Primary Domain controller dead

[maniac]

Can someone edit the title please! Comain should be Domain! DONE

 

We're in the middle of migrating to our new domain at the moment and It was going quite well until this afternoon. I'd copied across some data from the old to the new and managed to decomission 2 of our existing servers. So I removed them from the rack today and had a bit of a re-organisation which involved most of the server being removed from the rack in the end and moved to new positions to accomodate everything in some sort of logical order.

 

Came to powering everything back on again, all the new servers started up fine, I then started up server 1 for our old RMCC3 network and it started loading windows fine, so I switch on everything else. Come back to check on them about 2 or 3 minutes later, and notice that the old server 1 is just displaying a gray screen, looks like it got stuck on startup somewhere. So I leave it another ten minutes and still nothing so no other optiont but to hit the reset switch. Here's where it goes horribly wrong.

 

Server won't start, and the RAID controller is reporting 3 disc failures! Yes that's 3 discs out of a possible 6, absolutely great! So now it's officially dead. It's not a major major problem, as the backup completed sucessfully last night so I have all the data from it safe which happens to be all the staff files (Tape currently residing in my house for ultra safe keeping!) but it is a problem in that I still need to migrate exchange except the old exchange server can't authenticate to anything because that server was performing DNS for the old domain.

 

Would I get away with renaming the old Server 2 to server 1 and changing the IP number, then installing DNS onto that to replace the now missing DNS server? Would that work and allow exchange to log on (there are still 3 other domain controllers from the old domain, so there's servers there for it to autneticate to if it can find them)

 

Or could I add the old servers 1's IP number to my new network domain controller then add the old networks domain to that DNS server so it can resolve names? Would that work?

 

I only need it running so I can log onto the exchange box and migrate across all the e-mail, once that task is complete, then the old domain can be trashed - two more days that's all I needed!

 

Looks like staff won't be getting their laptops back as early as I hoped!

 

Any help appreciated!

 

Mike.

Edited July 22, 2009 by EduTech

[Michael]

Server won't start, and the RAID controller is reporting 3 disc failures! Yes that's 3 discs out of a possible 6, absolutely great!

 

If you've been physically moving these servers, have you checked the data cable SCSI/SATA that it hasn't become disconnected or half connected. I've seen it happen before!

 

What RAID level are running on the server which has failed? If you have other old domain controllers can you not take hard drives out of those, re-create the RAID array, restore from your backup and then migrate Exchange as originally planned?

 

If you need to install Windows beforehand, then you'd be correct in saying you need to give it the same computer name and IP configuration, but do not join it to the domain. Perform a direct restore at this point and the domain should be fully restored.

[maniac]

If you've been physically moving these servers, have you checked the data cable SCSI/SATA that it hasn't become disconnected or half connected. I've seen it happen before!

 

What RAID level are running on the server which has failed? If you have other old domain controllers can you not take hard drives out of those, re-create the RAID array, restore from your backup and then migrate Exchange as originally planned?

 

If you need to install Windows beforehand, then you'd be correct in saying you need to give it the same computer name and IP configuration, but do not join it to the domain. Perform a direct restore at this point and the domain should be fully restored.

 

I checked everything I could check inside the server, there doesn't seem to be anything obviously loose. The 6 disks were running as a single RAID5 configuration (courtesy of RM) all my new servers have seperate arrays for the OS and the data.

 

The other servers I've decomissioned are significantly older than the one that's gone wrong, so the drives are different capacities. I was planning to keep this server and re-use it on the new domain as it's only 18 months old! I'm still really intrigued as to how 3 disks can fail in one go, I was extremely careful with the server as I always am when moving them around.

 

I've been remoted into work this evening, and luckily for me the credentials for administrator were cached on the exchange server so I am at least able to log onto it. I've also reset the local admin password so I can't get locked out of the server now - This should allow me to migrate the mailboxes, as I'm pretty sure the migration tool will work without the directory services although I can't run system manager at the moment, so we'll have to see on that one. I'm sure they'll be a way round it somehow. I've also been told by one of my techs that Server 2 may have also have been running DNS even thou it wasn't actively used, so I'll be powering that on in the morning to see if he's right, this may give me a functional domain again if I'm lucky. If so then I think we'll be OK - just have to hope that the backup tape works, and we'll emerge relatively unscathed from this incident!

 

Mike.

[Michael]

Generally speaking all Domain Controllers do run DNS as it's recommended to have it integrated with Active Directory. You can of course run it seperately, but this is just more work.

 

It's easy enough to tell if DNS is running as I am sure you know. As for the backup, it is possible bad information could of been backed up, but hopefully you have multiple backups so this shouldn't be a problem.

[maniac]

All worked out well. All the domain controllers were running DNS luckily, so I was able to power on one of the others and get the domain functioning again. Exchange is working so I can migrate the mailboxes which was a great relief.

 

Backup tape was OK, put the AIT drive in another server installed backupexec and away it went. It's currently restoring, 22GGB done and another 6 hours of time remaining.

 

Mike.