Rubygem-Staypuft: After deploying HA neutron, neutron-openvswitch-agent reported failed. pcs resource cleanup <ID> removes it from the failed list. Environment: ruby193-rubygem-foreman_openstack_simplify-0.0.6-8.el6ost.noarch openstack-foreman-installer-2.0.20-1.el6ost.noarch rhel-osp-installer-0.1.10-2.el6ost.noarch openstack-puppet-modules-2014.1-20.el6ost.noarch Steps to reproduce: 1. Deploy HA neutron with Tenant Network Type=VXlan and ceph driver backend for both: glance and cinder. 2. Run pcs status on one controller. Result: Failed actions: neutron-openvswitch-agent_start_0 on mac047d7b61765c.example.com 'not running' (7): call=196, status=complete, last-rc-change='Thu Aug 14 18:55:23 2014', queued=4ms, exec=2001ms Note: Runing pcs resource cleanup makes this entry disappear. Expected result: No failures.
Created attachment 926902 [details] openvswitch-agent log from one controller.
I suspect this is another manifestation of the core rabbit messaging issues described in bug 1129242, but I'm not confident enough to mark it a dupe.
Based on Feedback from Eck, this is not reproducible. Please reopen if you reproduce it.