[ClusterLabs] Fence agent ends up stopped with no clear reason why

Casey Allen Shobe caseyandgina at icloud.com
Thu Aug 2 20:23:45 EDT 2018


The fence device starts fine, and randomly fails some time later.  In the first message I sent, you can see that the failure message had a different date/time on each cluster node, but ultimately it failed on all nodes.  My second E-mail on this thread has the log data from one node attached...

> On Aug 1, 2018, at 3:03 PM, Ken Gaillot <kgaillot at redhat.com> wrote:
> 
> For fence devices, a start first registers the device with stonithd
> (which should never fail). There should be a log message from stonithd
> like "Added 'vmware_fence' to the device list". The cluster then does
> an initial monitor. That is most likely what failed.
> 
> If you're lucky, the fence agent logged some detail about why that
> monitor failed, or has a debug option to do so.



More information about the Users mailing list