On 10/11/2012 12:43 AM, Kevin F. La Barre wrote: > I'm testing stickiness in a sandbox that consists of 3 nodes. The > configuration is very simple but it's not acting the way I think it should. > > > My configuration: > > # crm configure show > node hasb1 > node hasb2 > node hasb3 > primitive postfix lsb:postfix \ > op monitor interval="15s" > property $id="cib-bootstrap-options" \ > dc-version="1.1.7-6.el6-148fccfd5985c5590cc601123c6c16e966b85d14" \ > cluster-infrastructure="openais" \ > expected-quorum-votes="3" \ > no-quorum-policy="ignore" \ > stonith-enabled="false" \ > last-lrm-refresh="1349902760" \ > maintenance-mode="false" \ > is-managed-default="true" > rsc_defaults $id="rsc-options" \ > resource-stickiness="100" > > > > The test resource "postfix" lives on hasb1. > > # crm_simulate -sL > > Current cluster status: > Online: [ hasb1 hasb3 hasb2 ] > > postfix (lsb:postfix): Started hasb1 > > Allocation scores: > native_color: postfix allocation score on hasb1: 100 > native_color: postfix allocation score on hasb2: 0 > native_color: postfix allocation score on hasb3: 0 > > > On hasb1 I'll kill the corosync process. Resource moves over to hasb2 as > expected.
So cluster processes are killed and the resource keeps running on hasb1 ... and starts a second time on hasb2 as hasb1 is still running and you have no stonith ... > > # crm status > ============ > Last updated: Wed Oct 10 22:35:23 2012 > Last change: Wed Oct 10 21:30:12 2012 via crm_resource on hasb2 > Stack: openais > Current DC: hasb2 - partition with quorum > Version: 1.1.7-6.el6-148fccfd5985c5590cc601123c6c16e966b85d14 > 3 Nodes configured, 3 expected votes > 1 Resources configured. > ============ > > Online: [ hasb3 hasb2 ] > OFFLINE: [ hasb1 ] > > postfix (lsb:postfix): Started hasb2 > > > # crm_simulate -sL > > Current cluster status: > Online: [ hasb3 hasb2 ] > OFFLINE: [ hasb1 ] > > postfix (lsb:postfix): Started hasb2 > > Allocation scores: > native_color: postfix allocation score on hasb1: 0 > native_color: postfix allocation score on hasb2: 100 > native_color: postfix allocation score on hasb3: 0 > > > Now I'll start corosync & pacemaker. Postfix resource moves back to hasb1 > even though we have default stickiness. in your logs you will see the cluster detecting postfix running twice and do a stop all/start one by default ... really stop/reset a server if you want to test node failures, like: echo b >/proc/sysrq-trigger ... and use stonith! Regards, Andreas -- Need help with Pacemaker? http://www.hastexo.com/now > > # crm status > ============ > Last updated: Wed Oct 10 22:37:00 2012 > Last change: Wed Oct 10 21:30:12 2012 via crm_resource on hasb2 > Stack: openais > Current DC: hasb2 - partition with quorum > Version: 1.1.7-6.el6-148fccfd5985c5590cc601123c6c16e966b85d14 > 3 Nodes configured, 3 expected votes > 1 Resources configured. > ============ > > Online: [ hasb1 hasb3 hasb2 ] > > postfix (lsb:postfix): Started hasb1 > > > # crm_simulate -sL > > Current cluster status: > Online: [ hasb1 hasb3 hasb2 ] > > postfix (lsb:postfix): Started hasb1 > > Allocation scores: > native_color: postfix allocation score on hasb1: 100 > native_color: postfix allocation score on hasb2: 0 > native_color: postfix allocation score on hasb3: 0 > > > What am I missing? I'm pulling my hair - any help would be appreciated > greatly. > > Corosync 1.4.1 > Pacemaker 1.1.7 > CentOS 6.2 > > > -Kevin > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems >
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
