We need to come up with a new HA file server (user drive data) solution complete with DR. It needs to be 2008 R2, cater for about 25TB of data, and be suitable for 500 users (nothing high end on I/O). I don't want to rely on DFS for any form of resilience due to its limitations for open files. We have two active-active data centers (a third can be used for file share quorum).
We could entertain:
1)
Site1 - 2 x HP ProLiants with MSA storage, replicating with something like DoubleTake to a third HP Proliant at site 2 for DR.
2)
Site1 - 2 x HP ProLiants with local storage and VSA or HP StoreVirtual array (aka LeftHand), using SAN replication to site 2 where we could have a one or two node config of the same setup.
Ideally I would like all 3/4 nodes in these configurations to be part of the same multi-site cluster to ensure resources like file shares are in sync. With two pieces of storage across this single cluster (either a DoubleTake or SAN replication to local disks in DR) will this work? How will the cluster/SAN fail over the storage?
We do have VMWare 5.0/1 (not 5.5 yet). We don't have Hyper-V yet either. Any thoughts on the above, and possible alternatives welcome. HA failover RTO we'd like in seconds. DR longer, perhaps 30 mins.
Thanks in advance for any thoughts and guidance.