cancel
Showing results for 
Search instead for 
Did you mean: 

New veritas cluster setup for exchange.

Amin123
Level 4
Partner
If I need to do a new cluster setup for veritas storage foundations for microsfot exchange.
I need to install the cluster in the current exchange servers or I need to install a new setup.??
There is a microsoft cluster ruuning there, I need the highlited steps please.

Thanks,
AMin
1 ACCEPTED SOLUTION

Accepted Solutions

Wally_Heim
Level 6
Employee
Hi Amin,

I recommend setup a new Exchange server with SFW-HA clusters and migrate your users to it.  By doing this you will be able to test the new Exchange cluster and ensure that it is working completely before you start putting production users on it.  It also gives you the out if things are not working correctly then you just migrate your users back to the existing MSCS Exchange cluster.

You will need to leave the MSCS Exchange cluster up and running for awhile after you migrate your users off of it.  The reason for this is to get your users to automatically switch to the new Exchange server.  Clients that use Outlook or Exchange clients will automatically be redirected to the new Exchange server the first time that they log in after their mail box has been moved.  This is a big time saver in reconfiguring all of your clients.

Follow the VCS Exchange solutions guide on how to setup a new Exchange Server cluster.

Thanks,
Wally

View solution in original post

5 REPLIES 5

Wally_Heim
Level 6
Employee
Hi Amin,

I recommend setup a new Exchange server with SFW-HA clusters and migrate your users to it.  By doing this you will be able to test the new Exchange cluster and ensure that it is working completely before you start putting production users on it.  It also gives you the out if things are not working correctly then you just migrate your users back to the existing MSCS Exchange cluster.

You will need to leave the MSCS Exchange cluster up and running for awhile after you migrate your users off of it.  The reason for this is to get your users to automatically switch to the new Exchange server.  Clients that use Outlook or Exchange clients will automatically be redirected to the new Exchange server the first time that they log in after their mail box has been moved.  This is a big time saver in reconfiguring all of your clients.

Follow the VCS Exchange solutions guide on how to setup a new Exchange Server cluster.

Thanks,
Wally

Amin123
Level 4
Partner
Thank You Wally,
If the customer need to use the current exchange server and he inssist on that, could you please help me with high level steps and how to achieve that, Also how this need to include removing the current microsoft cluster.

Thanks,
Amin

Wally_Heim
Level 6
Employee
HI Amin,

There are a lot of ways that this can be done.  I just prefer to have a fully tested/solid cluster that I'm trying to move to.  I always recommend having an easy way back to a known working good configuration just in case something was missed during testing.  That is why I recommend setting up a completely new cluster with a new Exchange server instance. 

If you want to use the existing Exchange server and not have a migration then you need to destroy the existing MSCS cluster (reinstall the OS on both nodes) and install your Exchange server as a standalone server using the disasterrecovery switch.  Then you can bring the standalone Exchange server into a SFW-HA/VCS cluster.

This is a messy process and does not give any fail back capabilities to a known working Exchange server if you have problems.  I really don't like this approach because you are betting that you will not have any problems during this process and it takes your Exchange server down for the entire time that the new cluster is being built.

However, If hardware is a problem then you can remove 1 of the nodes from the MSCS cluster to create a single node SFW-HA/VCS cluster to create a new Exchange instance on.  This will not allow high availability (as you have two single node clusters -1 MSCS and SFW-HA/VCS) during this process but it can be done if hardware is a limiting factor.  Again, this is not my preferred way of doing this.

Thanks,
Wally

Amin123
Level 4
Partner
Excellent comments really thank you,
If they have exchange 2007 CCR cluster setup, please just comment or change the below steps if possible.
  1. Uninstall the exchange CCR cluster. This need to remove the all exchange 2007. From the first node. While the second node of cluster is running.
  2. install the storage foundations and cluster setup in the first node and remotely in the second node.
  3. Install exchange 2007 in the first node.
  4. Move the data to the shared storage using the storage foundation GUI.
  5. In the second node remove the exchange and this will include the CCR cluster.
  6. Install exchange in the second node as disasterrecover mode.
  7. Test the failover and failback.
apprciate your kind help and support.

Wally_Heim
Level 6
Employee
Hi Amin,

Exchange 2007 CCR only works with MSCS clusters.  If you are going to a SFW-HA/VCS cluster then CCR will not work.

You also need to be care when uninstall Exchange from the servers.  Depending on how you do this it can remvoe the Exchange server configuration from AD and you could loose your current mailbox configuration.  That is why I suggested rebuilding the OS on the servers so that you start from a point similar to when a server dies.  Then you can use standard recovery processes from Microsoft to recover the failed Exchange server. 

I have not done this with Exchange 2007 before.  I'm going off of what I have done in past Exchange versions which should be the same/similar to how Exchange 2007 server recovery is done.

The moving of the Exchange Information Stores to shared disk should be done by the VCS Exchange Setup Wizard's move Exchange Database option.  This will allow you to move all of the items needed to fail the Exchange Vitual Server (EVS) to another node.

Again, you should really push your customer to install a new Exchange server and migrate to it.  This will ensure a happy ending and covers failback to a working good configuration if needed.  By reusing the existing MSCS nodes you have no way to failback to a known working good configuration if there are problems (because the working MSCS cluster was destroyed in this process.)  This is the safest way to proceed and it also offers very little downtime to the customer (just the time it takes to migrate users from one EVS to another.)

Thanks,
Wally