Actions
Bug #16741
closedio getting stuck after advancing journal object set
Status:
Resolved
Priority:
Normal
Assignee:
-
Target version:
-
% Done:
0%
Source:
other
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
I can easily reproduce this issue running write bench on an image that has small object size journal:
rbd create --size 128 --image-feature layering,exclusive-lock,journaling ${image} --journal-object-size 64K
rbd --debug-rbd 30 --debug-journaler 30 bench-write ${image} --io-size 4096 --io-threads 1 --io-total 409600 --io-pattern rand
It hangs for me shortly after:
2016-07-19 21:42:22.323799 7f498affd700 10 JournalRecorder: object journal_data.31.37142ae8944a.0 now full 2016-07-19 21:42:22.323802 7f498affd700 20 JournalRecorder: close_and_advance_object_set: closing active object set 0
Note, when running bench-write with a smaller io count (but still large enough to advance object set) the object set is advanced using different code path:
2016-07-19 21:44:11.012319 7f16017fa700 20 JournalRecorder: handle_overflow: object journal_data.31.37142ae8944a.0 overflowed 2016-07-19 21:44:11.012322 7f16017fa700 20 JournalRecorder: close_and_advance_object_set: closing active object set 0
and it does not get stuck.
Updated by Jason Dillaman almost 8 years ago
@Mykola: do you think this is still and issue after the recent changes introduced in the reduce memory footprint branch?
Updated by Jason Dillaman over 7 years ago
- Status changed from New to Need More Info
Updated by Mykola Golub over 7 years ago
- Status changed from Need More Info to Resolved
Yes, it is not reproduced any more, after #16223 was resolved.
Actions