Hi,
I'm currently running two esx servers connected to a SanMelody storage server via iSCSI. I also have a second SANMelody box with lower grade storage (SATA rather than the SAS in the primary SAN). The LUNS are formatted with vmfs3
I want to replicate the contents of the first SAN to the second using AIM, so in the event of a failure on the primary node I can start up our standby ESX host which points to the second SAN and fire up the VM's.
My question is this:As it's asynch. replication, how do I ensure that the stuff tht's been replicated to the second SAN is consistent - as I see it, if SAN-1 suddenly dies, there will be writes that have completed on SAN-1 that haven't been replicated to SAN-2. What do I need to do to ensure the vmdk's that are replicated to SAN-2 are consistent - presumably snapshotting and then replicating the snapshots?.
Ideally I'd like to hear from someone who has done this before - I know quite a few people use this kind of setup for DR, although most of the stuff I have read involves backing up using something like vRanger and then replicating the backups.
Thanks for any advice.
Cheers,
Mark
