[Pacemaker] 4 node cluster help needed

Van Tonder, Raymond RVanTonder1 at fnb.co.za
Thu Nov 3 04:04:56 EDT 2011

On 2011-11-02 16:41, Van Tonder, Raymond wrote:

>Um, that's not consistent with the config. What you're running is a
>single-instance MySQL LSB primitive; for a Master/Slave replication
>managed by Pacemaker you would use the OCF resource agent
>(ocf:heartbeat:mysql) and wrap it in a master/slave set.

I'm running these in openVZ containers with shared disk space for the
Disk access is controlled by the openvz system.
All I need pacemaker to do is to control the active DB instance 

>> And 
>> 2 nodes running jboss server active/active 

>There is no trace of that in the config either.

As I said early stages .. so far I've done the db nodes and is starting
to add the jboss nodes

>> What do I do so that pacemaker wont look for a "mysqlcl_monitor_0" on
>> the servers that will be hosting the jboss.

>IIRC that's a no-can-do in a symmetric cluster; Pacemaker will pretty
>much always execute the probe.

On http://www.clusterlabs.org/ the picture of the first example,
active/passive has apps running on some hosts and not on others.
Does this mean all the host must potentially be able to run any of the
applications just so that the probe won't return a problem?

>> primitive ClusterIP ocf:heartbeat:IPaddr2 \
>> 	params ip="" cidr_netmask="32" \

>Er, no. That cidr_netmask is incorrect.
That is the netmask in "clusters_from_scratch" and is working fine atm

>> 	stonith-enabled="false"

>Bad idea.

This is a sandbox area where I'm starting to setup the cluster for proof
of concept.
Once I'm done with setup I'll enable it egain.


To read FirstRand Bank's Disclaimer for this email click on the following address or copy into your Internet browser: 

If you are unable to access the Disclaimer, send a blank e-mail to
firstrandbankdisclaimer at fnb.co.za and we will send you a copy of the Disclaimer.

More information about the Pacemaker mailing list