Failed to detach volume because of volume not found error prevents vm teardown
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Cinder |
Fix Released
|
High
|
Nikola Đipanov | ||
tempest |
Fix Released
|
High
|
Sean Dague |
Bug Description
When running the boto tests in the gate we get a periodic race on Cinder volumes that can be seen as follows:
Attach the volume (vdh) - http://
Detach the volume (vdh) - http://
Stack trace horribly because the volume is not found - http://
Stack trace horribly because we try again? - http://
Because of this we end up with a volume in an undeletable state, and the tests fail in making it go away (it remains marked in-use even though the guest that was using it is gone) - http://
tags: | added: volumes |
tags: | added: ec2 |
Changed in nova: | |
importance: | Undecided → High |
tags: | added: gate-failure |
Changed in nova: | |
assignee: | nobody → Nikola Đipanov (ndipanov) |
Changed in cinder: | |
assignee: | nobody → Nikola Đipanov (ndipanov) |
Changed in nova: | |
status: | New → Confirmed |
Changed in cinder: | |
importance: | Undecided → High |
Changed in tempest: | |
assignee: | nobody → Sean Dague (sdague) |
status: | New → Fix Released |
importance: | Undecided → High |
Possibly a Tempest bug as well, as maybe there is a better tear down path here.