topic Re: Hard disk failed on rx4640 in Operating System - Linux

Hard disk failed on rx4640

Di Xu — Sat, 14 Nov 2009 09:32:19 GMT

Hi all,
I am using a rx4640 box. The disks failed after a power down. Now it print error messages after loading the linux kernel which is like:
mount error 15 mounting ext3
mount error 2 mounting none
switchroot mount failed 22
umount /initrd/dev/ ...

I tried mount each disk on another box (rx2620), and using the command
fdisk -l /dev/hda
then it reported that the partition table is invalid. However, it can recognize the correct disk size and sectors.

What should I do to save the data on these disks? Dose this problem related with the smart array controller?

Re: Hard disk failed on rx4640

Di Xu — Sat, 14 Nov 2009 11:49:04 GMT

Now I realize that the two disks are organized as a disk array, so it may be normal for fdisk not recognize the partition table for each single one. Am I right?

If so, will the HP SmartStart be useful to fix the disks without destroy the data?

Re: Hard disk failed on rx4640

Matti_Kurkela — Sat, 14 Nov 2009 16:09:08 GMT

> Now I realize that the two disks are organized as a disk array, so it may be normal for fdisk not recognize the partition table for each single one. Am I right?

It depends on which RAID mode was used. RAID 1 disks should be accessible even alone, but if RAID 0 was used and one of the pair is damaged, all data is effectively lost.

In addition, these servers use EFI GPT partition table which is not supported by Linux's traditional fdisk command. Use the "parted" command instead.

> If so, will the HP SmartStart be useful to fix the disks without destroy the data?

I don't think so. It might be more useful to use boot the system from the installation CD/DVD of the original Linux distribution, and activate a rescue mode.

This would give you a Linux command prompt with necessary tools for fixing the filesystem manually. The first thing would be to examine /proc/partitions - can the system detect any partitions at all on the damaged disk/RAID set?

The error messages you listed in your original post would seem to indicate that the root partition is not accessible. How was the disk originally partitioned? If there are several other partitions, it might be possible to rescue data from them even if the root partition is completely lost. But if the disk is set up as one big partition, severe damage to the metadata of the partition may cause the loss of all data.

You may just have had a harsh practical lesson on the subject of "Why the backups are important."

MK

Re: Hard disk failed on rx4640

Michael Steele_2 — Sun, 15 Nov 2009 02:14:49 GMT

Hi

Any time you want to test a disk, internal or RAID, you dd it.

dd in=/dev/dsk/c1t2d4s4 of=/dev/null count=1000

If it hangs then its bad.

If it returns something like

1000 records in
1001 records out

Then its bad

The correct response is

1000 Records in
1000 records out

Re: Hard disk failed on rx4640

Di Xu — Wed, 18 Nov 2009 05:23:47 GMT

Thank you all!
Now it working well.

"You may just have had a harsh practical lesson on the subject of 'Why the backups are important.'"

Absolutely.

Re: Hard disk failed on rx4640

Di Xu — Wed, 18 Nov 2009 05:27:47 GMT

Boot system with Redhat install CD.
Rescue data.