Hi all,
I'm designing a solution for a customer and I'm running into a lot of problems about the replication.
The production storage array is an EMC DMX-4 Symmetrix array, the DR array is a DMX-3 and there is also another DMX-3 (bunker) at the production site.
Replication will be done by SRDF/STAR cascaded, synchronous between production and bunker and asynchronous between the bunker and the remote DR site.
The customer has a Wintel production environment with:
Physical servers with some volumes on the DMX-4
Virtual servers with vmdk disks
Virtual servers with vmdk disks and RDM LUNs on the DMX-4
Some of these servers are running complicated services, such as databases.
My understanding is that I can replicate both the VMFS and RDM datastores at storage level, and then mount the datastores on the DR ESX servers, having to resignature the VMFS volumes and
update the link files for the RDMs to point to the correct LUN IDs etc.
The problem is that I want to be sure that the data available after a disaster are reliable and consistent at OS and Application level, so I was considering the use of EMC Replication Manager,
but unfortunately it seems that it doesn't support the STAR solution (only SRDF-S).
Does anyone have a solution that allows me to have consistent replicated LUNs in this environment?
Thanks,
Daniele
