status is active while mysqld is not running

[Pascal Mazon]

I'm running a simple Openstack bundle (see attachment) that uses percona-cluster.
After everything is set up, I can normally access to my setup for a while.
But after a time (or maybe related to the number/type of commands typed?) I can no longer proceed with my `nova` commands:

$ nova list
ERROR (InternalServerError): An unexpected error prevented the server from fulfilling your request. (HTTP 500) (Request-ID: req-7590ef8b-d762-41d2-92d2-2c5e5502eb7c)

I noticed that on my percona-cluster container, the mysql service is marked as active (yet exited), but there's no longer any mysqld process running:

$ juju status mysql
App Version Status Scale Charm Store Rev OS Notes
mysql 5.6.21-25.8 active 1 percona-cluster jujucharms 246 ubuntu

Unit Workload Agent Machine Public address Ports Message
mysql/0* active idle 0/lxd/1 10.19.0.9 Unit is ready

$ juju ssh mysql/0
ubuntu@juju-896271-0-lxd-1:~$ sudo systemctl status mysql
● mysql.service - LSB: Start and stop the mysql (Percona XtraDB Cluster) daemon
   Loaded: loaded (/etc/init.d/mysql; bad; vendor preset: enabled)
   Active: active (exited) since Tue 2016-11-22 16:27:25 UTC; 17h ago
     Docs: man:systemd-sysv-generator(8)
  Process: 12766 ExecStop=/etc/init.d/mysql stop (code=exited, status=0/SUCCESS)
  Process: 13381 ExecStart=/etc/init.d/mysql start (code=exited, status=0/SUCCESS)
    Tasks: 0
   Memory: 0B
      CPU: 0

Nov 22 16:27:24 juju-896271-0-lxd-1 systemd[1]: Starting LSB: Start and stop the mysql (Percona XtraDB Cluster) daemon...
Nov 22 16:27:25 juju-896271-0-lxd-1 mysql[13381]: * Starting MySQL (Percona XtraDB Cluster) database server mysqld
Nov 22 16:27:25 juju-896271-0-lxd-1 mysql[13381]: ...done.
Nov 22 16:27:25 juju-896271-0-lxd-1 systemd[1]: Started LSB: Start and stop the mysql (Percona XtraDB Cluster) daemon.
Nov 23 09:20:50 juju-896271-0-lxd-1 systemd[1]: Started LSB: Start and stop the mysql (Percona XtraDB Cluster) daemon.
ubuntu@juju-896271-0-lxd-1:~$ ps aux | grep mysqld
ubuntu 5887 0.0 0.0 12944 784 pts/1 S+ 09:46 0:00 grep --color=auto mysqld

When starting the service, the mysqld process running was like this:
/usr/sbin/mysqld --basedir=/usr --datadir=/var/lib/percona-xtradb-cluster --plugin-dir=/usr/lib/mysql/plugin --user=mysql --wsrep-provider=/usr/lib/libgalera_smm.so --wsrep-new-cluster --log-error=/var/log/mysql/error.log --pid-file=/var/run/mysqld/mysqld.pid --socket=/var/run/mysqld/mysqld.sock --port=3306 --wsrep_start_position=00000000-0000-0000-0000-000000000000:-1

If I manually run this command in my container (I don't know how to daemonize it though), I can use my `nova` commands again:

$ nova list
+----+------+--------+------------+-------------+----------+
| ID | Name | Status | Task State | Power State | Networks |
+----+------+--------+------------+-------------+----------+
+----+------+--------+------------+-------------+----------+

[James Page (james-page)]

I think this is related to bug 1664025 where the unit is bootstrapped using the bootstrap-pxc action, but ends up in the wrong cgroup and can be shutdown and not restarted if the juju unit agent is restarted for some reason (a juju upgrade-juju operation for example).

This has just been fixed into the stable percona-cluster charm, so this should now be resolved.