I'm working on a failover setup (two DSS) and I was trying to setup the worst case scenario, aka:
First server is done for. You have to set up another DSS from scratch to act as a new primary node.
How does it work with the failback?
Do I just reconfigure the iscsi failover on this new server, set it as primary, plug it to the network, and hit "sync volumes" on the secondary node?
In the case of primary server completely died, then you need to stop the replication task from the secondary. Bring the old primary up, and create the failover task again and make it as secondary. When the 2 servers sync again you can modify the failover settings and switch the primary to be secondary and the secondary (old primary) to be primary.
your to fast i had this already typed up - i will post any way.
That is a very good question and we should post this on our kb.open-e.com
Ok, here is what will happen if your Primary "DIED" and you will place a new server with new hardware (just for the worst case).
Stop the replication from the Secondary/active server then from the Volume Manager / Vol. replication -> Function: Volume replication modes stop the volumes being replicated then set the iSCSI Logical Volumes to Source mode. Then set this as Primary / Active in the iSCSI Failover section of the GUI.
Then rebuild your new server setting the iSCSI Logical Volumes to be in destination mode, let the iSCSI Logical Volumes complete their replication process and make this your secondary / passive server in the iSCSI Failover section.
There's no ongoing replication process going on, as the failover kicked in and the secondary now appears as "Secondary/Active"
The volumes appear as "source" in the replication tab.
There are replication tasks being "replicatevolume_reverse" which led me to think that it "might" (emphasis on might) be possible to set up a new primary with the correct options and proceed to a syncback, then failback to said server.
Maybe there is not actual replication running but there is a replication task that is still active. And I recommend to stop that task again then create it again with the new LV.
Sorry if I sound slow to catch up, as it is definitely not my goal, but there is no running tasks whatsoever.
Unless you mean I should stop the failover service, delete the existing replication tasks, and set up some new replication tasks from secondary to primary?