Comment 8 for bug 1742683

Reading around it seems that "WSREP has not yet prepared node for application use" may indicate that another node in the cluster exited ungracefully. Sure enough looking at
the supplied crash dump shows:

grep -E "WSREP: caught exception in PC, s|Abort" mysql_*/var/log/mysql/error*
mysql_0/var/log/mysql/error.log:2018-01-11 09:17:39 93347 [ERROR] WSREP: caught exception in PC, state dump to stderr follows:
mysql_1/var/log/mysql/error.log:2018-01-11 09:17:39 67643 [ERROR] WSREP: caught exception in PC, state dump to stderr follows:
mysql_1/var/log/mysql/error.log:2018-01-11 09:17:39 67643 [ERROR] Aborting

$ grep "WSREP has not yet prepared node for application use" mysql_0/var/log/juju/unit-mysql-0.log | head -1
2018-01-11 09:27:16 DEBUG db-admin-relation-changed _mysql_exceptions.OperationalError: (1047, 'WSREP has not yet prepared node for application use')

So far the closest think I can find is: https://jira.percona.com/browse/PXC-1352 but this claims to be fixed in 5.6.15-25.4 and this deploy has 5.6.37-82.2