Some weeks ago we had a power failure which caused one of the RAID-1 disks in our primary domain controller (Windows Server 2003) to fail. This server is an IBM System x3550. Once the disk was replaced, the array rebuilt successfully, but since then an error message occurs in the event log:
One or more logical drives contain a bad stripe: controller 1.
Now I understand that the only way to get rid of this error is to break the RAID array and rebuild it - which means wiping all the data off the disk.
What would you recommend I do - should I go ahead and rebuild the array, or could I safely ignore the errors? Would ignoring the errors result in an eventual problem?
If I should rebuild the array, what's the best sequence to do this in? I was thinking the following - please advise if this is correct or if I should change something:
1. dcpromo the primary DC so it is no longer a DC - both primary and secondary DC's are global catalogs, so I assume that doing this will make the secondary DC become primary (as it'll be the only one).
2. Run NTBackup of system state and disk
3. Rebuild array, run a basic reinstall of the O/S, then restore the NTBackup
4. dcpromo the machine to make it a DC - but how do I make it the primary? Also, this machine is the primary DNS and WINS server, as well as DHCP - I assume NTBackup will take care of DHCP and dcpromo will take care of DNS/WINS replication?
Guidance would be appreciated! Thanks