Trying this again with a few updated details and a slightly different problem...
I have a 2 node cluster (xen1 and xen2) running on dl380 G3’s. There are 2 virtual machines running on these. From xen1 from the shell command line I can fence xen2 through the hp lights out port using /sbin/fence_ilo –a 192.168.1.53 –l aking –p password –v, or using the comoonics-bootimage-fenceclient-ilo-0.1-16.noarch.rpm. I can also do the same from the second node and successfully fence the first node. When I fence the node that has control of one or more virtual machines they are migrated to the other node.
If I fence a node in Luci this works too.
I have xen1-ilo and xen2-ilo set in /etc/hosts with their respective ip addresses
When I run an /sbin/reboot –f on xen2, xen1 successfully fences that node and if necessary takes control of any virtual machines that are on it.
When I run an /sbin/reboot –f on xen1, xen2 does not even try and fence node1 and just waits for it to come back up and rejoin the cluster. /sbin/fenced is running on both machines, I cannot see any difference in the running services between them.
The problem here is that if there is that problem with xen1, the virtual machines will be down. This kinda ruins the idea of a cluster for me.
Can anyone advise why this is happening?
contents of this message may be privileged and confidential. If you have
received this message in error, you may not use,
adam king intechnology com
Support 0845 120 7070
Telephone 01423 850000
Facsimile 01423 858866
This is an email from InTechnology plc, Central House, Beckwith Knowle, Harrogate, UK, HG3 1UG.