Microsoft, Exchange, 2003, Deploy a second exchange server Using a SAN for mailbox and log storage

I have a SAN pair that replicates over a high speed VPN to my DR site. I want to deploy a second exchange server without clustering that I can fail over to manually if neccesary. How do I install the second server so it can connect to the replicated (SAN) drives. I do not want to install a cluster at this time. I will do that later. I just need a warm server standing by.
Who is Participating?
j0lxConnect With a Mentor Commented:
For remote site warm DR of a single Exchange server without clustering, I recommend you consider solutions such as Double-Take and XOSoft WANSyncHA.

We use Double-Take in production to protect our Exchange system, and we have implemented WANSyncHA for several customers to protect a number of bespoke applications that required some scripting to create our own custom agents.

These solutiions will automate and manage the monitoring and failover of the source Exchange system, inlcuding the tasks you mentioned in article ID: 21861843.

You will need to change your replication architecture from SAN replication to server-based replication for these solutions to work.

These solutions implement a suite of components that work together to perform replication and disaster recovery.

The first is the replication component. On a Windows server, they generally use a filter driver at the file system level to capture the replication stream. The significance of this is that the data captured is file-system aware and enables the target DR server to maintain non-corrupted recovery points in the file system. In contract, a SAN based mirror or real-time replica will not have this consistency.

The second component is a state monitoring service that keeps a heartbeat between the two server, but more importantly, monitors the SERVICE being provided to the user. If the target server detects that the source server is not doing this, the event triggers a rule that determines what happens next.

The third component is the failover or takeover component. This component interacts with Active Directory, DNS, WINS, TCP/IP and ARP to allow one or more of the following:
- the target server takes over the NETBIOS name of the source server, does a dynamic DNS update and starts delivering the service
- the target server takes over the IP address of the source server, does a gratuitous ARP and starts delivering the service under its own name
- the target server does a dynamic update of the DNS records without taking over the NETBIOS name of the source server
The actual method will vary depending on the application being managed. Most failover applications now have optional prebuilt agents for file system, SQL and Exchange at the very least.

The advantages of using failvoer software are:
- cheaper and less complex than cluster or global cluster solutions
- the agents 'understand' the application they are managing and will do the necessary actions to gracefully or ungracefully take over with least risk of data corruption
- some have file system hooks to allow continuous protection, fire-drill and remote snapshot backup options
- they only replicate checkpointed changes rather than wholesale block writes at a SAN level (eg: user updates same block 10 times before committing the email by clicking 'send')

You'll need to be aware of requirements such as bandwidth and redundant network connectivity as well as your AD/DNS architecture.

I don't know how you are going to set it up as a "warm" server without a cluster.

I expect you will either need to cluster up, or set it up as a cold server for a bare metal restore.

I suppose you try it by setting up exchange as a normal second server, getting it up and running and "working" then when it all goes pearshaped you can manually point to the SAN for the databases and have them mount, but that is a pretty complex procedure - and it raises other problems - if the main server dies and you just remount the stores, you will need to reconfigure the firewall and outlook clients to point at the new server - an image of the original server, restored to the new server would just pick up natively
dyesjampAuthor Commented:
Correct me if I am wrong. I wouldneed to do the folowing:
1. install exchange
2. Do not connect tot he message stores yet.
3.Wait for a disaster and then attach the DR exchange server to the SAN cluster at DR site.
4. change firewall rule or preconfigure to allow mail through to that IP address from an externally NATed address.
5. Change the MX record in DNS
6. Have exchange on my DR laptops preconfigured to go to the DR exchange server since they will not be used for mail unless there is a problem withe the main site.

Would something like that work? I need to implememnt a shorterm solution next week when I travel to the DR site. I can work on a cluster for a permanent solution but I do not think I have time to get everything ready before my trip. I am the new network administrator and have been in place for a week now.
dyesjampAuthor Commented:
Thank you for your help.
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.