Unfortunately we are still seeing this with cs:trusty/ceph-radosgw-241 which is that last fix (a timeout bump). The evidence:
- juju status shows:
service-status: {current: blocked, message: 'Services not running that should
be: radosgw', since: '28 Jun 2016 05:26:55Z'}
- last entry in radosgw.log in all rados units:
$ tail ceph-radosgw-{0,1,2}/var/log/ceph/radosgw.log -n 1
==> ceph-radosgw-0/var/log/ceph/radosgw.log <==
2016-06-28 04:40:41.097768 7f4188335700 -1 Initialization timeout, failed to initialize
Unfortunately we are still seeing this with cs:trusty/ ceph-radosgw- 241 which is that last fix (a timeout bump). The evidence: {0,1,2} /var/log/ ceph/radosgw. log -n 1 0/var/log/ ceph/radosgw. log <==
- juju status shows:
service-status: {current: blocked, message: 'Services not running that should
be: radosgw', since: '28 Jun 2016 05:26:55Z'}
- last entry in radosgw.log in all rados units:
$ tail ceph-radosgw-
==> ceph-radosgw-
2016-06-28 04:40:41.097768 7f4188335700 -1 Initialization timeout, failed to initialize
==> ceph-radosgw- 1/var/log/ ceph/radosgw. log <==
2016-06-28 04:40:30.564641 7fabb583f700 -1 Initialization timeout, failed to initialize
==> ceph-radosgw- 2/var/log/ ceph/radosgw. log <==
2016-06-28 04:40:30.608160 7fd42d3e5700 -1 Initialization timeout, failed to initialize
Process listing on all radosgw units confirms no rados process running (file /var/log/ ps-fauxww. txt created by our log collector script)
I'm attaching a tarball with /var/log/* from the rados units, and next I will attach juju status.