Bug #46769
openqa: Refactor cephfs creation/removal code.
0%
Description
The 'CephFSTestCase' class creates the filesystem based on
'REQUIRE_FILESYSTEM' flag in 'setUp' but the corresponding
'tearDown' function doesn't cleanup the same.
The 'setUp' routine also cleans up stale filesystems if present.
So the filesystem created for a test case, doesn't get cleaned
up during tearDown of same test case but instead gets cleaned
up with the beginning of next test case. Though there is no
functional impact, cleaning it up in tearDown is neat.
There is a duplicate dead code of fs creation/removal code in
test_volumes.py. That also can be removed.
Updated by Kotresh Hiremath Ravishankar over 3 years ago
- Status changed from New to In Progress
- Assignee set to Kotresh Hiremath Ravishankar
Updated by Kotresh Hiremath Ravishankar over 3 years ago
- Component(FS) qa-suite added
Updated by Kotresh Hiremath Ravishankar over 3 years ago
- Pull request ID set to 36368
Updated by Kotresh Hiremath Ravishankar over 3 years ago
- Status changed from In Progress to Fix Under Review
Updated by Kotresh Hiremath Ravishankar over 3 years ago
- Subject changed from qa: Refactor cephfs creation/removal code. to qa: Remove dead code(volume creation part) in test_volumes.py
Updated by Kotresh Hiremath Ravishankar over 3 years ago
Based on comment https://github.com/ceph/ceph/pull/36368#pullrequestreview-458486627, retaining the behavior of cleaning up volumes at the beginning of the setUp rather than in the tearDown routine. So this tracker boils down to the removal of dead code which creates volume in test_volumes.py as that already happens in parent class.
Updated by Kotresh Hiremath Ravishankar over 3 years ago
- Subject changed from qa: Remove dead code(volume creation part) in test_volumes.py to qa: Refactor cephfs creation/removal code.
Updated by Patrick Donnelly over 3 years ago
Actually, here's a test failure where we get:
2020-11-25T03:40:33.794 INFO:teuthology.orchestra.run.smithi040.stdout:2020-11-25T03:39:55.686437+0000 mon.a (mon.0) 1793 : cluster [WRN] Replacing daemon mds.a as rank 0 with standby daemon mds.b
From: /ceph/teuthology-archive/pdonnell-2020-11-24_19:01:27-fs-wip-pdonnell-testing-20201123.213848-distro-basic-smithi/5654380/teuthology.log
During that time, we were just turning the daemons off. If we had deleted the fs first, the MDS replacement would not have happened.