[Pacemaker] Cleanup over secondary node

Andrew Beekhof andrew at beekhof.net
Wed Apr 17 03:27:39 UTC 2013


On 17/04/2013, at 11:28 AM, Daniel Bareiro <daniel-listas at gmx.net> wrote:

> Ho Andrew.
> 
> On Monday, 15 April 2013 14:36:48 +1000,
> Andrew Beekhof wrote:
> 
>>> I'm testing Pacemaker+Corosync cluster with KVM virtual machines. When
>>> restarting a node, I got the following status:
>>> 
>>> # crm status
>>> ============
>>> Last updated: Sun Apr 14 11:50:00 2013
>>> Last change: Sun Apr 14 11:49:54 2013
>>> Stack: openais
>>> Current DC: daedalus - partition with quorum
>>> Version: 1.1.7-ee0730e13d124c3d58f00016c3376a1de5323cff
>>> 2 Nodes configured, 2 expected votes
>>> 8 Resources configured.
>>> ============
>>> 
>>> Online: [ atlantis daedalus ]
>>> 
>>> Resource Group: servicios
>>>    fs_drbd_servicios  (ocf::heartbeat:Filesystem):    Started daedalus
>>>    clusterIP  (ocf::heartbeat:IPaddr2):       Started daedalus
>>>    Mysql      (ocf::heartbeat:mysql): Started daedalus
>>>    Apache     (ocf::heartbeat:apache):        Started daedalus
>>>    Pure-FTPd  (ocf::heartbeat:Pure-FTPd):     Started daedalus
>>>    Asterisk   (ocf::heartbeat:asterisk):      Started daedalus
>>> Master/Slave Set: drbd_serviciosClone [drbd_servicios]
>>>    Masters: [ daedalus ]
>>>    Slaves: [ atlantis ]
>>> 
>>> Failed actions:
>>>   Asterisk_monitor_0 (node=atlantis, call=12, rc=5, status=complete): not installed
>>> 
>>> 
>>> The problem is that if I do a "cleanup" of the Asterisk resource in the
>>> secondary, this has no effect. It seems to be Paceemaker needs to have
>>> access to the config file to the resource.
> 
>> Not Pacemaker, the resource agent.
>> Pacemaker runs a non-recurring monitor operation to see what state the
>> service is in, it seems the asterisk agent needs that config file.
>> 
>> I'd suggest changing the agent so that if the asterisk process is not
>> running, the agent returns 7 (not running) before trying to access the
>> config file.
> 
> I was reviewing the resource definition assuming there I might have made
> some reference to the Asterisk configuration file, but this was not the
> case:
> 
> primitive Asterisk ocf:heartbeat:asterisk \
>        params realtime="true" \
>        op monitor interval="60s" \
>        meta target-role="Started"
> 
> This agent is the one that is available in the resource-agents package
> from Debian Backports repository:
> 
> atlantis:~# aptitude show resource-agents
> Paquete: resource-agents
> Nuevo: sí
> Estado: instalado
> Instalado automáticamente: sí
> Versión: 1:3.9.2-5~bpo60+1
> Prioridad: opcional
> Sección: admin
> Desarrollador: Debian HA Maintainers <debian-ha-maintainers at lists.alioth.debian.org>
> Tamaño sin comprimir: 2.228 k
> Depende de: libc6 (>= 2.4), libglib2.0-0 (>= 2.12.0), libnet1 (>= 1.1.2.1), libplumb2, libplumbgpl2, cluster-glue, python
> Tiene conflictos con: cluster-agents (<= 1:1.0.4-1), rgmanager (<= 3.0.12-2+b1)
> Reemplaza: cluster-agents (<= 1:1.0.4-1), rgmanager (<= 3.0.12-2+b1)
> Descripción: Cluster Resource Agents
> The Cluster Resource Agents are a set of scripts to interface with several services to operate in a High Availability environment for both Pacemaker and
> rgmanager resource managers.
> Página principal: https://github.com/ClusterLabs/resource-agents
> 
> 
> 
> 
> Do you know if there is any way to get the behavior that you suggested
> me using this agent?

You'll have to edit it and submit the changes upstream.
If whatever it is looking for is not found when a monitor is requested, it should probably return 7 (STOPPED)

> 
> 
> Thanks for your reply.
> 
> 
> Regards,
> Daniel
> -- 
> Ing. Daniel Bareiro - GNU/Linux registered user #188.598
> Proudly running Debian GNU/Linux with uptime:
> 21:54:06 up 52 days,  6:01, 11 users,  load average: 0.00, 0.02, 0.00
> _______________________________________________
> 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





More information about the Pacemaker mailing list