Thanks for all the replies.
After I made my post, the "forum" did suggest similar posts that mentioned some of the stuff had has been mentioned.
Ok, let me give you some background on our infrastructure.
We are a large manufacturing facility and have 2 data centres (DC) linked by fibre.
Some of our systems are linux clusters using drbd and heartbeat.
Drbd keeps the data in sync via a private fibre connection between the 2 DC while Heartbeat allows the services and cluster IP address to follow the primary server and this is what clients connect to. So if the primary server fails, the secondary can take over the resources. With some of our clusters such as the nfs server, clients will notice the interruption but the main database service, there is no "service interruption" as far as the clients are concerned. This is because the failover can occur automatically within 30s and the application backend will just retry to connect. This has proved to work when the primary did fail.
So we are in an unusual position in that we have 2 DC on site that are active-active.
With it being such a large site, it is possible that we could loose one DC without effecting the other DC or factory operations.
Of course, there are a number of systems that are either duplicated across 2 DC and single sided so the idea of introducing vmware is
to improve availability.
So If we do put in a SAN, then despite the redundancy, if power fails to that part of the site, then the SAN is lost as is the data centre.
This would not be an issue if we just had the one DC on site i.e. vmware and the replication features would allow for quick recovery but with our 2 DC, we need to have a similar failover that we have with drbd/ha. Now I understand that drbd/ha is a great solution for 1 server while vmware/SRM etc are working at the site level, but still we want to say "how can we use the resources of 2 data centres to ensure automatic failover with no data loss if one DC goes down". This may be that we do use one DC as the primary and the other as a secondary.
I now see that it can be done but and may be at a considerable cost so I am trying to understand the possibilities before we speak to any vendors who may or may not offer the most suitable solution.
Thanks.