On Thu, Jun 24, 2010 at 2:35 PM, Alain.Moulle <[email protected]> wrote:
> Hi again,
>
> no way to make fence working on RHEL6 :
>
> I have 3 nodes in cluster and the status before the test is (crm_mon) :
>
> Online: [ node1 node2 node3 ]
>
> restofencenode1        (stonith:fence_ipmilan):        Started node2
> restofencenode2        (stonith:fence_ipmilan):        Started node3
> restofencenode3        (stonith:fence_ipmilan):        Started node1
>
> But when I do force an error on node2 (if down on heartbeat IF) so that
> node3 resource restofencenode3 fences the node2 with fence_ipmilan, it
> does not make the job, I can see that node2 is UNCLEAN (due to no more
> heartbeat)
> Node node2: UNCLEAN (offline)
> Online: [ node1 node3 ]
>
> restofencenode1        (stonith:fence_ipmilan) Started [    node2  node3 ]
> restofencenode2        (stonith:fence_ipmilan):        Started node3
> restofencenode3        (stonith:fence_ipmilan):        Started node1
>
> but it is very strange that restofencenode1 appears started on node2 and
> node3 whereas it should only be started
> on node 3, due to the constraints I set :
>      <rsc_location id="loc1-restofencenode1" node="node2"
> rsc="restofencenode1" score="+INFINITY"/>
>      <rsc_location id="forbiddenloc-restofencenode1" node="node1"
> rsc="restofencenode1" score="-INFINITY"/>
>      <rsc_location id="loc1-restofencenode2" node="node3"
> rsc="restofencenode2" score="+INFINITY"/>
>      <rsc_location id="forbiddenloc-restofencenode2" node="node2"
> rsc="restofencenode2" score="-INFINITY"/>
>      <rsc_location id="loc1-restofencenode3" node="node1"
> rsc="restofencenode3" score="+INFINITY"/>
>      <rsc_location id="forbiddenloc-restofencenode3" node="node3"
> rsc="restofencenode3" score="-INFINITY"/>
> but it seems that it is because of problems on the use of fence_ipmilan
> that crm_mon displays something not really right ...
>
> but I can also see in syslog :
> node3 daemon info stonith-ng [807]: info: can_fence_host_with_device:
> *restofencenode2 can not fence node2: dynamic-list

What happens if you run fence_ipmilan manually (with the necessary
authentication options) and -o list
Does it show node names?

> *
> I have quite the same configuration as before with external/ipmi on
> fc12, except all around the fence_ipmilan parameters which differs,
> and it worked fine on fc12, did someone here has already tried the
> pacemaker fencing with fence_ipmilan agent ?
>
> My restofencenode2 record in cib :
>
>      <primitive class="stonith" id="restofencenode2" type="fence_ipmilan">
>        <instance_attributes id="restofencenode2-instance_attributes">
>          <nvpair id="restofencenode2-instance_attributes-ipaddr"
> name="ipaddr" value="BMC address of node2"/>
>          <nvpair id="restofencenode2-instance_attributes-login"
> name="login" value="mylogin"/>
>          <nvpair id="restofencenode2-instance_attributes-passwd"
> name="passwd" value="mypasswd"/>
>          <nvpair id="restofencenode2-instance_attributes-action"
> name="action" value="reboot"/>
>        </instance_attributes>
>        <meta_attributes id="restofencenode2-meta_attributes">
>          <nvpair id="restofencenode2-meta_attributes-target-role"
> name="target-role" value="Started"/>
>        </meta_attributes>
>      </primitive>
>
> If someone has an idea ...
> Thanks a lot
> Regards
> Alain
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to