[Pacemaker] Question Regarding Ping Node - Pingd

Andrew Beekhof andrew at beekhof.net
Mon Jul 26 12:24:31 UTC 2010


On Fri, Jul 16, 2010 at 11:40 AM, Rainer Lutz <rainer.lutz at realtech.com> wrote:
> Dear all,
>
> i have a question regarding the ping nodes and the resource ocf:pacemaker:pingd
> or ocf:pacemaker:ping
>
> I configured the resource ping on one of my two nodes. Pointing the resource (
> to a virtual ip adress i configured with hb_gui. XML:
>
> <primitive class="ocf" id="Ressource_ping" provider="pacemaker" type="ping">
>  <meta_attributes id="Ressource_ping-meta_attributes">
>  <nvpair id="Ressource_ping-meta_attributes-target-role"  name="target-role"
> value="Started"/>
>  </meta_attributes>
>  <operations id="Ressource_ping-operations">
>  <op id="Ressource_ping-op-monitor-10" interval="10" name="monitor" timeout="60"/>
>   </operations>
>   <instance_attributes id="Ressource_ping-instance_attributes">
>   <nvpair id="Ressource_ping-instance_attributes-host_list" name="host_list"
> value="10.0.2.132"/>
>   </instance_attributes>
> </primitive>
>
> Now if the stop the virtual IP Adress with the hb_gui the resource still shows
> "green lights"
>
> Am i missing here something?

The resource doesn't fail just because the host(s) are no longer reachable.
Its job is to maintain various host attributes.

Try this when the IP is active:
    cibadmin -Ql | grep nvpair.*ping

Now deactivate the IP address and compare the output.




More information about the Pacemaker mailing list