LAB02: Site Recovery Manager (SRM)

LAB02: Site Recovery Manager (SRM)

(Drill Instructor Dan Anderson and Tim Meyers)

This Lab was great. When entering class and finding your seat, there was this huge American dude standing in front of class. When taking us through the first powerpoint slides, he was constantly walking from left to right and back again, remembering us that “We’re talking smokin’ craters here!!!” when talking about what happened to our production site. And he made also very sure that we wouldn’t press the “RUN fail over now” button, which would mess up the lab. Yeah, don’t mess with the instructor.

I’ve seen a number of presentations of SRM already, which all turned out to be rather dull slides about what the product could do, but I wanted to get some hands on and this was very well possible during this lab. We got to configure two sites, create the replication and do a test fail over. It gave me a really good idea about what SRM is going to offer to me and even more to the customers I work for.

With SRM you define a number of protection groups that include the items you want to fail over, things like VMs (wow really???), resource pools, datastores, networks. On your primary site you will have your Virtual Center with SRM. SRM can be installed on the VC server or on a separate machine, it can use the same SQL server as your Virtual Center, just make sure it uses a different database, because it will happily clean your VC database on install if you by accident give the VC database name. On your secondary or recovery site, you will also install the same combination (VC and SRM). SRM heavily relies on the storage replication by your SAN. You need to make sure that the LUNs you want to fail over on your primary site are always in sync with the LUNs on the secondary site. It is essential for your recovery.

An important part of the replication and fail over mechanism is the “Storage Replication Adapter”. This is not a hardware like adapter, but a software plug in for SRM that should be created by your storage vendor. What SRM globally does when a fail over is initiated:

– Stop the running VMs on your primary site, because the disaster happening might not be a total disaster yet and you might have VMs that are still running. You want to shut them cleanly first.

– Then SRM tells the Storage Replication Adapter to break the replication of the LUNs from primary to secondary site and make the secondary site LUNs available to the ESX hosts on the secondary site.

– Next it will start the VMs on the ESX hosts on the secondary site, using the replicated LUNs.

Important to know is that SRM will not manage your LUNs, it will not talk to your storage, the Storage Replication Adapter will do this. SRM just kicks off this SRA script and tell it to do its job. This is why your storage vendor should write the Storage Replication Adapter. They know best how to talk to their SAN. So for your its essential to ring your storage vendor tomorrow and ask them when they will be ready for SRM. Without their support, nothing is gonna happen !!! Let them know you want it.

It was a great lab although there was not really anything I hadn’t heard already in earlier presentations, but getting a feel of it was great. I was even able to draw their attention to the fact that the small printed text in green and red, was very difficult to read for a colorblind guy like me. Love to see if they listened and changed this in the final product.