>
> Hi,
>
> I still haven't been able to boot up my FC4 box, beyond repair mode
> without a live distro CD. The attempt gave me errors on /home with
> something along the lines of (from my notes):
> Buffer I/O error on device hdb1
> /dev/hdb1 reaad failed after 0 of 2048 at 0
> No Volume group found
>
> Using the live distro, I was able to determine that all of
> hdba will mount
> (except hdba3 which is swap space) along with hdb2, my /data
> directory.
> All mounted and I could see the files on /dev/hda1 (/boot), /dev/hda2
> (/1), /dev/hda4 (/data1), /dev/sda1 (usb drive) and /dev/hdb2
> (/data).
> The plan was to have OS files on hda and most/all data on hdb
> so that if
> the OS disk got hosed data wouldn't be lost.
>
> I tried "mount /dev/hdb1 -vt ext3 /mnt/hdb1" and got
>
> mount: wrong fs type, bad option, bad superblock on /dev/hdb1,
> missing codepage or other error
> In some cases useful info is found in syslog - try
> dmesg | tail or so
>
> root@2[root]# dmesg | tail
> Buffer I/O error on device hdb1, logical block 7
> hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hdb: dma_intr: error=0x40 { UncorrectableError }, LBAsect=63, high=0,
> low=63, sector=63
> ide: failed opcode was: unknown
> end_request: I/O error, dev hdb, sector 63
> FAT: unable to read boot sector
> hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hdb: dma_intr: error=0x40 { UncorrectableError }, LBAsect=65, high=0,
> low=65, sector=65
> ide: failed opcode was: unknown
> end_request: I/O error, dev hdb, sector 65
> EXT3-fs: unable to read superblock
>
> I can't remember the command I used next, but I wound up getting the
> result of a "bad magic number in superblock"
>
> When I went back to repair mode, the file system start up gives clean
> report on /1, /boot, /data and /data1.
>
> I tried e3fsck, e2fsck with both the -cc and -t options followed by
> similar commands with plain fsck. This last try at fsck -t
> ext3 /dev/hdb1
> gave me "Attempt to read block from file system resulting in
> short read
> while trying to open /dev/hdb1. Could this be a zero-length
> partition?
> I've even tried these commands on /dev/hdb and all I got was a little
> longer wait time before the bad news.
>
> I'm getting to the end of everything I can think of to do.
> My next bright
> idea is to try to change the partition names from /data1 to /home and
> making hdb1 /<something else> to verify that the OS is still
> working and
> perhaps get me back and running (although at a diminished
> capacity) while
> I learn and research. I've even toyed with the idea of
> installing windoze
> on the /data1 drive to see if I could get some sort of
> windoze disk data
> rescue software to retreive what I could of hdb1. Unless
> there has been
> some magic discovery of a software retrival package for linux
> since the
> last time it was discussed on this list.
>
> Can someone please help me save my /home directory or is it
> totally hosed?
>
> Thanks,
>
> Nancy
>
>
I have seen these messages on drives that have gone bad.
hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hdb: dma_intr: error=0x40 { UncorrectableError }, LBAsect=63, high=0,
If you have a good backup, you may be best off reloading. If not,
you might try turning off DMA and seeing if you can get the drive to
read well enought to take a backup. The newest version of SpinRite
supports Linux file systems but it costs money and if the board on the bottom
of the drive has gone bad, it will not help.
If there are no backups and the data is critical and you have an identical
drive available (same firmware issue, same everything), you can try swapping the
board with the good drive and see if that fixes it. Several manufacturers connect
the board to the drive with contacts rather than soldered wires so you can
just unbolt the card and swap it. We did this in one obscure case.
Good Luck
Bob Styma
