On Friday we successfully executed a 20-server move from one colocation facility to another across town.  The new site has much more room for expansion, is more secure, and will save the company a few hundred thousand dollars a year in related costs.  So it was a *good thing*.

Technically, the best part about it was being able to perform the move during business hours.  There is enough redundancy in the systems now to allow the full shutdown of DCs, an Exchange server, a huge portion of the Cisco Call Center environment, other support and security systems, and the end users never notice.  That is a fantastic feeling, knowing that designs are coming together properly to allow such controlled failures, without affecting the business.

Over the course of this move, I’ve learned a few things:

  • Exchange servers appropriately update themselves upon an IP subnet change with no errors.  I expected a 2nd reboot to be required after the server recognized it was in an completely different AD site.
  • Physically moving DCs to a new site is also extremely easy, provided you update DNS / WINS appropriately.
  • Cleaning up bad subnets in AD Sites and Services is a pain, because it’s so tiresome doing all the subnet calculations over and over again.  Worth it, but boring.
  • Security guards at building docks can be real jerks, or can be really easy to work with.  Leaving the building, the movers kept having to drive around the building as we went back in for another pallet’s worth of equipment.

Now it’s time to rebuild servers that have been brought back to the office to be re-deployed with new OS’s in the DMZ, and finish building that environment properly!