Comment 14 for bug 1398128

Revision history for this message
Adam Gandelman (gandelman-a) wrote :

So the merged devstack patch does not seem to help much, still the same failure altho ssh key creation has moved much earlier in the timeline:

# devstack
2014-12-09 18:13:42.602 | ++ configure_ironic_auxiliary # ssh key setup
2014-12-09 18:16:11.869 | ++ enroll_nodes
2014-12-09 18:16:36.082 | ++ '[' 3 -ge 3 ']' # nova reports hypervisor count=3

# Tempest first instance fail
Details: Server 6b7c6c7e-1e86-4f8f-a19f-8d93c350d2a0 failed to build and is in ERROR status
Details: {u'message': u'No valid host was found. There are not enough hosts available.', u'code': 500, u'created': u'2014-12-09T18:17:20Z'}

# first n-cpu periodic to pick up node *count*
2014-12-09 18:16:34.322 30099 DEBUG nova.openstack.common.periodic_task [-] Running periodic task ComputeManager.update_available_resource run_periodic_tasks

# first n-cpu periodic to pick up node *resources*
2014-12-09 18:17:34.841 30099 DEBUG nova.openstack.common.periodic_task [-] Running periodic task ComputeManager.update_available_resource run_periodic_tasks

It looks as if the first periodic sync that picks up the enrolled nodes is still missing parameters that prevent the driver from including its node resources in nova's resource tracker. Some debug logging in Ironic would help here if its proving to be this non-obvious.