Datacenter Move Best Practices

by Jeff Ouellette

Very often in the course of our business, a client will come to iuvo and ask us to assist them in moving their IT equipment to a new office or handle the move of their IT equipment from one colocation datacenter to another.  Having just completed our own datacenter move, her are 10 best practices and tips for a successful move.  Many of these also apply to setting up new IT environments as well as migrating existing services to the cloud.

1. Start early - It goes without saying that  you can never have too much lead time before the move.   Many aspects of a move like getting a new circuit from an ISP can take a month or two of lead time, be sure you understand which tasks will require some advanced planning.
2. Plan ahead -  A little planning goes a long way. Well before the day of the move, you should be able to create a plan that details all of the changes that will be made , by whom, when and in what order.  Aspects such as changing IP Addresses, DNS changes, shutdown and startup orders are all important… and remember that some changes like DNS take time to propagate.
3. Understand the impact of downtime - Most companies incur a significant cost with downtime, whether their customers  are directly affected or not.  Lost employee productivity is a significant expense.  If your customer will be impacted by the downtime, you may need to think about how you will continue to service customers during the outage.
4. Over communicate - We have come to rely on technology and expect it to be working all the time.  If you have a decent budget, you might be able to avoid downtime altogether, but many times that is cost prohibitive, especially for smaller companies.  In either case, communicate the plan early and often, and remember that your ability to send out email and  in some cases use the telephones (if you use VoIP) will cease during the move.  Be sure to communicate the impact of changes to end users as well as the status during and the results after.  Additionally, have a fall back strategy for communication during the move.  Some common methods are cell phones and chat room.
5. Verify backups - While you should be doing regular backups and verifying them regularly with a test restore, now is a good time to test that process, before you need them.    Too many times we have seen machines that were running fine for years, were shut down for a move and never spin back up.  Remember too that you need to both backup the data as well as the configuration of applications.
6. Document everything -  If you don't have good documentation about your environment, you are asking for trouble.  Verify that you have up to date network maps, runbook processes and configurations.  Take pictures of the current setup and label everything.   If you have been doing this all along, it will make formulating a plan much easier  and minimize the potential for surprises.
7. Work smarter, not harder - Utilize technology where you can.    Utilize DNS names instead of IP addresses, so that you only need to change one item during the move instead of having to go around and fix hard coded IP Addresses.
8. Avoid the temptation to make improvements -  The move itself is significant .  It may sounds like a good time to replace a piece of equipment , but even the slightest changes can add considerable time to troubleshooting because you aren't sure if the cause of something is move related or the other changes you made.
9. Test - Formulate a step by step plan that will allow you to verify that services are working correctly and validate the test plan before the move.
10. Be prepared for Issues that arise - even the best plan can't cover every possible scenario.  So, be prepared for issues that come up and have someone from each application /specialty on call for issues that come up.  A new environment lends itself to unexpected problems like power consumption spikes during the initial power on and CPU/memory load when servers are all starting at once.