Closed jacobwagner closed 10 years ago
Problem exists on centos as well Monit not configured to monitor rpcdaemon:
# Do not edit, changes will be overwritten
check process rpcdaemon matching "^((/usr/bin/)?python\d? )?(/usr/bin/)?rpcdaemon\b"
start program = "/sbin/service start"
stop program = "/sbin/service stop"
if 5 restarts within 5 cycles then timeout
On Centos: Rpcdaemon started on one of the nodes and crashed on the other. Scheduled subnets, but never scheduled router. Crashed RPCDaemon never restarted because of monit issue above.
This bug was due to missing attributes in the nova-network
cookbook (where rpcdaemon is located) for the recipe in openstack-monitoring
. It was fixed in master
. Will's merge above should fix it!
This has been fixed, verified by QE
https://gist.github.com/jacobwagner/0d9a3ede8a3a3d5d0c1b
To recreate: install ha with quantum on precise, check for any process with rpcdaemon, also check the log, it will look like the above gist
This causes routers and networks to not be properly scheduled for creation on either controller