tripleo-quickstart multinode jobs fail with openvswitch 2.8
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tripleo |
Fix Released
|
Medium
|
Alfredo Moralejo |
Bug Description
When trying to update openvswitch to 2.8 in RDO, which is needed by networking-ovn team, multinode oooq jobs are failing in step "[undercloud-setup : Run the TripleO-CI VXLAN networking script]", when installing openvswitch in subnodes with following error [1]:
2017-11-06 09:47:04 | + ssh -tt -o PasswordAuthent
2017-11-06 09:47:04 | Warning: Permanently added '192.168.100.103' (ECDSA) to the list of known hosts.
2017-11-06 09:47:04 | Loaded plugins: fastestmirror, priorities
2017-11-06 09:47:04 | Repository centos-ceph-jewel is listed more than once in the configuration
2017-11-06 09:47:04 | Loading mirror speeds from cached hostfile
2017-11-06 09:47:05 | 2 packages excluded due to repository priority protections
2017-11-06 09:47:05 | Package 1:openvswitch-
2017-11-06 09:47:05 | Nothing to do
2017-11-06 09:47:05 | A dependency job for openvswitch.service failed. See 'journalctl -xe' for details.
2017-11-06 09:47:05 | Connection to 192.168.100.103 closed.
Checking at subnode logs, i've found that openvswitch is failing to start openvswitch services [2]. The root cause is that we are installing openvswitch but not openstack-selinux wich is required to start openvswitch services since version 2.8.
[1] https:/
[2] https:/
Proposed fix in https:/ /review. openstack. org/#/c/ 518037/