[ClusterLabs] Issues with crm_mon or ClusterMon resource agent

[Ken Gaillot]

On 03/06/2016 08:36 AM, Debabrata Pani wrote:
> Hi,
> 
> I would like to understand if anybody has got this working recently.
> 
> Looks like I have missed something in the description and hence the
> problem statement is not clear to the group.
> 
> Can I enable some logs in crm_mon to improve the description of the
> problem ?
> 
> Regards,
> Debabrata Pani
> 
> 
> 
> On 04/03/16 11:48, "Debabrata Pani" <Debabrata.Pani at mobileum.com> wrote:
> 
>> Hi,
>>
>> I wanted  to configure ClusterMon resource agent so that I can get
>> information about events in the pacemaker cluster.
>> *Objective is to generate traps for some specific resource agents and/or
>> conditions*
>>
>>
>> My cluster installation details :
>> pacemakerd --version
>> Pacemaker 1.1.11
>> Written by Andrew Beekhof
>>
>>
>> corosync -v
>> Corosync Cluster Engine, version '1.4.7'
>> Copyright (c) 2006-2009 Red Hat, Inc.
>>
>>
>> crm_mon --version
>> Pacemaker 1.1.11
>> Written by Andrew Beekhof
>>
>>
>>
>> I followed the following documentation
>> http://clusterlabs.org/doc/en-US/Pacemaker/1.1-crmsh/html/Pacemaker_Explai
>> n
>> ed/ch07.html
>> http://floriancrouzat.net/2013/01/monitor-a-pacemaker-cluster-with-ocfpace
>> m
>> akerclustermon-andor-external-agent/
>>
>> My current cluster status shows:
>>
>> MonitorClusterChange    (ocf::pacemaker:ClusterMon):    Started stvm4
>>
>> The resource agent is configure
>>
>> The crm_mon is indeed running as a daemon on the node stvm4
>>
>> [root at stvm4 panidata]# ps -ef | grep crm_mon | grep -v grep
>> root      2908     1  0 10:59 ?        00:00:00 /usr/sbin/crm_mon -p
>> /tmp/ClusterMon_MonitorClusterChange.pid -d -i 0 -E
>> /root/panidata/testscript.sh -e anonymous -h
>> /tmp/ClusterMon_MonitorClusterChange.html

The "-i 0" refers to the "update" property of your resource, which is
how often crm_mon should recheck the cluster. I'm not sure what zero
would do, but it would be better around 10-30 seconds.

Is /root/panidata/testscript.sh executable? Does it work when run from
the command line?

>> My test script is the following
>> #!/bin/bash
>> echo "running" >> /root/running.log
>> echo "CRM_notify_recipient=$CRM_notify_recipient"
>> ..
>>
>>
>>
>> As I trigger events by shutting down one or the other service, I see the
>> html file "/tmp/ClusterMon_MonitorClusterChange.html² getting updated each
>> time an event is triggered.
>> So the timestamp of the file keeps on changing.
>>
>> But I am not sure if the script is getting executed. Because I don¹t see
>> any ³/root/running.log² file.
>>
>> Things I have tried:
>> * Using ³logger² command instead of echo.
>> * Running the crm_mon command with -d and other parameters manually to
>> check , if it is the problem with resource agent etc.
>>
>> Queries:
>> * Is this a know issue ?
>> * Am I doing something incorrect ?
>>
>>
>>
>>
>> Regards,
>> Debabrata Pani