[Pacemaker] How to serialize/control resource startup on Standby node

[Andrew Beekhof]

Try setting the interleave meta option for the ms resources.
This should tell ms_pimd to care about the instance of ms_Tmgr on the same node.

On Thu, Dec 29, 2011 at 2:43 PM, neha chatrath <nehachatrath at gmail.com> wrote:
> Hello,
>
> I have  cluster with 2 nodes with multiple Master/slave resources.
> The ordering of resources on the master node is achieved using order option
> of crm. When standby node started, the processes are started one after the
> another.
> Following is the configuration info:
> primitive ClusterIP ocf:mcg:MCG_VIPaddr_RA \
>         params ip="192.168.113.67" cidr_netmask="255.255.255.0" nic="eth0:1"
> \
>         op monitor interval="40" timeout="20"
> primitive Rmgr ocf:mcg:RM_RA \
>         op monitor interval="60" role="Master" timeout="30"
> on-fail="restart" \
>         op monitor interval="40" role="Slave" timeout="40" on-fail="restart"
> primitive Tmgr ocf:mcg:TM_RA \
>         op monitor interval="60" role="Master" timeout="30"
> on-fail="restart" \
>         op monitor interval="40" role="Slave" timeout="40" on-fail="restart"
> primitive pimd ocf:mcg:PIMD_RA \
>         op monitor interval="60" role="Master" timeout="30"
> on-fail="restart" \
>         op monitor interval="40" role="Slave" timeout="40" on-fail="restart"
> ms ms_Rmgr Rmgr \
>         meta master-max="1" master-max-node="1" clone-max="2"
> clone-node-max="1" notify="true"
> ms ms_Tmgr Tmgr \
>         meta master-max="1" master-max-node="1" clone-max="2"
> clone-node-max="1" notify="true"
> ms ms_pimd pimd \
>         meta master-max="1" master-max-node="1" clone-max="2"
> clone-node-max="1" notify="true"
> colocation ip_with_Rmgr inf: ClusterIP ms_Rmgr:Master
> colocation ip_with_Tmgr inf: ClusterIP ms_Tmgr:Master
> colocation ip_with_pimd inf: ClusterIP ms_pimd:Master
> order TM-after-RM inf: ms_Rmgr:promote ms_Tmgr:start
> order ip-after-pimd inf: ms_pimd:promote ClusterIP:start
> order pimd-after-TM inf: ms_Tmgr:promote ms_pimd:start
> property $id="cib-bootstrap-options" \
>         dc-version="1.0.11-db98485d06ed3fe0fe236509f023e1bd4a5566f1" \
>         cluster-infrastructure="Heartbeat" \
>         no-quorum-policy="ignore" \
>         stonith-enabled="false"
> rsc_defaults $id="rsc-options" \
>         migration_threshold="3" \
>         resource-stickiness="100"
>
> I have a system requirement in which start of resource (e.g. pimd) is
> dependent on successful start of  another resource (e.g. Tmgr)
> Everything run smoothly on the master node. This is due to ordering and few
> seconds delay untill a resource is promoted as Master.
> But on the standby node since the resources are started one after the
> another without any delay , Standby node in the cluster behaves erratically
>
> Is there a way, through which I can serialize/control resource start up on
> the standby node.
>
> Thanks and regards
> Neha Chatrath
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>