Ok, so I'm "new" to SRM, even though I've been working on this project for a considerable amount of time. It seems straightforward, but there are lots of gotchas I'm learning about. My goal is to test SRM as it was set up by consultants a few years ago, but has not gone through a proof of concept, and I've been given that task.
We only have eight critical servers on our Recovery site, out of 50 or so that aren't protected. If there is a disaster, we will be running a bare-bones environment. Just a few app servers, a db server and View so people can remote in, assuming our Protected site is unavailable. I'm in LA, so I'm assuming an earthquake, for instance.
In any event, I've been trying to plan for an actual test that is the least disruptive as possible, over a weekend, and am at the point where I've been running a test of the recovery plan, but it is pretty restrictive and not very representative of an actual DR situation. For instance, I bring up the VMs, usually all of them report the error of VMWare Tools timing out. Then I try to test our db server by consoling in (MSSQL) and can open the SQL Management console, but can't actually get into any of the databases. We run SnapDrive, which doesn't mount the LUNs, but they are mounted in Storage Manager and are presented.
Since the Test environment uses an isolated network, I was wanting to use IP customization to bring the VMs up so they could talk to the AD server that already resides in production on the Recovery Site (it's a small branch office with a vCenter server, AD server, SRM server and a couple others). I'd like to create NAT rules for the connection servers that will be spun up, and be able to test those.
Another issue is the VMs are old. We SnapMirror multiple times a day, but I didn't understand the explanation from the VMWare engineer of why the latest image/snapmirror isn't being used when I run the Test Recovery Plan. It was explained that in Test, everything is brought up as read only, so every time I run a new test, it reverts back to the same (old) image.
I'm doing all this to have a more solid feeling before the ACTUAL test of a Planned Migration. That's what VMWare advised me to do to run a more realistic, robust test.
Am I on the right track? I know this is kind of long winded and not very detailed about our environment, and I'm being asked to write up a detailed project plan, but I don't know what I don't know and frankly the documentation is not the clearest.
I realize I'm not asking this in the clearest terms either, mostly because I'm a bit confused and somewhat wary that I'm going to miss some critical setup items, and will hit potentially damaging snags when doing the "real" test.
Any good documentation for best practices would be helpful, or some clarification on the items I mentioned would be great. And yes, I have been searching the community and The Google.
TIA!
Steve