cinder.volume.manager: Error encountered during initialization of driver: RBDDriver

[Anastasia Palkina]

"build_id": "2014-08-14_19-52-36",
"mirantis": "yes",
"build_number": "170",
"ostf_sha": "09b6bccf7d476771ac859bb3c76c9ebec9da9e1f",
"nailgun_sha": "af3d1922bfc21345f81be3454115ab6139675c35",
"production": "docker",
"api": "1.0",
"fuelmain_sha": "fd58828f404e4298ed338e8f44c6a326cebd31de",
"astute_sha": "6db5f5031b74e67b92fcac1f7998eaa296d68025",
"release": "5.0.1",
"fuellib_sha": "a31dbac8fff9cf6bc4cd0d23459670e34b27a9ab"

1. Create new environment (CentOS, simple mode)
2. Choose nova-network, vlan manager
3. Choose both Ceph
4. Add 1 controller+ceph, 2 compute+ceph, 1 cinder+ceph
5. Start deployment. It was successful
6. Start OSTF tests. It was successful
7. But there are errors and warnings in cinder-volume.log on node-10:

14-08-15 10:16:11 WARNING

cinder.volume.manager [-] Unable to update stats, RBDDriver -1.1.0 driver is uninitialized.

2014-08-15 10:15:11 WARNING

cinder.volume.manager [req-b2b37809-7f57-48da-85d4-5f188a1ae05c - - - - -] Unable to update stats, RBDDriver -1.1.0 driver is uninitialized.

2014-08-15 10:14:51 ERROR

cinder.volume.manager [req-65ff4739-4e3a-4bb1-8558-91f20667d4b5 - - - - -] Bad or unexpected response from the storage volume backend API: error connecting to ceph cluster

2014-08-15 10:14:51 ERROR

cinder.volume.manager [req-65ff4739-4e3a-4bb1-8558-91f20667d4b5 - - - - -] Error encountered during initialization of driver: RBDDriver

2014-08-15 10:14:51 ERROR

cinder.volume.drivers.rbd [req-65ff4739-4e3a-4bb1-8558-91f20667d4b5 - - - - -] error connecting to ceph cluster

[Stanislaw Bogatkin (sbogatkin)]

Problem is that node with cinder not calling ceph to create keyring.
Workaround:
# ceph auth get-or-create client.volumes > /etc/ceph/ceph.client.volumes.keyring
# /etc/init.d/openstack-cinder-volume restart

After that cinder working correct:
[root@node-10 ~]# tail -n 15 /var/log/cinder-all.log
<158>Aug 18 14:22:50 node-10 cinder-cinder.openstack.common.service INFO: Caught SIGTERM, stopping children
<158>Aug 18 14:22:50 node-10 cinder-cinder.openstack.common.service INFO: Waiting on 1 children to exit
<158>Aug 18 14:22:50 node-10 cinder-cinder.openstack.common.service INFO: Caught SIGTERM, exiting
<158>Aug 18 14:22:50 node-10 cinder-cinder.openstack.common.service INFO: Child 1066 exited with status 1
<158>Aug 18 14:22:51 node-10 cinder-cinder.openstack.common.service INFO: Starting 1 workers
<158>Aug 18 14:22:51 node-10 cinder-cinder.openstack.common.service INFO: Started child 1520
<156>Aug 18 14:22:51 node-10 cinder-cinder.service AUDIT: Starting cinder-volume node (version 2014.1.1)
<158>Aug 18 14:22:51 node-10 cinder-cinder.volume.manager INFO: Starting volume driver RBDDriver (1.1.0)
<158>Aug 18 14:22:51 node-10 cinder-cinder.volume.manager INFO: Updating volume status
<158>Aug 18 14:22:51 node-10 cinder-oslo.messaging._drivers.impl_rabbit INFO: Connected to AMQP server on 192.168.0.2:5672
<158>Aug 18 14:22:51 node-10 cinder-oslo.messaging._drivers.impl_rabbit INFO: Connected to AMQP server on 192.168.0.2:5672
<158>Aug 18 14:22:51 node-10 cinder-oslo.messaging._drivers.impl_rabbit INFO: Connected to AMQP server on 192.168.0.2:5672
<158>Aug 18 14:23:00 node-10 cinder-cinder.volume.manager INFO: Updating volume status
<158>Aug 18 14:24:00 node-10 cinder-cinder.volume.manager INFO: Updating volume status
<158>Aug 18 14:25:00 node-10 cinder-cinder.volume.manager INFO: Updating volume status

[Andrew Woodward (xarses)]

We won't be fixing low or medium bugs for 5.0.x any longer. If this needs to be addressed we need to increase priority