Comment 11 for bug 1283035

Reproduced the same issue on ISO # 338 (4.1):

{"build_id": "2014-06-03_00-40-36", "mirantis": "yes", "build_number": "338", "nailgun_sha": "a828d6b7610f872980d5a2113774f1cda6f6810b", "ostf_sha": "c959aa55f83fe2555cf2d382559271c7a9b17467", "fuelmain_sha": "15805b14f36fc0a896f136051a3be10fea4611d9", "astute_sha": "55df06b2e84fa5d71a1cc0e78dbccab5db29d968", "release": "4.1B", "fuellib_sha": "377f27df4cb008b400b69013dbd606e6225f5c04"}

HA + Ubuntu + NeutronVlan + Ceph for Glance
3 Controllers + 1 Compute + 1 Cinder + 2 Ceph

After successful deployment all controllers have errors in cinder logs:

http://paste.openstack.org/show/82565/

and creating of new volume fails with the following errors:

<158>Jun 3 09:38:20 node-1 cinder-eventlet.wsgi.server INFO: 192.168.0.2 - - [03/Jun/2014 09:38:20] "GET /v1/dfa6efa33ac945d586de917d3b4ae151/limits HTTP/1.1" 200 266 0.009824
<155>Jun 3 09:38:20 node-1 cinder-cinder.scheduler.filters.capacity_filter ERROR: Free capacity not set: volume node info collection broken.
<155>Jun 3 09:38:20 node-1 cinder-cinder.scheduler.filters.capacity_filter ERROR: Free capacity not set: volume node info collection broken.
<155>Jun 3 09:38:20 node-1 cinder-cinder.scheduler.filters.capacity_filter ERROR: Free capacity not set: volume node info collection broken.
<155>Jun 3 09:38:20 node-1 cinder-cinder.volume.flows.create_volume ERROR: Failed to schedule_create_volume: No valid host was found.

I guess we should backport fix for this bug to the 4.1 branch.