[Pacemaker] ocf:heartbeat:Route failover not working

[Volker Dormeyer]

Hi,

On Thu, Oct 04, 2012 at 11:00:13AM +0100,
tony at intermodesystems.com <tony at intermodesystems.com> wrote:
>  ClusterIP	(ocf::heartbeat:IPaddr2):	Started node01.local
>  ClusterRoute	(ocf::heartbeat:Route):	Started node02.local
> (unmanaged) FAILED
> 
> Failed actions:
>     ClusterRoute_start_0 (node=node02.local, call=4, rc=5,
> status=complete): not installed
>     ClusterRoute_stop_0 (node=node02.local, call=5, rc=5,
> status=complete): not installed

Pacemaker tries to distribute your resources through the available
nodes, by default.

> here is the output from crm configure show:
> 
> [root at node01 ~]# crm configure show
> node node01.local
> node node02.local
> primitive ClusterIP ocf:heartbeat:IPaddr2 \
> 	params ip="192.168.231.217" cidr_netmask="22" nic="bond0.248" \
> 	op monitor interval="30s"
> primitive ClusterRoute ocf:heartbeat:Route \
> 	params destination="0.0.0.0/0" device="bond0.248"
> gateway="192.168.231.209" \
> 	op monitor interval="10" timeout="20" depth="0"
> property $id="cib-bootstrap-options" \
> 	dc-version="1.0.12-unknown" \
> 	cluster-infrastructure="openais" \
> 	expected-quorum-votes="2" \
> 	stonith-enabled="false" \
> 	no-quorum-policy="ignore"
> 
> hope someone can help with this as I have no idea why it tries to
> start the route part on the other node?

You will need a colocation contraint to let pacemaker know to start the
resources together on a single node. Like this:

  colocation colIPwithRoute inf: ClusterIP ClusterRoute

Also, a mandatory order-constraint might make sense to first set the
IP and then the Route. Like this:

  order ordIPthenRoute inf: ClusterIP ClusterRoute

An alternative to the order-constraint might be using a group, instead.

Best Regards,
Volker