Actions
Bug #11339
closedceph-fuse failing to drop cap on unlinked dir
Status:
Resolved
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
% Done:
0%
Source:
other
Tags:
Backport:
Regression:
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(FS):
Labels (FS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
Shown up in the test_strays:
2015-04-02 12:27:31,551.551 INFO:tasks.cephfs.filesystem:_json_asok output: { "mds_cache": { "num_strays": 1, "num_strays_purging": 0, "num_strays_delayed": 0, "num_purge_ops": 0, "strays_created": 3201, "strays_purged": 3200, "strays_reintegrated": 0, "strays_migrated": 0, "num_recovering_processing": 0, "num_recovering_enqueued": 0, "num_recovering_prioritized": 0, "recovery_started": 0, "recovery_completed": 0 } } 2015-04-02 12:27:31,552.552 ERROR:tasks.cephfs_test_runner:Traceback (most recent call last): File "/Users/john/git/ceph-qa-suite/tasks/cephfs/test_strays.py", line 24, in test_files_throttle self._test_throttling(self.FILES_THROTTLE) File "/Users/john/git/ceph-qa-suite/tasks/cephfs/test_strays.py", line 114, in _test_throttling raise RuntimeError("Timeout waiting for {0} inodes to purge, stats:{1}".format(total_inodes, mdc_stats)) RuntimeError: Timeout waiting for 3201 inodes to purge, stats:{u'strays_migrated': 0, u'strays_purged': 3200, u'strays_created': 3201, u'strays_reintegrated': 0, u'recovery_started': 0, u'num_purge_ops': 0, u'num_recovering_enqueued': 0, u'recovery_completed': 0, u'num_recovering_prioritized': 0, u'num_strays_delayed': 0, u'num_strays_purging': 0, u'num_recovering_processing': 0, u'num_strays': 1} 2015-04-02 12:27:31,552.552 ERROR:tasks.cephfs_test_runner:Error in test 'test_files_throttle (test_strays.TestStrays)', going interactive Ceph test interactive mode, use ctx to interact with the cluster, press control-D to exit...
In this instance, the test was tripped up by the client failing to give up its cap on the /delete_me directory after "rm -rf" was run on it.
This used to work, so it's either the result of a recent client change, or it's intermittent.
Updated by Zheng Yan about 9 years ago
looks like commit 0c6bd2775fe43274e07c5b10586565d6170d8ea9 client: Hold on to exclusive caps on directories we "own"' changes the time sequence how caps are issued,
Updated by Zheng Yan about 9 years ago
- Status changed from New to Fix Under Review
Updated by Greg Farnum about 9 years ago
- Status changed from Fix Under Review to Resolved
https://github.com/ceph/ceph/pull/4302, merged to master in commit:c1de23ff45aa3d40d55706bba4a8221d9f3b189d
Actions