[Pacemaker] resource failover in active/active cluster

[Andrew Beekhof]

On 03/04/2013, at 4:07 AM, Charles Mean <charles.mean at gmail.com> wrote:

> Hello guys,
> 
> I am running corosync 1.4.2 and pacemaker 1.1.7 on Debian environments trying to deploy an active/active cluster with two nodes.
> The main problem is that I have two resources that depend on each other, so I have the VIP cloned over the two node and nginx daemon which depends on VIP, but when nginx goes down the VIP onde the failed node is not removed from cluster and that node still answer to the requests.

Nothing in your configuration says that it should.
Why would you clone the IP if nginx is only running on a single host?

> Follow my current configuration:
> node host01
> node host02
> primitive ClusterIP ocf:heartbeat:IPaddr2 \
> 	params ip="192.168.2.100" cidr_netmask="24" nic="eth0" clusterip_hash="sourceip" \
> 	op monitor interval="1s"
> primitive WebSite lsb:nginx \
> 	op monitor interval="1s"
> clone WebIP ClusterIP \
> 	meta globally-unique="true" clone-max="2" clone-node-max="2"
> clone WebSiteClone WebSite
> colocation website-with-ip inf: WebSiteClone WebIP
> order nginx-after-ip inf: WebIP WebSiteClone
> property $id="cib-bootstrap-options" \
> 	dc-version="1.1.7-ee0730e13d124c3d58f00016c3376a1de5323cff" \
> 	cluster-infrastructure="openais" \
> 	expected-quorum-votes="2" \
> 	stonith-enabled="false" \
> 	last-lrm-refresh="1364509824"
> 
> I have followed the pacemaker doc(http://clusterlabs.org/doc/en-US/Pacemaker/1.1-crmsh/pdf/Clusters_from_Scratch/Pacemaker-1.1-Clusters_from_Scratch-en-US.pdf) but I haven't used the distributed file system because, in this case, nginx is not sharing the configuration files and is playing like a load balancer.
> Can you tell me how to "link" the resources ?
> 
> Thank you
> _______________________________________________
> 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