Comment 4 for bug 1492028

I've dug a little into two instances of this
====
http://logs.openstack.org/98/201398/15/check-tripleo/gate-tripleo-ironic-overcloud-f21puppet-nonha/ceda1a6/logs/
root 5776 5148 0 08:27 ? 00:00:00 /bin/systemctl start openstack-nova-compute.service
In this one puppet was stuck starting the nova-compute service for about 24 minutes

=====
http://logs.openstack.org/57/213957/15/check-tripleo/gate-tripleo-ironic-overcloud-f21puppet-nonha/f927ef4/logs/
root 4942 4314 0 08:31 ? 00:00:00 /bin/systemctl start openstack-nova-compute.service
In this one puppet was stuck starting the nova-compute service for about 23 minutes,

I the nova compute logs we see these every 2 seconds for less then a minute
Sep 04 09:32:15 overcloud-novacompute-0 nova-compute[4955]: 2015-09-04 08:32:15.178 4955 ERROR oslo.messaging._drivers.impl_rabbit [req-bbd45241-28d2-4ba0-ad2f-5565442c29a8 - - - - -] AMQP server on 192.0.2.5:56

then followed by 20 minutes of
Sep 04 09:33:03 overcloud-novacompute-0 nova-compute[4955]: 2015-09-04 08:33:03.491 4955 WARNING nova.conductor.api [req-bbd45241-28d2-4ba0-ad2f-5565442c29a8 - - - - -] Timed out waiting for nova-conductor. Is

I don't see any sign of any nova services running on the controller