[Pacemaker] Trouble swinging a Shared IP (VIP) upon active mysql instance shutdown

[Attila Megyeri]

 <PATRICKZOBLISEIN at ...> writes:

> 
> Hi All,
> I currently have a 2-node Corosync+Pacemaker cluster configured with a 
Shared IP (VIP).  When the active node fails, the VIP swings to the passive 
node and becomes active - this works very well so far.  
> 
> Also - on each node is a MySQL instance - a Master/Master replication 
configuration.
> 
> My issue at the moment, is getting Pacemaker configured properly in order to 
swing the shared ip (VIP) when the active MySQL instance is down.
> 
> Whenever I shutdown the mysql service on EITHER node - Pacemaker restarts 
that instance - instead of restarting, I'd really like to see the VIP swing to 
the other node.
> 
> Below is my crm configuration - I understand that the example_mysql1 
primitive is configured to start the mysql instance - but is there a way to 
say instead of "op start" - "op shared_ip_one"?
> 
> I've only started playing with Pacemaker and Corosync and am still digesting 
concepts - so I'm hoping what I'm trying to accomplish is 'simple' and I'm not 
able to see the forest for the trees.
> 
> 
> Many thanks in advance,
> Patrick
> 
> 
> `crm configure edit`
> 
> 
> node mycluster1.oz.com
> node mycluster2.oz.com
> primitive example_mysql1 ocf:heartbeat:mysql \
>         params binary="mysqld_safe" datadir="/var/lib/mysql/data" 
pid="/var/lib/mysql/data/mysql.pid" socket="/var/lib/mysql/data/mysql.sock" 
log="/var/lib/mysql/data/mysql_cluster.log" \
>         op start interval="0" timeout="120" \
>         op stop interval="0" timeout="120" \
>         op monitor interval="10" timeout="30" depth="0"
> primitive shared_ip_one ocf:heartbeat:IPaddr \
>         params ip="10.4.1.100" cidr_netmask="255.255.255.0" nic="eth0" \
>         op monitor interval="10s"
> ms example_mysql1_master example_mysql1 \
>         meta target-role="Slave"
> location cli-prefer-example_mysql1 example_mysql1_master \
>         rule $id="cli-prefer-rule-example_mysql1" inf: #uname eq 
mycluster2.oz.com
> location cli-prefer-shared_ip_one shared_ip_one \
>         rule $id="cli-prefer-rule-shared_ip_one" inf: #uname eq 
mycluster1.oz.com
> property $id="cib-bootstrap-options" \
>         dc-version="1.0.11-1554a83db0d3c3e546cfd3aaff6af1184f79ee87" \
>         cluster-infrastructure="openais" \
>         expected-quorum-votes="2" \
>         stonith-enabled="false"
> 
> 

Hi Patrick,

I am having the exact issue with Pacemaker + Mysql, namely, that mysql is 
restarted instead of the VIP being moved to the passive node.

I tried all possible conbinations of migration-threshold, no help.

Were you able to resolve the issue with Andrew's assistance?

Thanks.