RAID 1 Disk Failure has killed my system!

[depaulo]

Hi Can anyone help?
I have a raid 1 setup with 2 disks using Intel raid controller built into the mother board. One disk has gone wrong (at bootup you can see 1 disk says 'Error Occurd' and the other say 'Normal', and the array is in state 'Degraded'). This has had the effect that my system (wk2k3 server) won't boot as it can't find boot.ini etc. I have tried going into the recovery console and it wont even do a dir on C:\. It says 'unable to iterate directories' or some such.
As i dont understand RAID that much, could it be that maybe the second disk was never being used? or that the boot partition was on the first so now it can't boot? Any help appreciated.....

 

[Daveyd123]

The first place I would start is in the RAID BIOS. Read your motherboard manual to see how to access it and how to repair a faied mirrored array.

 

[wahnula]

Another thing to try would be to disconnect each drive and try booting from the other.

If that does not work, rebuild the array and recover from backup.

 

[depaulo]

I have tried unplugging each disk in turn but no joy. As for backup, i stupidly thought raid could not fail me and have been rather slack lately. shame on me.

 

[JGALEY]

Maybe I am missing something, but your computer should not care if one of your RAID 1 disks is down - they are copies of each other, and the server should just see them as drive C.

So your other hard drive should have the boot.ini, and your server should automatically use that. If all that is wrong is one bad drive, you should be able to operate as normal, and hopefully even rebuild your array while staying online, if your RAID hardware/software allows it. Dell Poweredge servers allow me to rebuild degraded arrays inside Windows, for example.

If you cannot boot, you might have another problem, like two bad drives (not that likely, imo) or a bad controller, or maybe a bad backlplane or something else like that.

Of course, you could have file corruption that got put on both drives, which would require a restore.

HTH,


Jason

http://www.catotek.com

 

[depaulo]

Thats exactly what I thought. I think they must both be knackered even though the bios says the second is ok. I have been kind of clutching at straws thinking maybe I had to do something to get it booting using the second. Thanks anyway.

 

[wahnula]

Hopefully not too much data has been added since your last backup. If never, then perhaps forensic recovery of the data is in order. If that is the case you should stop using the drives immediately. It is an expensive option but how valuable is your data on the array?

I agree look into the RAID BIOS and try to effect a repair/rebuild. Since the array is 'degraded' it may need to rebuild (speculation here as I am not familiar w/ that controller) before it can be used. Jason is correct that it should not matter if one drive is bad.

My experience w/ RAID1 is if I remove/replace a drive it starts to rebuild automatically and still boots into Windows. This reinforces the 'corrupt data on both drives' scenario.

Wish I could be more helpful.

 

[SteveMCSE]

If the RAID 1 is Windows RAID (not hardware raid) you will have to either move the good drive to slot 0 (the first slot on the controller) or modify the boot.ini file to look to disk 1 (slot 1).

I have plenty of experience with those crappy IOmega NAS 300m servers. Servers!? HA! More like frizbies! When the mirror fails, (when not if - they are junk servers!) you have to modify the boot.ini file. It's Windows RAID so it has to be done.

Look into this option. Although - it sounds like you are using hardware raid (better way to go imho).



Steve
Systems Engineer

 

[depaulo]

I think I must be using hardware raid (you enable it in the bios) though i do have to install a windows raid driver so windows setup can see(ish) the disk. Maybe i will try fixmbr and fixboot and see if that brings it to life. First i will by a new hard disk and see if it will rebuild. Thanks for the help.

 

[wahnula]

I just thought of another option...Google "Bart PE" and make a bootable Windows CD. This will allow file recovery to a backup drive or flash drive. There are also Nero drivers you can install to burn CD's w/ BartPE.

Once you have recovered your data, feel free to play away with the disks. My recommendation would be a fresh install of Windows AFTER configuring your RAID1 array in RAID BIOS. You will probably need to F6 the RAID drivers from floppy during boot from CD.

 

[depaulo]

OK, got it sorted. Thought i'd best post what happened.
Removed the drive which said 'Error Occurrd' in the bios (rather than the one that said 'Normal'. The one that said 'Normal' was useless) and plugged it as slave into another pc.
Downloaded PC Inspector and tried that on it; no joy, just garbage returned.
Downloaded Getdataback NTFS and ran that over the disk and lo and behold there were all my files!!!! copied off and i am a very happy man.
In case anyone is googling this, the exact message I was getting from recovery console when doing dir was: An error occurred during directory enumeration.
Many thanks everyone.

 

[SteveMCSE]

http://www.grc.com

has a GREAT tool called SpinRite 6.0. It's used for testing/detecting/correcting hard drives and errors. It's an 89.00 tool, but it works really well!

Steve
Systems Engineer

 

[wahnula]

Great news!

Now, about that backup plan...I use (2)80GB mini-USB 2.0 laptop hard drives (less than $100 each) and SyncBack SE ($24) as my data backup plan. I swap drives occasionaly so I have a recent copy that stays in my backpack for offsite media.

I know techies like robust software like Veritas, but for a small-office scenario my plan is better. Veritas (and similar BU s/w) compresses and saves your data in a proprietary format, requiring a working install of Veritas to recover.

SyncBack actually copies the entire file & format to backup, recoverable from any PC without special software. If your server is stolen, you can connect the drive to a client, re-name or map it, and be up again in minutes.

This dual-backup plan has saved my bacon more than once, the first time is when a departing employee deleted all her files from the server on a Sunday, I did not discover this until mid-week, last week's drive had all the missing files. They were promptly replaced. Another instance was a corrupted .xlb file on a client corrupted a file on the server, again the BU drive replaced the file.

For archiving, I spin off a few DVDs of pertinent files every month.

I hope you take this event to heart and back up that data!!!