[Pacemaker] DRBD resources show as running when they don't exist

[Stephen Marsh]

Hi all,

I'm using Corosync 2.3.1 with Pacemaker 1.1.10 (final release) and DRBD 
8.4.3.

I've got a strange problem with the way Pacemaker handles DRBD resources 
that don't exist.

I'm testing with this config:

primitive vs1-1-DRBD ocf:linbit:drbd \
         params drbd_resource="vs1-1" \
         operations $id="vs1-1-DRBD-ops" \
         op start interval="0" timeout="60s" on-fail="block" \
         op stop interval="0" timeout="60s" on-fail="block" \
         op monitor interval="20" role="Master" timeout="50" \
         op monitor interval="30" role="Slave" timeout="50"

ms vs1-1-MS vs1-1-DRBD \
         meta resource-stickiness="100" master-max="2" notify="true" 
interleave="true" target-role="Started" is-managed="true"

DRBD resource vs1-1 does not exist at all, and "crm_mon -1" correctly 
shows that it has failed... but when I run "crm resource show vs1-1-MS" 
(or "crm resource show vs1-1-DRBD"), it reports that it's running on all 
members of the cluster.

If, on the other hand, I create a DRBD config file for vs1-1 but specify 
a non-existent backing device so that it still can't start, but for a 
different reason, Pacemaker behaves as expected: it registers the 
resource as failed and doesn't show the resource as running at all.

Any help would be most appreciated.

Thanks

Regards,
Stephen Marsh