mysqlchk for galera

Based on ClusterControl's mysqlchk script but built for Galera. It detects the Galera status on the cluster node as per below:

• if wsrep_cluster_status is Primary,

• return 'MySQL is running.'

• else

• return 'MySQL is *down*'

How to use

1. Install HAProxy on the selected node.

2. Grab the file galerachk.mysql from this repo and place it inb /usr/local/sbin/

3. Update required informations :

MYSQL_HOST="localhost"
MYSQL_PORT="3306"
MYSQL_USERNAME='user'
MYSQL_PASSWORD='password'

4. You are good to go. Now ensure in HAProxy you have 2 listeners (3307 for write, 3306 for read) and use option tcp-check & tcp-check expect to distinguish master/slave similar to example below:

listen  galera
        bind *:3306
        mode tcp
        #timeout client  10800s
        #timeout server  10800s
        balance leastconn
        option tcp-check
        tcp-check expect string is\ running.
        default-server port 9200 inter 2s downinter 5s rise 3 fall 2 # slowstart 60s maxconn 64 maxqueue 128 weight 100
        server node1 192.168.55.111:3306 check
        server node2 192.168.55.112:3306 check
        server node3 192.168.55.113:3306 check

maxscalechk

Based on ClusterControl's mysqlchk script but built for Maxscale. It detects the Maxscale status on the proxy node as per below:

• if maxctrl list servers has return code 0,

• return 'Maxscale is running.'

• else

• return 'Maxscale is *down*'

How to use

1. Install HAProxy on the selected node.

2. Grab the maxscalechk.service file from this repo and place it in /usr/local/sbin/ of the maxscale servers

3. Update required informations :

MXS_USERNAME='user'
MXS_PASSWORD='password'

4. Deploy the xinetd service on the maxscale machines as per below : add maxscalechk file to /etc/xinetd.d directory on maxscale servers

5. link the new service and a port edit /etc/services and add the content of etc.services.maxscale on maxscale servers

6. You are good to go.use option tcp-check & tcp-check expect :

listen  maxcale
        bind *:3306
        mode tcp
        #timeout client  10800s
        #timeout server  10800s
        balance leastconn
        option tcp-check
        tcp-check expect string is\ running.
        default-server port 9200 inter 1s downinter 5s rise 3 fall 2 # slowstart 60s maxconn 64 maxqueue 128 weight 100
        server node1 192.168.55.111:3306 check
        server node2 192.168.55.112:3306 check