tempest errors in logs

[Brant Knudson]

This is regarding a tempest run on a Keystone change, here's the log: http://logs.openstack.org/73/111573/3/check/check-tempest-dsvm-full/f4e3313/console.html

All the tempest tests ran successfully. Then it runs the log checker and there are several errors in the logs.

- Log File Has Errors: n-cond

nova.quota - Failed to commit reservations ...

- Log File Has Errors: n-cpu

There's several errors here:

glanceclient.common.http -- Request returned failure status 404.
(there's several of these)

oslo.messaging.rpc.dispatcher -- Exception during message handling: Unexpected task state: expecting (u'powering-off',) but the actual state is None
(this generates a lot of logs and there are several of them)

- Log File Has Errors: n-api

glanceclient.common.http - Request returned failure status 404.
(there's several of these)

- Log File Has Errors: g-api

glance.store.sheepdog [-] Error in store configuration: [Errno 2] No such file or directory
swiftclient [-] Container HEAD failed: http://127.0.0.1:8080/v1/AUTH_3c05c27e027f451b9837e04c9d8ae1e5/glance 404 Not Found

- Log File Has Errors: c-api

ERROR cinder.volume.api Volume status must be available to reserve
(There's 4 of these)

- Log File Has Errors: ceilometer-alarm-evaluator

ceilometer.alarm.service [-] alarm evaluation cycle failed
(several of these)

- Log File Has Errors: ceilometer-acentral

ERROR ceilometer.neutron_client [-] publicURL endpoint for network service not found
(there's several errors related to endpoints)

- Log File Has Errors: ceilometer-acompute

ceilometer.compute.pollsters.disk [-] Ignoring instance instance-00000087: internal error: cannot find statistics for device 'virtio-disk1'
(there's a few of these)

[Davanum Srinivas (DIMS) (dims-v)]

Brant, the commit reservations has its own bug. fyi - https://bugs.launchpad.net/nova/+bug/1353131

[Joe Gordon (jogo)]

this catch all bug isn't very useful IMHO, I think we should track each one of these with separate bugs

[gordon chung (chungg)]

i think this is resolved in ceilometer (at least what is mentioned above)... should've been resolved by https://review.openstack.org/#/c/117051/ as well as various other patches that check service before polling. marking as commited.

[John Griffith (john-griffith)]

Some of these have been cleaned up recently under this bug:
https://bugs.launchpad.net/cinder/+bug/1370773

[Sean McGinnis (sean-mcginnis)]

Closing stale bug. If this is still an issue please reopen.