[ClusterLabs] <Outage Resend> [Q] Cluster failovers too soon

Sebish sebish at sebish.de
Mon Jan 4 09:57:34 EST 2016

< I am resending this mail, because of the outage of clusterlabs during 
the weekend, a received error message and my timelimit until next week>
<Please excuse that this mail will be kept two times in the january archive>

Hello guys,

happy new year to all of you!

I have a little (/understanding?/) problem regarding Heartbeat/Pacemaker 
and deadtime/timeout.
I know that corosync is the way the go, but atm I have a heartbeat 
cluster and need to adjust it's time before a failover is initiated.

My cluster and resources completely ignore the heartbeat deadtime raise 
and the timeout in pacemaker resource agents definitions.
When I shut him off, the node gets shown as offline and the services are 
failovered after 4-9 seconds. But I want 20 seconds.

What do I have to adjust, to make the cluster failover after +- 20 
seconds instead of 9? Do I miss a parameter apart from 
deadtime(deadping) and timeout?
Every hint would be a great help!

Thank you very much



keepalive 2
warntime 6
deadtime 20
initdead 60

*_crm (pacemaker)_:*

node $id="6acc2585-b49b-4b0f-8b2a-8561cceb8b83" nodec
node $id="891a8209-5e1a-40b6-8d72-8458a851bb9a" kamailioopenhab2
node $id="fd898711-4c76-4d00-941c-4528e174533c" kamailioopenhab1
primitive ClusterMon ocf:pacemaker:ClusterMon \
         params user="root" update="30" extra_options="-E 
/usr/lib/ocf/resource.d/*myname*/*script*.sh" \
         op monitor interval="10" timeout="40" on-fail="restart"
primitive FailoverIP ocf:heartbeat:IPaddr2 \
         params ip="*ClusterIP*" cidr_netmask="18" \
         op monitor interval="2s" timeout="20"
primitive Openhab lsb:openhab \
         meta target-role="Started" \
         op monitor interval="2s" timeout="20"
primitive Ping ocf:pacemaker:ping \
         params host_list="*ClusterIP*" multiplier="100" \
         op monitor interval="2s" timeout="20"
location ClusterMon_LocationA ClusterMon -inf: kamailioopenhab1
location ClusterMon_LocationB ClusterMon 10: kamailioopenhab2
location ClusterMon_LocationC ClusterMon inf: nodec
location FailoverIP_LocationA FailoverIP 20: kamailioopenhab1
location FailoverIP_LocationB FailoverIP 10: kamailioopenhab2
location FailoverIP_LocationC FailoverIP -inf: nodec
colocation Services_Colocation inf: FailoverIP Kamailio Openhab
property $id="cib-bootstrap-options" \
dc-version="1.1.7-ee0730e13d124c3d58f00016c3376a1de5323cff" \
         cluster-infrastructure="Heartbeat" \
         expected-quorum-votes="2" \
         last-lrm-refresh="1451669632" \
         stonith-enabled="false" \
rsc_defaults $id="rsc-options" \

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20160104/8643dc63/attachment-0003.html>

More information about the Users mailing list