[Pacemaker] How to make the 1st resources to failover is 2nd resources is being killed?

[Andrew Beekhof]

On Fri, Mar 15, 2013 at 2:21 PM, Paul Sun <paul.sun at syniverse.com> wrote:
> I noticed that I need to configure the "failover-timeout" together with "migration-threshold" in order to achieve the expected failover behavior, is this the correct implementation?

Yes.

> Can advice?
>
>
>     <nodes>
>       <node id="HA_NODE1" uname="HA_NODE1"/>
>       <node id="HA_NODE2" uname="HA_NODE2"/>
>       <node id="HA_NODE3" uname="HA_NODE3"/>
>     </nodes>
>     <resources>
>       <primitive class="lsb" id="PDNS-Recursor" type="pdns_recursord">
>         <meta_attributes id="PDNS-Recursor-meta_attributes">
>           <nvpair id="PDNS-Recursor-meta_attributes-migration-threshold" name="migration-threshold" value="1"/>
>           <nvpair id="PDNS-Recursor-meta_attributes-failure-timeout" name="failure-timeout" value="30"/>
>         </meta_attributes>
>         <operations>
>           <op id="PDNS-Recursor-monitor-30s" interval="30s" name="monitor" on-fail="restart"/>
>         </operations>
>       </primitive>
>       <primitive class="ocf" id="FloatingIP_02" provider="heartbeat" type="IPaddr2">
>         <instance_attributes id="FloatingIP_02-instance_attributes">
>           <nvpair id="FloatingIP_02-instance_attributes-ip" name="ip" value="192.168.206.160"/>
>           <nvpair id="FloatingIP_02-instance_attributes-cidr_netmask" name="cidr_netmask" value="32"/>
>           <nvpair id="FloatingIP_02-instance_attributes-nic" name="nic" value="eth2"/>
>           <nvpair id="FloatingIP_02-instance_attributes-iflabel" name="iflabel" value="vip2"/>
>         </instance_attributes>
>         <operations>
>           <op id="FloatingIP_02-monitor-30s" interval="30s" name="monitor"/>
>         </operations>
>         <meta_attributes id="FloatingIP_02-meta_attributes">
>           <nvpair id="FloatingIP_02-meta_attributes-resource-stickiness" name="resource-stickiness" value="1000"/>
>         </meta_attributes>
>       </primitive>
>     <constraints>
>       <rsc_colocation id="PDNS-Recursor-Services" rsc="PDNS-Recursor" score="INFINITY" with-rsc="FloatingIP_02"/>
>       <rsc_location id="PDNS-Recursor-prefer1" node="HA_NODE1" rsc="PDNS-Recursor" score="-INFINITY"/>
>       <rsc_location id="PDNS-Recursor-prefer2" node="HA_NODE2" rsc="PDNS-Recursor" score="400"/>
>       <rsc_location id="PDNS-Recursor-prefer3" node="HA_NODE3" rsc="PDNS-Recursor" score="300"/>
>       <rsc_order first="FloatingIP_02" id="PDNS-Recursor-Order" score="INFINITY" then="PDNS-Recursor"/>
>       <rsc_location id="PDNS-IP-prefer1" node="HA_NODE1" rsc="FloatingIP_02" score="-INFINITY"/>
>     </constraints>
>
>
>
> -----Original Message-----
> From: Paul Sun
> Sent: Friday, March 15, 2013 10:14 AM
> To: 'The Pacemaker cluster resource manager'
> Subject: RE: How to make the 1st resources to failover is 2nd resources is being killed?
>
> Thanks for the information, however, it only works for 1 times. The behavior is exactly same as the previous posts " http://www.mail-archive.com/pacemaker@oss.clusterlabs.org/msg00415.html"
>
> When I killed the application for the 1st time, it will failover to other node, but when I terminate the application in the new node, it will never failover back, and the status is "stopped".
>
> Any idea?
>
>
>> Anyone can help?
>
> Set migration-threshold.
>
> And: this is a public community mailing list. Pinging after a few hours is not good style. If you need help that urgently, some people sell consulting for this.
>
>
> Regards,
>     Lars
>
> --
> Architect Storage/HA
> SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 21284 (AG Nürnberg) "Experience is the name everyone gives to their mistakes." -- Oscar Wilde
>
> -----Original Message-----
> From: Paul Sun
> Sent: Thursday, March 14, 2013 6:20 PM
> To: 'The Pacemaker cluster resource manager'
> Subject: RE: How to make the 1st resources to failover is 2nd resources is being killed?
>
> Anyone can help?
>
> -----Original Message-----
> From: Paul Sun
> Sent: Thursday, March 14, 2013 4:16 PM
> To: The Pacemaker cluster resource manager
> Subject: How to make the 1st resources to failover is 2nd resources is being killed?
>
> Hi
>
> I have 3 node setup, the FloatingIP_02 and PDNS-Recursor serve for the same purpose. The PDNS-Recursor can only be started if FloatingIP_02 is presented on the same node. I would like to configure such that if I terminate PDNS-Recursor, the FloatingIP_02 together with PDNS-Recursor will be failed over to node2.
>
> Below is my configuration.
>
> Thanks
>
> - Paul
>
> node HA_NODE1
> node HA_NODE2
> node HA_NODE3
> primitive FloatingIP_02 ocf:heartbeat:IPaddr2 \
>         params ip="192.168.206.160" cidr_netmask="32" nic="eth2" iflabel="vip2" \
>         op monitor interval="30s" \
>         meta resource-stickiness="1000"
> primitive PDNS-Recursor lsb:pdns_recursord \
>         op monitor interval="30s"
> location PDNS-IP-prefer1 FloatingIP_02 -inf: HA_NODE1 location PDNS-IP-prefer2 FloatingIP_02 400: HA_NODE2 location PDNS-IP-prefer3 FloatingIP_02 300: HA_NODE3 location PDNS-Recursor-prefer1 PDNS-Recursor -inf: HA_NODE1 colocation PDNS-Recursor-Services inf: PDNS-Recursor FloatingIP_02 order PDNS-Recursor-Order inf: FloatingIP_02 PDNS-Recursor property $id="cib-bootstrap-options" \
>         dc-version="1.1.8-7.el6-394e906" \
>         cluster-infrastructure="classic openais (with plugin)" \
>         expected-quorum-votes="3" \
>         stonith-enabled="true" \
>         no-quorum-policy="ignore" \
>         last-lrm-refresh="1363236700"
>
> _______________________________________________
> 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