Exchange 2007 LCR in Disaster Recovery

By rmashadi ·
We are in middle of upgrade or basically in coexistence mode with exchange 2000. I have activated LCR for my Exchange 2007 storage group on a separate drive. My Exchange server has the following config for its logs and databases;
C: (Array 1 Local to machine) OS +Exchange 2007 SP1
(Array 2 Local to machine) storage group Databases public folders an mailbox
E: (Array 3 Local to machine) Transaction logs.
My idea is to use LCR as an in house recovery method of total disaster. Meaning Exchange server totally gone. In other words, No C:, no and no E: drives available. I am planning to create a VM version of current Exchange and activate it if exchange server crashes. Obviously the host that will run VM version of Exchange will have virtual path to E: and even though the info in those drives might be outdated, as well as a G: drive to ISCSI SAN to access LCR copy of Database and logs.
My only worry is; when I activate the VM, would it simply ignore what it has in its logs folder and read the LCR copy? Or will complain that has no clue on what logs are not valid and basically dies on me.
Would be nice to tell exchange to forget what logs have been applied and replicated and basically blindly read the LCR copy.

Anyone has run into this situation before?

Many thanks for reading this :)

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Answers

Collapse -

Thought of the same thing

by adam.stepien In reply to Exchange 2007 LCR in Disa ...


I, independently, have thought about exactly the same scenario so maybe it's an indication we're not crazy.

What's driving this solution is my trying to achieve good recources utilization and have HA/DR with limited cost.

With this solution, I will have only 1 physical Exchange 2007 server, with all the roles on it (no Edge Transport).

I have some 2000 users in 1 location + e-mail archiving system (mailboxes are quite small) + anti-spam/virus appliance for Internet mail (inbound/outbound traffic not huge on Exchange) and if I get 2 socket quad-core server, I can easily run the whole Exchange infrastructure on it.

In order to have redundancy/DR I would p2v the Exchange server (without the store drives) with Platesping on regular basis (even a few times a day) to VMWare cluster in my secondary data center and LCR would copy data to an iSCSI drive in my secondary data center (I have Gigabit connectivity).

In case of disaster/server failure, I bring up my VM on ESX in my secondary data center and I'm back in business. With 4 vCPUs the server might be a bit busy but it will be up only until I can set up a replacement physical box. Server name, IP remains the same so I don't have to worry about repointing anything.

Richardo99, I would think you'd want to use MS software iSCSI initiator on the physical server to access the iSCSI drive (LCR target) to avoid zoning issues on your iSCSI SAN. Any thoughts on this?

Are you planning to do any testing of this scenario?

Related Discussions

Related Forums