Initialization of realm, zonegroup, zones etc. may also be necessary in a single-site scenario

[Frode Nordahl]

In a focal-ussuri deploy using -next charms with a single ceph-radosgw unit I got into a situation where neither realm, zonegroup or any zones existed.

The charm currently assumes that realm and zonegroup will always exist in a single-site scenario and as such it was not able to recover.

It may be that we need to move the logic in the master-relation-joined hook into the regular mon-relation-changed hook so that the charm can deal with the cluster regardless of its state.

[Frode Nordahl (fnordahl)]

Unit log:
2021-08-26 05:47:26 INFO juju unit_agent.go:253 Starting unit workers for "ceph-radosgw/0"
2021-08-26 05:47:26 INFO juju.worker.apicaller connect.go:158 [bb06d3] "unit-ceph-radosgw-0" successfully connected to "10.246.115.255:17070"
2021-08-26 05:47:26 INFO juju.worker.apicaller connect.go:255 [bb06d3] password changed for "unit-ceph-radosgw-0"
2021-08-26 05:47:27 INFO juju.worker.apicaller connect.go:158 [bb06d3] "unit-ceph-radosgw-0" successfully connected to "10.246.115.255:17070"
2021-08-26 05:47:27 INFO juju.worker.migrationminion worker.go:140 migration phase is now: NONE
2021-08-26 05:47:27 INFO juju.worker.logger logger.go:120 logger worker started
2021-08-26 05:47:27 INFO juju.worker.upgrader upgrader.go:219 no waiter, upgrader is done
2021-08-26 05:47:27 INFO juju.worker.meterstatus runner.go:89 skipped "meter-status-changed" hook (missing)
2021-08-26 05:47:27 INFO juju.worker.uniter uniter.go:328 unit "ceph-radosgw/0" started
2021-08-26 05:47:27 INFO juju.worker.uniter uniter.go:629 resuming charm install
2021-08-26 05:47:27 INFO juju.worker.uniter.charm bundles.go:79 downloading cs:~openstack-charmers-next/ceph-radosgw-407 from API server
2021-08-26 05:47:27 INFO juju.worker.uniter uniter.go:346 hooks are retried true
2021-08-26 05:47:27 INFO juju.worker.uniter.storage resolver.go:127 initial storage attachments ready
2021-08-26 05:47:27 INFO juju.worker.uniter resolver.go:148 found queued "install" hook
2021-08-26 05:47:38 INFO unit.ceph-radosgw/0.juju-log server.go:314 Installing python3-psutil with options: ['--option=Dpkg::Options::=--force-confold']
2021-08-26 05:47:42 INFO unit.ceph-radosgw/0.juju-log server.go:314 Registered config file: /etc/haproxy/haproxy.cfg
2021-08-26 05:47:42 INFO unit.ceph-radosgw/0.juju-log server.go:314 Registered config file: /etc/ceph/ceph.conf
2021-08-26 05:47:44 INFO unit.ceph-radosgw/0.juju-log server.go:314 Installing Apache
2021-08-26 05:47:44 INFO unit.ceph-radosgw/0.juju-log server.go:314 Installing ['apache2'] with options: ['--option=Dpkg::Options::=--force-confold']
2021-08-26 05:47:54 INFO unit.ceph-radosgw/0.juju-log server.go:314 Disabling unused Apache sites
2021-08-26 05:47:54 INFO unit.ceph-radosgw/0.juju-log server.go:314 Restarting Apache
2021-08-26 05:47:54 WARNING unit.ceph-radosgw/0.install logger.go:60 Job for apache2.service failed because the control process exited with error code.
2021-08-26 05:47:54 WARNING unit.ceph-radosgw/0.install logger.go:60 See "systemctl status apache2.service" and "journalctl -xe" for details.
2021-08-26 05:47:54 INFO unit.ceph-radosgw/0.juju-log server.go:314 Installing ['haproxy', 'radosgw', 'apache2'] with options: ['--option=Dpkg::Options::=--force-confold']
2021-08-26 05:48:12 WARNING unit.ceph-radosgw/0.juju-log server.go:314 Package libapache2-mod-fastcgi has no installation candidate.
2021-08-26 05:48:12 WARNING unit.ceph-radosgw/0.install logger.go:60 radosgw.service is not a native service, redirecting to systemd-sysv-install.
2021-08-26 05:48:12 WARNING unit.ceph-radosgw/0.install logger.go:60 Executing: /lib/systemd/systemd-sysv-install disable radosgw
2021-08-26 05:48:13 WARNING unit.ceph-radosgw/0.install logger.go:60 Cre...