I mention this stuff a lot and people are always surprised. So I thought I would share this here so that more of you can see this.
A software package that helps you with test plans, and do failovers is a DR orchestration tool that can use replicas or backups as the source for failovers in a crisis. As a professional DR tool it allows you to test your failovers to make sure your applications recovery is as you might hope. It also has great documentation - mostly automatic too that is very helpful. But there is other things you can do with this very useful tool that are not necessarily obvious.
- You can use a test failover to test out a few things - not just does the app work but also:
- Application Patching - if it works you know it will work in production as you are working in a copy of production, not just a lab built copy.
- OS patching or updates - same here, you will know if it works in production or not.
- Security vulnerability scanning - if you break something in the test only you know and you can prepare for that in production by maybe tweaking the scanning tool.
- Plan, and test, a migration to a new data center. It is very good indeed to test your migration before it becomes permanent.
- You can pull data into a test failover from physical machines if necessary. So like an AIX server that is running Oracle. How it would look is:
- During a test failover a PowerShell script executes and talks to AIX
- A new LPAR is deployed,
- A virtual nic is attached to the LPAR that is connected to the private and non-routing test network.
- Another script runs that copies a subset of the Oracle database to the new LPAR.
- So now in the test failover an app can find Oracle. Which is important to support a test failover.
- Now at the end of the test there is a PowerShell script that runs and notes it is the end of the test so it talks to AIX and deletes the LAPR that has the subset of Oracle database on it. This is important to make sure no test data gets into the production Oracle.
- I have seen this done with AIX, and I have had customers tell me they have done this with HPUX.
- You can protect physical machines.
- Use a tool like VMware Converter to do lunchtime and night time conversions. Then on a different schedule it is replicated and protected. This means if the physical machine is lost, a virtual version will be available, and that is quite usable even if the app owner didn’t want to virtualize for some reason as the alternative is total loss.
- I am working on a blog to do this as part of your backups. When I have that figured out and documented I will share it out.
- One that I heard from a customer is to do organized and auditable restores for people. Since it is a DR activity there is a nice audit trail, and you can test it too.
So I hope that this all helps you understand that there is a lot that a DR orchestration tool like SRM, VAO, and HCC can do over and above the typical DR stuff.
Let me know if you have questions or comments,
=== END ===