Skip to content

Latest commit

 

History

History
64 lines (38 loc) · 1.71 KB

restart-mariadb-ha.md

File metadata and controls

64 lines (38 loc) · 1.71 KB

Restarting MariaDB in an HA environment

Before running your database schema upgrades you will need to bring up enough of your environment to have a functioning database. The following commands will bring up the database, associated VIP, and haproxy on a single host so that the database is available at the expected address.

  1. Edit /etc/my.cnf.d/galera.conf on one of your database nodes, and replace:

    wsrep_cluster_address="gcomm://XXX.XXX.XXX.X,XXX.XXX.XXX.X,XXX.XXX.XXX.XX"

    With:

    wsrep_cluster_address="gcomm://"

  2. Start the database resource:

    pcs resource debug-start mysqld

  3. Start the database VIP.

    Look in /etc/keystone/keystone.conf for the connection setting in the database section:

    connection=mysql://keystone:[email protected]/keystone

    The IP address used in that connection string is the database VIP. Start that resource:

    pcs resource debug-start ip-192.168.100.11

  4. Start HAProxy:

    pcs resource debug-start haproxy

  5. Edit /etc/my.cnf.d/galera.conf to undo the change you made it your first step.

At this point, your database service should be running. Confirm that you can connect to it locally:

# mysql -e 'select 1'

And confirm that you can connect to it using the credentials from your Keystone configuration:

# mysql -u keystone -psecret -h 192.168.100.11 -e 'select 1'

After you have restarted your cluster

Once you have successfully restarted all of your HA resources, you will need to revert the change you made earlier to /etc/my.cnf.d/galera.conf.

Replace:

wsrep_cluster_address="gcomm://"

With the original:

wsrep_cluster_address="gcomm://XXX.XXX.XXX.X,XXX.XXX.XXX.X,XXX.XXX.XXX.XX"