The server has an 80GB disk in it. Not much, but enough for our purposes.

So, I buy another 80GB disk so that I can put in RAID 1. It’s only software RAID, but it is better than nothing.

But the disks have different size heads and # of cylinders… so they won’t match up block for block. Not possible, even if I don’t use any space on the new disk for swap (who RAIDs swap anyway?).

So, the order of events is as follows:

1. Waste a day trying to do the impossible.
2. Get another disk identical to the new disk (can’t find an identical disk to the production one).
3. Copy all data from production disk to the new disk.
4. Remove the old production disk; promote the new disk to the production disk.
5. Make sure everything boots.
6. Put in the mate to the new production disk; partition, mate via RAID 1, sync.
7. Test the RAID to make sure it works.
8. Burn in.
9. Send to colo facility, hope for the best.

We are on number 2. And we haven’t even touched on migrating the SQL and LDAP databases.

Why, yes, there will be more downtime. Why do you ask?