[Linux-cluster] fence_cisco_ucs working from command line.but not from cluster.conf

[Digimer]

What is the exit code from the manual fence call?

/usr/sbin/fence_cisco_ucs -a 172.22.90.10 -l admin -p duc2Cisco -o 
status -v -n node1 -z; echo $?

The exit codes must follow the API rules. Read; 
https://fedorahosted.org/cluster/wiki/FenceAgentAPI#agent_ops

digimer

On 09/24/2012 01:29 AM, Ben .T.George wrote:
> Hi
>
> from command line i issued :
>
> /usr/sbin/fence_cisco_ucs -a 172.22.90.10 -l admin -p duc2Cisco -o
> status -v -n node1 -z
>
> is working and i got the output, even i give reboot also working
>
> but i configured in the cluster.conf file with same configuration, like:
>
> <clusternode name="node1.test.net <http://node1.test.net>" nodeid="1">
> <fence>
> <method name="ucs-node1">
> <device name="ucs-node1" port="node1"/>
> </method>
> </fence>
> </clusternode>
>
>
> <fencedevices>
> <fencedevice agent="fence_cisco_ucs" ipaddr="172.22.90.10" login="admin"
> name="ucs-node1" passwd="duc2Cisco"/>
> </fencedevices>
>
>
> but to test fencing property.i turned off heartbeat network on node
> node1. but i am getting error on /var/log/messages.
>
>
> Sep 24 07:20:13 node1 fenced[6589]: fence node2.test.net
> <http://node2.test.net> dev 0.0 agent fence_cisco_ucs result: error from
> agent Sep 24 07:20:13 node1 fenced[6589]: fence node2 . test .net failed
> Sep 24 07:20:16 node1 fenced[6589]: fencing node node2 . test .net Sep
> 24 07:20:17 node1 fenced[6589]: fence node2 . test .net dev 0.0 agent
> fence_cisco_ucs result: error from agent Sep 24 07:20:17 node1
> fenced[6589]: fence node2 . test .net failed Sep 24 07:20:20 node1
> fenced[6589]: fencing node node2 . test .net Sep 24 07:20:20 node1
> fenced[6589]: fence node2 . test .net dev 0.0 agent fence_cisco_ucs
> result: error from agent Sep 24 07:20:20 node1 fenced[6589]: fence node2
> . test .net failed
>
> please help me to solve this
>
> regards,
> Ben
>
>


-- 
Digimer
Papers and Projects: https://alteeve.ca