Exchange 2003 Hot Standby Server

[wrathyimp]

Hi all,

We are implementing a disaster recovery plan for exchange 2003, we would like to setup a hot standby server for exchange 2003 (different box), which will be used when there is a disaster or a crash happens to our actual exchange 2003 server. I need to know how can i do exchange replication between the hot standby server and the primary exchange server.

Our Current Setup is Exchange 2003 Cluster with 2 nodes (Active/Passive).I have the following issues for this setup:
1. Standby Exchange should be connected to our Disk Array.
2. Not changes required at the users end.
3. acceptable down time of 2 hrs.


If anyone could point me in the right direction as to how to set this up.

 

[Zelandakh]

You want a hot standby server when you already have a cluster? What are you trying to achieve?

You think 2 hours is acceptable? That isn't hot, that's only warm.

Not sure what you are after but I think Exchange 2007 SP1 standby cluster replication might be your answer. If not, continuous clustered replication is the answer.

 

[58sniper]

Zelandakh's right. That's certainly not what I would consider a hot standby.

If a problem takes down both nodes of a cluster, then you have to look at what's common to both of them - the shared storage. I mean - let's face it, today's servers have redundant RAM, processors, power supplies, hard drives, fans, NICs, etc.

Since this isn't Exchange 2007, you don't have the CCR, SCR, or LCR options. You'd have to look into a log shipping solution provided by a third party company such as Double Take or Never Fail.

Pat Richard
Microsoft Exchange MVP

 

[danarf1]

we use Teneros

http://www.teneros.com.

Its a hot stand by appliance that installs in less than an hour with no downtime and no apps on the server/cluster.

 

[xmsre]

It sounds like you're trying to implement a standby cluster. In Exchange 2003, the standby cluster can be a single node cluster if the Primary cluster is a two node A/P cluster. Generally you would put the standby cluster in an alternate data caenter and use third party replication tools to replicate the data to it. In the event the primary data center went down, you would bring up the standby cluster in the alternate data center. An example of such a design can be found in the October 2006 issue of technet magazine.


Moving to Exchange 2007 opens up additional replication possibilites that leverage application level replication and, depening on your RPO (you state the RTO as 2 hours but don't state an RPO) may eliminate the need for 3rd party replication tools. If you want to cluster in the local data center, then you'd likely go CCR + SCR to the remote data center or SCC + SCR to the remote data center. You can do a standby cluster and use recovercms/clearcms, or leverage Exchange 2007 database portability and recover to alternate server (and use the cmdlet to rehome the mailboxes)