How to force cluster failover

[wakathi]

Hello,

I would like to know what kind of response is expected from a SCSI target that would force the cluster to failover to another node in the cluster ?

Eg:

node1 node2
| /
| /
-----
| lun |
-----

node1 is the current owner and application that is accessing the lun is running on node1.

What kind of SCSI response (Check Condition/Sense codes or reservation conflict etc) is required to be returned by the lun to the I/Os from node1, so that MSCS forces a failover/failback from node1 to node2 and node2 now becomes the current owner ?

Thanks!