Deleting an instance doesn't check task_state properly
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Confirmed
|
Undecided
|
Unassigned |
Bug Description
Issuing a delete on an instance with any task_state other than None should either fail, or do cleanup as required. For example, when taking a snapshot of an instance, task_state goes to image_snapshot, but during the snapshot you can issue a delete on the instance, and task_state goes to deleting as the resources are removed. If you do it quickly, while the snapshot is running, it ends up deleting the backing disk before the snapshot is complete and the snapshot just hangs in SAVING status. I also saw state transition go image_snapshot -> image_pending_
Steps to reproduce:
1. Create instance
2. Create snapshot
3. Delete instance while snapshot is running
4. Hung snapshot
This is same as bug 1143659