With iSCSI failover running just fine, I cannot find a way to go back and add more LVs/LUNS without having to turn everything off, impacting the Xen nodes, adding the new LVs/LUNS, and then re-enabling failover?
I can get to the point of creating the LVs/LUNS, but you cannot add the task to replicate to the "Failover Tasks" section, the apply is greyed out.
Am I missing something? Do properly implement failover you need to have all your LVs/LUNS defined up front or else shut everything down, add new ones, bring it all back up?
You are correct that this is the proper procedure that you will need to stop the replication then add the new replicated volumes in to the failover task.
This is also similar when increasing the iSCSI Logical volumes that use the manual replication where you have to stop the replication and then expand the logical volume then re-enable the replication task for both the source and destination.
In future releases we will look into this to make it seamless.
Just so I am 100% clear, stopping the failover services also stops the VIP and means everything has to be taken down (complete outage) while doing this, or am I missing something else?
A complete outage of both the primary and redundant "SAN", right?
... and this is IMHO not good for a fail over cluster! If I'm running a clustered HA SAN, I want to do everything (! LV expansion, RAID changes, adding LUNs ...) WITHOUT stopping the depending services (VMWare, XEN, Hyper-V, ...). This is the reason why I want to use 2 replicated/mirrored machines ...
I hope, open-e is working on this issue! It's really a great product, but these small things are very painful.
Is there any workaround for this virt. IP problem? Maybe MPIO with 2 NICs (one virtual, one physical)??
adding LV's after sync?
Reply With Quote
