4.0 System Design

Essential Steps

When you setup GroupWise Disaster Recovery there are a few things you need to do:

Designing for Disaster

Just adding a GWDR server to your network and backing up the domain and post offices is not necessarily going to help all that much in a disaster, depending on what type of disaster your system might experience. Proper system design will make all the difference. The following scenarios will guide you in your system design goals.

At its most fundamental level an email system must:

Setting up a system that does this is easy, but it can be fragile. However, setting up a system that will be able to be robust in the face of disaster is not hard. We will go through some of the considerations so you can decide how you want to design your system.

We have noticed that some customers, out of ignorance, set up their system in a fragile manner. They place the Domain, Post office, and GWIA all on one server, and have their clients connect via an IP address to the GroupWise server. This is an easy and common configuration, but is also very fragile.

When you suffer a failure of the GW server, you can switch the GWDR server to Disaster Recovery mode, but you will have to tell each user to change the IP address the client uses, and they will only be able to email between users on the system. However, without a GWIA your users will not be able to send or receive mail from the Internet. So this will work if just the MTA or POA goes down and won’t come up for some reason.

We can make this system more robust by making a few minor changes. You can add robustness by distributing functions across servers so that if one goes down another can be used.

You can set up an A record in your DNS so that the GroupWise client will connect to the server you specify with a hostname instead of an IP address. This way you only have to change the A record in the DNS when you need to have your users connect to the GWDR server.

Then you setup the GWIA on a separate server. This way when the GW server fails the GWDR server can use the GWIA for mail flow to the Internet.

This allows your system to deal with the GW server going down. This is better, but this design will not be able to withstand multiple server failures. For that we need to design for site robustness.

Site robustness is having a secondary, off-site, or disaster recovery site.

If you suffer a site-level disaster there are several things that need to happen to insure business continuity.

You may want a secondary DNS where you can update the A record that will direct the users to the proper location. You may want to set up a webaccess server so that users can use that if they don’t have access to their normal client computer.

GW allows you to set up a secondary GWIA that it can automatically failover to during a failure of the primary GWIA.

GWDR is designed around the concept of the human being the final authority. There can be many minor interruptions of the primary system that do not require a disaster recovery failover, it might be as minor as the primary server going into nightly maintenance mode. That is not something worth going into disaster recovery mode for. So we let the human make the final decision to enable disaster recovery mode.

GWDR can backup multiple domains and post offices and run them in disaster recovery mode. It might run like a herd of turtles through a field of peanut butter but it will run, depending on the resources available. If you wish to run multiple post offices at the same time, you will need to bind an IP address for each to use.

How you implement your GroupWise Disaster Recovery system also depends on what your GroupWise is hosted on. If hosted on SLES Linux then you can use the Collector model, if Windows then you must use the Server Only model.

GroupWise Disaster Recovery is a backup system and recovery system. It is built around the idea that two is one and one is none, in other words. if you have two copies of your data then if something happens you still have one copy available.

GroupWise Disaster Recovery can be configured save a backup of your production system, send a copy to an off-site server and to a cloud host to maximize data safety.