Ask Your Question

Why is cinder trying to talk to glance on my compute node?

asked 2014-03-28 00:41:55 -0500

tmac_elasticbox gravatar image

Hello all,

I'm a first time Openstack installer here so apologies if this is a simple question. After following the install instructions and searching for hours I'm asking here.

I've created a 2 node install of Openstack.

Node 1 ( Controller that hosts Nova, Glance, Heat, Ceilometer, Cinder Node 2 ( Compute node that hosts nova-compute, nova-network, cinder-volume

Everything works fine until I try to create a cinder volume from an image (I can create an empty cinder volume)

When I attempt to create a volume from an image I get this error in /var/log/cinder/cinder-scheduler.log on the controller host:

2014-03-27 22:16:06.456 1452 ERROR cinder.scheduler.filter_scheduler [req-5bb090ad-7557-4b14-9bc8-b91b4b72b902 84199e9581e34ce488de770bc40c88f2 3866e3db3b8745e5be727088f8c63cfc] Error scheduling None from last vol-service: compute1 : [u'Traceback (most recent call last):\n', u' File "/usr/lib/python2.7/dist-packages/cinder/taskflow/patterns/", line 172, in run_it\n result = runner(context, args, *kwargs)\n', u' File "/usr/lib/python2.7/dist-packages/cinder/taskflow/", line 260, in __call__\n self.result = self.task(args, * kwargs)\n', u' File "/usr/lib/python2.7/dist-packages/cinder/volume/flows/create_volume/", line 1163, in __call__\n \'image_meta\':, image_id),\n', u' File "/usr/lib/python2.7/dist-packages/cinder/image/", line 224, in show\n _reraise_translated_image_exception(image_id)\n', u' File "/usr/lib/python2.7/dist-packages/cinder/image/", line 222, in show\n image =, \'get\', image_id)\n', u' File "/usr/lib/python2.7/dist-packages/cinder/image/", line 178, in call\n raise exception.GlanceConnectionFailed(reason=str(e))\n', u'GlanceConnectionFailed: Connection to glance failed: Error communicating with [Errno 111] ECONNREFUSED\n'] 2014-03-27 22:16:06.457 1452 ERROR cinder.volume.flows.create_volume [req-5bb090ad-7557-4b14-9bc8-b91b4b72b902 84199e9581e34ce488de770bc40c88f2 3866e3db3b8745e5be727088f8c63cfc] Failed to schedule_create_volume: No valid host was found. Exceeded max scheduling attempts 3 for volume None

It's clearly trying to contact glance on the compute node (, which will fail since glance is installed on the controller node instead.

My question is should it be doing that? I.E. is the problem that a) I need to install glance on the compute node or is the problem that b) I've got cinder misconfigured somehow and it should be trying to talk to the controller instead.

If its "a" then what should I install. The openstack docs left that out. If its "b" then can anyone give me a hint on what I need to change to fix this?

Thanks Everyone!


edit retag flag offensive close merge delete

2 answers

Sort by ยป oldest newest most voted

answered 2014-03-28 01:03:08 -0500

unmesh-gurjar gravatar image

IMO, you need to re-configure cinder and set the value of 'glance_host' parameter (in cinder.conf) to your controller IP (i.e. I guess, this parameter is not specified in the cinder.conf; so its ending up with the node #2 IP. After re-configuring, restart the cinder services so that the configuration change is picked up!

edit flag offensive delete link more

answered 2015-01-26 16:02:34 -0500

flovax gravatar image

That was exactly my error. Thanks a lot :-)

edit flag offensive delete link more


Great! Can you please mark the above answer as 'correct' by clicking on the 'tick' button below that answer and close the question! Thanks.

unmesh-gurjar gravatar imageunmesh-gurjar ( 2015-01-27 06:13:38 -0500 )edit

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2014-03-28 00:41:55 -0500

Seen: 2,267 times

Last updated: Mar 28 '14