Smooth Moving Ahead
Fortunately, shutting down services is avoidable today due to the availability of hosted data centers and cheap hardware. But hosting applications requires extensive preplanning so that any staging and production requirements are met without a hitch. While there is a huge downside to transferring critical systems onto foreign environments, administrators have to bite the bullet and make hard choices to make sure applications continue to run.
Plan Of Attack
Ultimately, it's the system administrators that end up doing the grunt work, so good communication between the different IT groups, including architects and engineers, can lead to a headache-free relocation.
Recently, the CRN Channel Test Center lab moved between floors at our headquarters in Manhasset, N.Y. To get there, a "miniplan" was created so that our infrastructure would only be down for a maximum of two to three days. In the middle of mapping it all out, floor planning changed as well as power sources, UPS service and network layout. To minimize downtime, we quickly devised a plan of action. Here's a step-by-step look at how this was done:
1. Coordinate your move with facilities, electricians and IT staff.
From the beginning, we told each group precisely what infrastructure equipment and racks we needed to transfer. The new server room was smaller, so there was a potential space problem. We figured out the minimum number of racks and shelves that could both fit in the server room and house our equipment.
2. Shut down, but stay up.
On the day of the move, we came up with a simple but effective plan of action so that our editorial priorities would be met. Our staff here began removing equipment and dismantling without shutting down key network access.
3. Manage cooling and other services.
Although the Test Center is growing and our service requirements will change, the new server room requires less cooling infrastructure because of its design. When you move a data center, you will also need to talk to electricians about issues such as maximum loads in new server rooms to make sure future growth can be accommodated. Be sure to keep CIOs and CEOs in the loop.
4. Maintain a data center heartbeat.
Data transfers and keeping critical applications was never an issue. However, based on our experience, the least costly course of action is to split the job. In other words, have two data centers functional while the move is taking place. VARs should recommend splitting up locations to maintain service continuity. This step eliminates the need for hosted infrastructures.
5. Save yourself time when getting back up and running.
Don't cut network wires from racks and network panels unless it's absolutely necessary. The less cutting, the less costly the job and the faster a client can be back up and running. To speed up the process for the Test Center's move, staff here kept most of the wiring intact and labeled any network wires we unhooked.
6. Test, test, test.
Don't be satisfied with a minimum setup. Test network connections, external services and servers while the relocators are still under contract. Relocating the data center can be an eye-opening experience, as many small details that are daily dismissed become vitally important.
Capacity planning also became crucial. Despite some hiccups, we were able to keep disruption minimal and we're confident that any unforeseen power and growth requirements during the move shouldn't be a problem.