asked on

Clustering Exchange 2010 in Vsphere 4 ESX environment...

We have Vsphere 4 Up1 and I have an Exchange server (2010). However, I would like to run something like clustering so that there is no risk of failure of the mail system. I was wondering what kind of setup should I be using...

Vshere 4 has features like High Availability and DRS, etc. that can failover things automatically but should I be using the built in features of the guest OS here to use Microsoft clustering. What's the best way to rule out hardware/software/host failure when deploying Exchange 2010 in a virtual environment?

coolsport00

Using MS Clustering is your ultimate best bet to severly minimize or discount any kind of downtime to your server. My org doesn't use it, but that is your best bet. What I have set up in my shop is an ESX Cluster. That way, if 1 host goes down, my Exchg VM won't. For potential Exchg VM failure, what I have set up is a 'stand-by' VM at a DR location. If my Exchg VM crashes for whatever reason, I will power up this DR location VM, restore the DBs with my b/u solution, apply guest OS updates and I'm good to go....less that 30mins of downtime. For my b/u solution, I use EMC's Avamar, in which I have 2 nodes...1 at my main branch location, and a "b/u node" and my DR location. That certainly plays a part in minimizing downtime as 1. it's at the DR location and, 2. Avamar uses de-dup'ing.

Hope that help.
Regards,
~coolsport00

wlandymore

ASKER

yeah, I have the Exchange box on a Netapp SAN and then running on an ESX cluster, I'm just trying to figure out if I can run the regular MS cluster as well so that I've got two Exchange boxes answering on one IP so if one goes does it still stays up.

ASKER CERTIFIED SOLUTION

coolsport00

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

wlandymore

ASKER

Thanks. That's really helpful.

one other question....what happens in the event of a database corruption when you're running things like this? (IE - Database Availability Group, etc.)

Does it replicate the corruption across to the other node so then you would have to use the SAN snapshot, etc. to get it back or is there anything that guards against this?

Thanks.

coolsport00

Unfortunately I'm not entirely sure as I've never run any form of MS Clustering. Since it's pretty much the same server across several, my guess would be yes, it would corrupt the others. I would submit this question on the Exchange Zone to see what they say (I'm on there, too, but obviously can't answer your Clustering question) :) You might also add this Cluster question to the Server Zone, as well.

Regards,
~coolsport00

wlandymore

ASKER

I did find another post with this:

"During day to day operations, the database isn't replicated, it is the transaction logs. The other servers then build a database from those logs. Therefore it is possible that in some scenarios corruption may be replicated, but in others it may not.

If you discover a database has become corrupted and choose to replace it, then that will resolve the initial issue, but if the source has some unidentified corruption, then that may well be replicated across. "

So I guess the best way to keep exchange running is with the Database Availability Group (cluster) in Exchange 2010 and then to take frequent snapshots with the SAN (Snapmirror for Exchange with Netapp) and then you can roll back the db if there is some form of corruption)

Thanks for the help.