Zaqar service failed, causing node registration to hang
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tripleo |
Triaged
|
Medium
|
Unassigned |
Bug Description
I've seen this a number of times in my local dev environments, but it doesn't reproduce consistently. For some reason the zaqar service is in a failed state, which causes the node registration workflow to hang. In the zaqar logs is the following:
2016-11-09 18:38:08.539 19123 INFO zaqar.transport
2016-11-09 20:34:15.923 19123 WARNING keystonemiddlew
2016-11-09 20:34:16.442 19123 CRITICAL zaqar [(None,) 0b0279336e324cb
2016-11-09 20:34:16.442 19123 ERROR zaqar Traceback (most recent call last):
2016-11-09 20:34:16.442 19123 ERROR zaqar File "/usr/bin/
2016-11-09 20:34:16.442 19123 ERROR zaqar sys.exit(run())
2016-11-09 20:34:16.442 19123 ERROR zaqar File "/usr/lib/
2016-11-09 20:34:16.442 19123 ERROR zaqar _fail(1, ex)
2016-11-09 20:34:16.442 19123 ERROR zaqar File "/usr/lib/
2016-11-09 20:34:16.442 19123 ERROR zaqar print(ex, file=sys.stderr)
2016-11-09 20:34:16.442 19123 ERROR zaqar IOError: [Errno 32] Broken pipe
2016-11-09 20:34:16.442 19123 ERROR zaqar
One possible cause is that I attempted a deployment without registering nodes. This correctly failed, but maybe the error did something that crashed zaqar?
I'll keep bug #1661227 which looks similar.