Re: Salvaging data from a RAID-5 where two disks failed

[Alvin Oga <aoga@ns.Linux-Consulting.com>]

hi ya danie

if you replaced the prev bad disk with a new one..watched for a bit
and went home..

and than it was dead when yougot back ...
and now its complaining about another disk

and now, if you have a 4 drive raid5 setup, you probably s.o.l.

hope you have backup before the drive was going bonkers

and hint... its probably not the drives...

but since you replaced the 1st drive, and used a new one,... the old
disk is probably way out of sync from the current raid setup

too late now ...

c ya
alvin

-- those dma_intr errors are probably not drive problems..
	- just cable problems ( 18" max )
	- if its ibm deskstar... its probably the disk
	- if you changed hdparm options ... ooppss..
	- if you changed your kernel ... ooopppss ...


On Mon, 26 Jan 2004, Danie Roux wrote:

> Hi All,
> 
> I have 4 drives in a RAID-5 configuration.
> 
> One of the drives broke. I replaced said drive, did a 
> 'raidhotadd /dev/md0 /dev/hdb2'. Watched it for a while and then went
> home. 
> 
> And then /dev/hde failed too. The syncing wasn't done:
> 
> Jan 25 19:17:23 clio kernel: hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> Jan 25 19:17:23 clio kernel: hde: dma_intr: error=0x40 { UncorrectableError }, LBAsect=80464841, sector=79214920
> Jan 25 19:17:23 clio kernel: end_request: I/O error, dev 21:03 (hde), sector 79214920
> Jan 25 19:17:23 clio kernel: raid5: Disk failure on ide/host2/bus0/target0/lun0/part3, disabling device. Operation continuing on 2 devices
> Jan 25 19:17:23 clio kernel: md: updating md1 RAID superblock on device
> Jan 25 19:17:23 clio kernel: md: (skipping faulty ide/host2/bus0/target0/lun0/part3 )
> Jan 25 19:17:23 clio kernel: md: md_do_sync() got signal ... exiting
> 
> And then ext3 went bonkers:
> 
> Jan 25 19:29:11 clio kernel: EXT3-fs error (device md(9,1)): ext3_get_inode_loc: unable to read inode block - inode=32997377, block=65994779
> Jan 25 19:29:12 clio kernel: EXT3-fs error (device md(9,1)): ext3_readdir: directory #17055779 contains a hole at offset 0
> Jan 25 19:32:56 clio kernel: EXT3-fs error (device md(9,1)): ext3_get_inode_loc: unable to read inode block - inode=26214401, block=52428827
> 
> Ok, so my question is: 
> 
> Any way in which I can salvage at least a part of the data?
> 
> -- 
> Danie Roux *shuffle* Adore Unix
> 
> 
> 
> -- 
> To UNSUBSCRIBE, email to debian-user-request@lists.debian.org 
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
>