Have question concerning migration of an existing SQL 2000 (2-node, active/passive) Cluster to a new SAN disk array. My plan is to do the following:
1. Existing SQL cluster environment is as follows:
- Q: current quorum drive
- F: current data drive
- G: current log drive
2. Add new drives from new SAN to the cluster using Cluster Admin
- X: future quorom drive
- H: future data drive
- I: future log drive
3. Move quorum from Q: to X: using Cluster Admin. Verify quorum is okay then remove Q:
4. Add H: and I: drives to SQL Group in Cluster Admin
5. Backup databases on Cluster
6. Set SQL database offline in Cluster Admin (shutdown)
7. Copy all SQL data from F: to H: and logs from G: to I: (old to new)
8. Add new drives H: and I: as dependancy for SQL in Cluster Admin
9. In disk administrator, rename F: to Y: and G: to Z: (old to vacant)
10. In disk administrator, rename H: to F: and I: to G: (new to old)
11. In cluster admin, rename F: to Y:, G: to Z:, H: to F:, and I: to G:
12. Remove drives Y: and Z: dependency from SQL in Cluster Admin
13. Start SQL (online) and hope all goes well..
As always, I plan on testing this in the QA environment prior to actual migration in production. I know I can also detach, copy, and re-attach each database from the existing drives to the new drives instead of renaming drive letters, but I did not know if this would work for the system databases (master, msdb, model).
Does any one have comments and/or a better way to approach this? Setting up a new SQL instance on the cluster is not an option because of the number of applications and remote-users that would have to be modified.