SEO

ancient Compaq Deskpro w/ 10G hard drive
ROMset : 486Y , flashable, soldered in place
Ethernet NIC embedded on mobo , AMD FlexNet
RedHat 6.0

After 10 years of faithful service (go Tux!) I upgraded from
dialup to cablemodem. Discovered I did not compile ethernet
support into the kernel way back when. Discovered I did not load
kernel source code either way back when. Tried to load source code but
not enough room on hard drive. Grrr

Buy new 80G drive, install as slave, set up same number of partitions
as on old drive just proportionally bigger except for /dev/hda1 which
I kept the same size (10 cyl). mounted new partitions. So far so good.

Began copying from old partition to new partition using
dd if=/dev/hda1 of=/dev/hdb1
dd if=/dev/hda3 of=/dev/hdb2
dd if=/dev/hda5 of=/dev/hdb3
dd if=/dev/hda7 of=/dev/hdb4
and then things started to go horribly wrong. Started getting numerous
messages about 'ext-fserror ext_readdir bad entry in directory' all
pointing to the same inode. System won't respond so switched the
machine off with the Off switch.

Now I can't boot up, LILO starts to load from hard disk but fails with
a 'kernel panic: no init found' message.

Possible side effect of ancient Compaq boxes' BIOS setup is it may
have set the hard disk type to 65 instead of 83

Boot up with Knoppix, which mounts the errant partitions (as seen by
peeking in /etc/fstab) as ext2.

I cd to /mnt/hda1 , /mnt/hda3 /mnt/hda5 and /mnt/hda7 in succession
and all are EMPTY!!! What gives? And more importatly, how can I
recover? There's only one pathtree I =really= need to recover but it'd
be nice to get the old drive back to the way it originally was. Then I
can start the migration process again. Another good question - what
=should= my migration process be?

D-oH!

TIA,

Steve






__________________________________________________ _________
Get your own Web-based E-mail Service at http://www.zzn.com

followup;

while running Knoppix "sudo su" to become root.
looking in /etc/fstab Knoppix found my hard drive partitons.
mounting /dev/hda1 ( /boot ) /dev/hda5 ( /home) and /dev/hda7 ( / )
works.
looking in /mnt/hda1 shows some files including the kernel , good
But looking in /mnt/hda5 and /mnt/hda7 only show one subdir ,
lost+found, and
within this dir (on both partitions) there are no files -- "ls -ltra"
turns up only "." and ".."

Now what?

TIA,

Steve


__________________________________________________ _________
Get your own Web-based E-mail Service at http://www.zzn.com

Anonymous Sender wrote:
>
> Buy new 80G drive, install as slave, set up same number of partitions
> as on old drive just proportionally bigger except for /dev/hda1 which
> I kept the same size (10 cyl). mounted new partitions. So far so good.
>
> Began copying from old partition to new partition using
> dd if=/dev/hda1 of=/dev/hdb1
> dd if=/dev/hda3 of=/dev/hdb2
> dd if=/dev/hda5 of=/dev/hdb3
> dd if=/dev/hda7 of=/dev/hdb4
> and then things started to go horribly wrong. Started getting numerous
> messages about 'ext-fserror ext_readdir bad entry in directory' all
> pointing to the same inode. System won't respond so switched the
> machine off with the Off switch.

When you started copying whole partitions with dd, you were copying them
byte for byte including the filesystem metadata. IOW, your filesystems
on the new disk are now the same size as the ones on the old disk. Your
larger partitions won't have access to the extra space. Moreover, since
you state that you mounted the new partitions, I presume you had already
created filesystems on them, which were overwritten with the smaller
filesystems from the older drives. Then when you shutdown with the
power button, things were probably in a wildly inconsistent state. Who
knows what happened with the partitions on /dev/hdb.


> Another good question - what =should= my migration process be?

See the "Hard Disk Upgrade HOWTO" at http://www.tldp.org/