Stuck in vm_state SHUTOFF and task_state rebuild_spawning - stuck in to_xml() ?
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Fix Released
|
High
|
Unassigned |
Bug Description
See http://
The first test to fail is test_rebuild_
1. rebuild with new image id
2. wait to transition to SHUTOFF/None
3. Test is done, run cleanup operations ...
4. Rebuild to old image id
5. Wait to hit SHUTOFF/None
6. Start the instance again
It is step 5 we get stuck and time out. We make this transition:
2014-02-06 22:29:51,724 State transition "SHUTOFF/
and never transition to ACTIVE/powering-off ... which suggests we get stuck in spawn() somewhere
The last log entry for req-60fa2fbb-
[instance: a3ac8847-
We never get the corresponding 'End to_xml'
Now ... interestingly, there's a resume operation happening around the same time (req-aeedee18-
"An error occurred while trying to launch a defined domain with xml:"
Hmm ... also interestingly, there isn't another "Start to_xml" in the logs after this one - perhaps the stuck thread is holding a lock.
tags: | added: libvirt |
tags: | added: testing |
Changed in nova: | |
importance: | Undecided → High |
status: | New → Confirmed |
Changed in nova: | |
milestone: | none → juno-rc1 |
status: | Fix Committed → Fix Released |
Changed in nova: | |
milestone: | juno-rc1 → 2014.2 |
Smells quite similar to bug #1273478 - i.e. deadlock caused by holding a lock while waiting on the async network info operation to complete, and then the async operation tries to take the same lock