We are setting up iSCSI Failover on two DSS V6 servers (colocated). If the power goes out for more than a few minutes (enough to drain the UPSs), both nodes will shut down. In testing, I have seen that the iSCSI Failover is not running when the servers come back up. Initiators cannot connect through the virtual IP. I asked Todd at Open-E about this, and he explained that there is a reason for that, you have to get volume replication back to a consistent state before failover can be re-enabled. That makes sense, but it defeats the purpose of 'highly-available' if an administrator has to re-enable failover before clients (initiators) can connect.

Is there a way to automatically provide access to the primary node in this situation? That way the primary node could still be available even before replication and failover are reestablished by the administrator. Ideally the primary node would still be available through the virtual IP, or if that won't work, maybe there is a way to configure initiators to point to the primary node's physical IP as a backup?

Any ideas would be appreciated.

Thanks,
Dave