5.5 nodes won't gracefully shutdown

Hi,

I have a 3 node test cluster that I’m trying to upgrade to 5.6.

Unfortunately, I can’t even get the nodes to stop gracefully so I can upgrade them.

  • ‘/etc/init.d/mysql stop’ returns “fail” after 5 minutes.

  • ‘service mysql stop’ just seems to call /etc/init.d/mysql stop

  • ‘mysqladmin shutdown’ returns immediately without any message

  • ‘kill -TERM’ doesn’t work

The only thing that works is kill -9 which doesn’t leave things in a good state for an upgrade.

This cluster has practically no load at all, maybe 1-5 queries per second.

It doesn’t appear to be a problem with innodb either:

Innodb_buffer_pool_bytes_dirty=0

I’m using the latest packages for Ubuntu trusty.

Source: percona-xtradb-cluster-5.5
Version: 5.5.39-25.11-816.trusty

Any ideas?

I suspect the wrong /etc/mysql/debian.cnf credentials that break the Ubuntu’s init.d script. Any sign of shutdown sequence in the MySQL error log (or syslog if separate log is not enabled)?

No, there’s nothing in the logs. It’s like the command is being discarded or ignored. I did verify the debian.cnf credentials. I opened a bug report on it since I haven’t been able to make any progress:

https://bugs.launchpad.net/percona-xtradb-cluster/+bug/1375897

Ah, so this seems to be the handlersocket problem then. Can you post your full my.cnf? Did actually use the handlersocket port before the problems with shutdown started?

I’m having the same issue - http://www.percona.com/forums/questions-discussions/mysql-and-percona-server/percona-server-5-6/29379-5-6-21-70-1-not-stopping-cleanly - was this ever resolved?