drac5 fencing issue for Linux 6.7

[chindi]

We are not able to complete our regression test for our 2 node cluster

we are powering off our one node directly , after which the resources are not moving to second node.

In /var/log/cluster/fenced.log we can see below messages :

node fencing failed : error from agent

We are using fence_dra5 n our cluster.conf file.

[chindi]

<fencedevices>
                <fencedevice agent="fence_drac5" cmd_prompt="/admin1" ipaddr="10.1.4.146" login="fence" name="Dell_Drac" passwd="P@ssw0rd" retry_on="10" secure="on"/>
                <fencedevice agent="fence_drac5" cmd_prompt="/admin1" ipaddr="10.1.4.145" login="fence" name="Dell_Drac1" passwd="P@ssw0rd" retry_on="10" secure="on"/>
                <fencedevice agent="fence_idrac" auth="password" ipaddr="10.1.4.146" lanplus="on" login="fence" name="test1" passwd="cedge123" privlvl="ADMINISTRATOR"/>
        </fencedevices>

 Our fence devices entry is as above.

[Arunabha Banerjee]

As I know drac5 fence agent is being used for Dell server. If you have HP server then please use fence_ipmilan agent for HP iLO3/4.

[chindi]

We have dell servers R530 , but fencing command from ssh works , but not through cluster.conf 😔

[Arunabha Banerjee]

Please refer Red Hat Solution: https://access.redhat.com/solutions/18915

[chindi]

We have drac8

[Arunabha Banerjee]

Then you should configure your fencedevice agent="fence_ipmilan". You should not use drac5.

[chindi]

We were able to use drac5