Bug #42917
ceph: task status not available
% Done:
0%
Source:
Q/A
Tags:
Backport:
nautilus
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(FS):
MDS
Labels (FS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
2019-11-20T20:08:49.706 INFO:tasks.cephfs_test_runner:====================================================================== 2019-11-20T20:08:49.706 INFO:tasks.cephfs_test_runner:ERROR: test_scrub_abort (tasks.cephfs.test_scrub_checks.TestScrubControls) 2019-11-20T20:08:49.707 INFO:tasks.cephfs_test_runner:---------------------------------------------------------------------- 2019-11-20T20:08:49.707 INFO:tasks.cephfs_test_runner:Traceback (most recent call last): 2019-11-20T20:08:49.707 INFO:tasks.cephfs_test_runner: File "/home/teuthworker/src/github.com_ceph_ceph_master/qa/tasks/cephfs/test_scrub_checks.py", line 58, in test_scrub_abort 2019-11-20T20:08:49.707 INFO:tasks.cephfs_test_runner: self._check_task_status("idle") 2019-11-20T20:08:49.707 INFO:tasks.cephfs_test_runner: File "/home/teuthworker/src/github.com_ceph_ceph_master/qa/tasks/cephfs/test_scrub_checks.py", line 35, in _check_task_status 2019-11-20T20:08:49.707 INFO:tasks.cephfs_test_runner: self.assertTrue(task_status['0'].startswith(expected_status)) 2019-11-20T20:08:49.708 INFO:tasks.cephfs_test_runner:KeyError: '0'
From: /ceph/teuthology-archive/pdonnell-2019-11-20_19:35:29-fs-master-distro-basic-smithi/4525848/teuthology.log
This was broken by the fix to #42635. The MDS no longer registers itself as a service daemon because the mons would get confused by printing two instances for the MDS in the `ceph versions` output, one the correct version and the other "unknown".
The fix for this ticket could reintroduce the call by reverting 046137f819aae72f1423e3feb213f0e46c97c9ce (and moving the code to mgr init in MDSDaemon.cc) However, it needs to make sure `ceph fs status` is not broken.
Related issues
History
#1 Updated by Patrick Donnelly over 4 years ago
- Related to Bug #42635: mgr: daemon state for mds not available added
#2 Updated by Patrick Donnelly over 4 years ago
- Status changed from New to Duplicate
#3 Updated by Patrick Donnelly over 4 years ago
- Duplicates Bug #42835: qa: test_scrub_abort fails during check_task_status("idle") added