ILM “2” RC0 + Named SQL Instance = Bad

A couple of weeks ago I was busy setting up ILM “2” RC0 for my session demo at TEC Europe. I’ve installed ILM “2” a bunch of times, and if you have the prerequisites properly installed, it’s pretty much a no-brainer. It normally takes me less than an hour to get from a new machine image to a running ILM “2”. This time was no different. I built my ILM “2” image, populated it, and tested the Quest PowerShell cmdlets for ILM “2” to make sure my session demos were all functioning and ready to go to Europe. Everything was fine. I shut down the machines, copied their Hyper-V images to my external USB hard drive, and moved on to cleaning up my Powerpoint slides to reflect the significant changes in FIM 2010 RC1.

When I got to Berlin, I rounded up a server from our A/V provider (an adventure in itself), cobbled up the networking to work with the hotel wired internet, copied the images, and started fiddling with them to get them to work on the new network. When I got everything sorted, I started testing my Powershell->ILM “2” demos. I couldn’t even connect to the ILM “2” web service. And as a bonus, the ILM “2” Sharepoint portal wouldn’t even start, failing with the infamous “unexpected error”. Hmmm. Why would previously working VMs suddenly start failing?

I had a lots of other things going on leading up to TEC, so I didn’t get to spend a lot of concentrated time on the problem, but ultimately I deduced that the ILM “2'” service wasn’t starting. the event log indicated that SQL “might not be installed”, but SQL clearly was installed and running. I rolled back the images to an earlier snapshhot that only had the prerequisites installed, reinstalled ILM “2”, and everything was swell. Mystery bug, fixed by reinstall, case closed. Or maybe not.

When I came back to my room to walk through my demos one last time before my session, I encountered the same problem. I couldn’t establish a connection to the web service, the portal wouldn’t run, and the ILM “2” service wouldn’t start. I rolled back the images, reinstalled, and everything was fine. I rebooted the ILM “2” server, and I encountered the same failure. I tried the sequence again, just to make sure I was seeing what I thought I was seeing. ILM “2” would run properly after install, but would fail to start after a reboot. Very curious. And I was running out of time.

Maybe there was something wrong with the prerequistes? I rolled back to a bare WS2008 image, and reinstalled the prerequisites, and reinstalled ILM “2” RC0. It worked. I rebooted. It failed. Damn! At this point I started working out how I could move the server down to the conference area without rebooting the images. But as I thought through what was happening, I realized that I was doing something different this time around compared to other times I had installed ILM “2”. In earlier versions, I installed SQL with a default instance. For some reason, this time around, I specified an instance name. So I rolled back to the base OS image, reinstalled the prerequisites, but this time specifying the default SQL instance. ILM “2” started and worked properly. I rebooted the image. And it continued to work. Aha!

So word to the wise: Use the default SQL instance with ILM “2” RC0.