Being a bit of a geek I run a server at home so laptops, workstations and media computer are all connected to a domain.

The server runs Windows Small Business Server (SBS) 2011 (upgraded from the earlier version) it has (had) 2 640GB hard drives, mirrored for recoverability, and it is backed up every day, so you would think that when something goes wrong it is easy to recover.

On Wednesday night the server went down and it turned out to be a hard drive failure. Since they are several years old I shouldn’t be too surprised when one of them give up the ghost but in this case it turned out to be both of them; which probably means I didn’t pay attention to the warning events.

Now the idea of having mirrored drives is that you take out the failed one, put a new one in and they synchronise without loss of service, apart for the time it takes to do the swap. Unfortunately when they both go you are left with no option but to rebuild; I hear you asking ‘what about the backup?’ yup, that would work fine if your server wasn’t the domain controller(DC). The beauty of SBS is that it is an all in one server with DC, Exchange, SQL Server and internal SharePoint web site. The down side is that when you have to rebuild, in my case using 2 shinny new 2GB drives, you are creating a new domain. New domain means that all the users and other computers have to join the new domain -and redo their settings.

So the backup – yes all of the file based data was recovered; E-mail was recovered by exporting to a pst then importing it after the rebuild but I couldn’t recover the SharePoint site because all of the running credentials were from the old domain. Not really an issue as it wasn’t used much so I decided it was too much trouble to adjust the credentials and I took the opportunity install SharePoint 2013.

Anyway 2 days later I am just about getting everything back to normal. I have decided to knock up a secondary DC using a virtual machine on my workstation so if it happens again I will not have to go through the pain of sorting out all of the user systems again; and I recommend anyone running a single DC do the same.