mon-relation-changed hook failed, disk initialization failure

[Alberto Donato]

During a 3-node ceph deployment, I got a "mon-relation-changed" hook error on one unit:

2015-01-15 10:48:33 INFO mon-relation-changed OSError: [Errno 16] Device or resource busy: '/var/lib/ceph/tmp/mnt.CnkNwB'
2015-01-15 10:48:33 ERROR juju-log mon:0: Unable to initialize device: /dev/sdb
...
2015-01-15 10:48:33 INFO mon-relation-changed subprocess.CalledProcessError: Command '['ceph-disk-prepare', '--fs-type', u'xfs', '--zap-disk', u'/dev/sdb']' returned non-zero exit status 1

Full log is attached.

The /dev/sdb node was present, I manually ran the ceph-disk-prepare command afterwards, and it worked.

[chrone (chrone81)]

I'm having same issue with Ubuntu Landscape Autopilot OpenStack (Ubuntu 14.04.02 LTS). Funny is, the /dev/sdb is there and so is ceph health cluster showing it's healty.

[James Page (james-page)]

I think this may be related to bug 1475247 which splits the zapping of disks into two steps both in ceph itself and in the charms for general disk cleaning.