There are a few ways to migrate a server to a new piece of hardware – backup and restore or manual. Depending on how old the server is and how complex of an environment it is, one method may be favored over the other. I almost always perform migrations manually. As long as your old server is still running, it will give you the opportunity to do some “cleaning up” of the old data. It’s like moving to a new house – do you just up and bring everything blindly arranging things the exact same way in your new home, or do you take some extra time and decide to get rid of Granma’s dentures and arrange your furniture in a way that makes more sense for the new home? I vote for the latter.
Here’s a sample of what my general outline would be. Make sure to review this and fill in any blanks for your organization.
- IP Address for new server – will it be the same or new?
- Name for the new server
- What services are running on the old server? Do you need all of them running on the new server? Can any services be retired?
- What applications are installed on the old server? Do you need all of them installed on the new server? (Look in Add\Remove programs)
- What data needs to be migrated? Should some of the old data be archived to tape or optical media?
- What should the folder structure be for the new server? Should the old data be reorganized?
- What should the permissions be for the data structure?
- Have you notified end users of your plan to migrate?
- Have you taken a backup before the migration?
- Did you take a backup after the migration?
This list could go on, but these should give you a good start. Once you’re ready to do the migration, follow your list manually moving data and installing applications and services that are needed. The hardest part of doing a server migration isn’t the migration itself – it’s the planning! With proper planning, the migration should go off without a hitch.