ServerRaid questions

[yert33]

Net5000 running NT 4.0 server. Three 9.1 GB drives in RAID 5 array.
Box won't boot - says: \winnt\system32\config\system is missing or corrupt.

Upon reboot, ServerRaid reports:
Controller 1, Slot 5, Logical Drive = 1, Other = 0
Firmware = 3.72.00, Status = OK
Warning: 1 logical drive is critical
Warning: 1 online drive defunct
Press Ctl-I to access mini-config program

The IBM p/n for all 3 drives is 25L1953

Additional fact: the drive has been partitioned into a C: and a D: drive.

Last fact: there are no (amber) lights on any of the 3 hotswap drives.

Question1: Does the fact that there are no (amber) lights on the drives tell me that the drives may be good, but the raid config is bad?

Question2: I have similar hot swap drives in stock but another IBM p/n 25L1893. Where can I find out if these are compatible enough to rebuild?

Thanks a million in advance....

 

[yert33]

New info: top drive DOES have an amber light on it.

 

[yert33]

erase, erase, erase. I had my guy go into the Ctl-I program and read to me what he saw. Looks like drive 0 is defunct and the other 2 are OK. At least they say "Online".

New question: when I locate another drive do I stick it in and boot up the computer? Will ServerRaid rebuild that drive before loading the OS? Or how do I accomplish this?

TIA...

 

[Stiddy]

If you add the same type of drive, it should rebuilt itself. That does not mean it will be bootable. I am sure someone will correct me if I am wrong, but it is my understanding that if you have a RAID 5 with 3 drives and no hotspare then you are dead in the water. You will have to let the drives rebuilt and then perform a restore. Had you had a hotspare, you would still be running

HTH - Stiddy

 

[yert33]

Thanks, Stiddy.
We obtained an identical drive, replaced the defunct one with it and booted up. Of course it's still hanging at the "missing or corrupted file" but the lights on all the drives are solid green and the new drive also has a flashing amber light.

I hope this means it's rebuilding the drive.

BTW - I wish we had done a hotspare way back when. I have had various other RAID 5 arrays tell me there was a drive going bad, replaced it with a hot swap and it rebuilt it on the fly. Was rebootable after that, too. This is the first time that we've had a drive go bad before we noticed it, though. For my sake, I hope you're wrong about the no-boot thang. ;-)

 

[ibmtech65]

You have two things going on, 1st you have some corrupt OS files that needs to be fixed. 2nd you have a failed drive.

Most likely when the drive went defunct the server was in the middle of some sort of write operation to that drive and that corrupted your files.

The only way you are going to get the drive to rebuild correctly is to fix the OS issue so that you can boot the machine completly to the OS. Once you can boot the machine to the OS you would want to replace the drive by removing the failed drive, wait several minutes, then put the new drive in. If you try to do a rebuild outside of the OS it is very good possibility that you will not be able to recover the OS without reloading it.

As far as a raid 5 goes you can lose a drive and still keep on running. That is the whole point of having raid 5. The only advantage of a hotspare is that if you lose a drive it will automatically rebuild and then you just replace the defunct drive and it will become the hotspare.

In theory with raid 5 you should never have to reboot the machine or do anything other then replace the drive, the machine should keep right on running.

 

[yert33]

Thanks for all the replies. Here's how it all got fixed.

We obtained NTFS4DOS from

http://www.bootdisk.com

and booted the problem server up. Very glad to see all the data files on the d: partition. Attempted to copy the system file from another NT$ server onto the C: drive but there was no system32 folder!

Removed the tape drive from this problem server and installed on another box. Restored last good tape found to a CD (6 months old...). Just the WINNT folder.
Put CD drivers on NTFS4DOS diskette and rebooted problem server. Copied system32 folder to problem server. Rebooted....voila!: Ctl-Alt-Del to logon.....

Just had to update a few users and obsolete TCPIP settings on the problem server and we were back in business.

How did the system32 folder disappear?
That is the question...