RAID 5 configuration on Dell PowerEdge 2800 hard disk failure
 

hi friends,

Using DELL Poweredge 2800 with a Raid 5 configuration of 2 hard drives capacity 85GB
Ultra 320 SCSI. Hard disks have been showing failure.

I tried to bring the hard drives ONLINE by FORCE option provided in the utility manager to configure RAID and got the hard drives up but couldn't login to windows server 2003. It started showing NTLDR and NTDETECT missing messages. I also copied the files using recovery console with Windows Server installation CD but still no use. I still can not login to Windows with the message ntoskrnl.exe is missing or corrupt.

Can i get to the data without a backup?

............................................

Your best chance would probably be to use a live cd like Ubuntu or something that supports your RAID controller, and at the same time having your backup drive connected directly to the motherboard and not the RAID controller. Using the live CD you could then copy the files accross without worry that the OS you are in is corrupt in any manner.

But I'd double check your RAID configuration as it is. I'd assume you've already had a failure since RAID5 requires a minimum of 3 drives to run/be configured at the beginning, and if it's reporting 2... sounds like a failure already.

Issues with NTLDR, NTOSKRNL, etc, generally are symptom of corrupting partitions/drives or (obviously) failures. If you keep track of which ones specifically it is complaining about, it could be possible to fix the issues using the recovery console in the Windows Install disks. The repair would basically be replacing the corrupted files with the originals from the disk and you should be able to boot and relogin to the system.
Here is the Microsoft support page on how to replace files like ntoskrnl. While it says Windows XP, it still applies to Win2003.

Can I just chip in another question?

It it shows a failure does not it mean that the data are already corrupted?

==Using the existing example above==
That's where varying RAID controllers have different definitions regarding failures and the words they display related. From a raw point of view, independent of the RAID controller's chosen context, you'd be looking at something like this:

3 drive RAID5, offers 1 drive redundancy(All drives good means array OK!)
1 of 3 fail, no data loss should occur, no corruption should either(Array Degraded)
2 of 3 drives fail, technically the system shouldn't be bootable since RAID5 is built on a more complicated version of 3 + 1 = 4. You can determine the third number using any two knowns, be it half the data and a checksum/hash or both halfs of the data and not the checksum/hash. In this state data corruption should have already occurred. This could be identified several ways, Array Degraded, Array Failure, Array Dead, Array Missing.

What your more possibly looking at is partition corruption on top of a drive failure. Partition corruption can occur due to outside causes as well as just from a software point of view.

If you can still see the file system on your RAID configuration you can boot to a boot CD and see the files and move the files to another drive. You can use UBCD4Win to create a boot CD.

Just a side note, RAID 5 is not designed for anything less than 3 drives, so your best bet for redudancy using 2 drives is RAID 1.

Thank you for the##informative##posts.

Thank you for the##informative##posts.