Hello, again guys. I`m trying to setup MDT fail over with drbd and pacemaker/corosync. I have OST fail over similar to that how-to http://wiki.lustre.org/index.php/Using_Pacemaker_with_Lustre. It works really nice, whenever OSS becomes unavailable all of its OSTs are mounted on the fail over OSS. I`m trying to accomplish same with drbd and pacemaker/corosync for the MDS/MDT. I have two servers each with dedicated drbd device. Does anyone can point me to the right direction for drbd and pacemaker configurations. I think that both drbd devices should be in primary/primary mode. So that when one is down, and the other becomes active MDT, the sync should also be started from the newly active device to the previously active. When using primary/primary configurations split brain can occur, there are some handlers which can be defined on how to proceed when split brain is detected ... I guess I should look more carefully into them. The other thing is how to define the resource in pacemaker. I see that there is a resource agent for drbd : ocf:linbit:drbd. So I guess it should be responsible for checking the drbd device availability and then another ocf:heartbeat:Filesystem resource for the actual mount .... here is where I get lost. Cheers, _______________________________________________ Lustre-discuss mailing list Lustre-discuss-aLEFhgZF4x6X6Mz3xDxJMA@public.gmane.org http://lists.lustre.org/mailman/listinfo/lustre-discuss