fail during thrashing - Ceph - Ceph

Actions

Copy link

Bug #6179

closed

ceph_test_rados user_version checks fail during thrashing

Added by Sage Weil over 10 years ago. Updated over 10 years ago.

Status:

Resolved

Priority:

Urgent

Assignee:

Sage Weil

Category:

OSD

Target version:

% Done:

Source:

Q/A

Tags:

Backport:

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

ceph-qa-suite:

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

2013-08-30T01:33:42.532 INFO:teuthology.task.rados.rados.0.out:[10.214.132.15]: Writing 44 current snap is 201
...
2013-08-30T01:33:42.646 INFO:teuthology.task.rados.rados.0.out:[10.214.132.15]: finishing write tid 1 to plana635252-44
2013-08-30T01:33:42.677 INFO:teuthology.task.rados.rados.0.out:[10.214.132.15]: finishing write tid 2 to plana635252-44
2013-08-30T01:33:42.678 INFO:teuthology.task.rados.rados.0.out:[10.214.132.15]: finishing write tid 3 to plana635252-44
2013-08-30T01:33:42.678 INFO:teuthology.task.rados.rados.0.out:[10.214.132.15]: update_object_version oid 44 is version 8222
...
2013-08-30T01:33:43.777 INFO:teuthology.task.rados.rados.0.out:[10.214.132.15]: Reading 44
...
2013-08-30T01:33:44.713 INFO:teuthology.task.rados.rados.0.err:[10.214.132.15]: oid: 44 version is 8224 and expected 8222

Actions

Copy link

Updated by Sage Weil over 10 years ago

Subject changed from ceph_test_rados gets user_version of 0 from thrashing to ceph_test_rados user_version checks fail during thrashing
Description updated (diff)
Category set to OSD
Status changed from New to 12
Priority changed from Normal to Urgent
Source changed from other to Q/A

Actions

Copy link

Updated by Sage Weil over 10 years ago

flab:teuthology 10:17 AM $ teuthology-schedule a.yaml --name sage-bug-6179-a -n 1
Job scheduled with ID 13732

should run shortly

Actions

Copy link

Updated by Greg Farnum over 10 years ago

Unfortunately I didn't set up any debug output for user versions because...well, because I didn't want to confuse things and they shouldn't match. That was stupid and we should fix it.

However, there is plenty of output on the involved versions. Both of the ops associated with versions 17658 and 17660 (there are two different ops, yes!) get replayed (no idea why yet), and 17658 first appears as part of a make_writeable clone. I'm having trouble parsing more than that, though. :(

Actions

Copy link