Ceph » RADOS

Running latest master without a manager daemon makes `osd destroy` commands hang.

ceph version 14.0.0-1906-g637bb23 (637bb23580a98489a60ef56a09fd4cb39923eee0) nautilus (dev)

Seems that having a manager daemon is absolutely required (wasn't for ceph-volume tests up until July). After this was merged: https://github.com/ceph/ceph/pull/22717
ceph-volume tests started hanging on this command:

 /usr/bin/python2.7 /bin/ceph osd destroy osd.2 --yes-i-really-mean-it

There are a couple of problems with this:

• There is no output at all
• The command hangs forever

Running the hanging command with --verbose doesn't have a human readable reason why this hangs:

bestcmds_sorted:
[{u'cmd026': {u'avail': u'cli,rest',
              u'flags': 8,
              u'help': u'mark osd as being destroyed. Keeps the ID intact (allowing reuse), but removes cephx keys, config-key data and lockbox keys, rendering data permanently unreadable.',
              u'module': u'osd',
              u'perm': u'rw',
              u'sig': [argdesc(<class 'ceph_argparse.CephPrefix'>, name=prefix, req=True, n=1, prefix=osd),
                       argdesc(<class 'ceph_argparse.CephPrefix'>, name=prefix, req=True, n=1, prefix=destroy),
                       argdesc(<class 'ceph_argparse.CephOsdName'>, name=id, req=True, n=1),
                       argdesc(<class 'ceph_argparse.CephString'>, name=sure, req=False, n=1)]}}]
Submitting command:  {'prefix': u'osd destroy', u'sure': u'--yes-i-really-mean-it', u'id': 2, 'target': ('mgr', '')}