I was checking the features list for something like failover but couldn't find something like that. A through the forum on failover gave me this http://forum.open-e.com/showthread.p...light=failover topic...
In here there's an anouncement for july/august 2006 for automatic failover...is it already there and if not when will it be available ?

For my iscsi solution I had two types of setup in my mind, depending on what will be possible, could someone tell me if they're possible :

Setup 1
- Two storage servers with local disks.
- Both nodes export half of the total available lun's to spread the load.
- Both nodes replicate their active lun's to the other node to keep in sync.
- In case of failure the still active node exports all lun's. (possibly automatically)

Setup 2
- One or more FC storage arrays with redundant controller
- Two DDS storage servers as FC > iSCSI gateways
- Both nodes export half of the lun's that reside on the FC arrays, to spread load.
- In case of failure the still active node exports all lun's. (possibly automatically)

In the second case I use Open-E as a storage gateway to virtualize underlying arrays and to provide snapshotting and a decent gui for management.
For both setups automatic failover would be very nice...manual failover in my case means all xen rootfilesystems running from SAN will be dead first and all vm's have to be restarted.

I was also wondering how the manual failover procedure works and if it's possible to script it from the outside ?

Regards,

Werner Reuser
XL-Data Hosting