can't mount ext3 filesystem

[alhambra]

hi,

i have installed a gentoo system on a SAMSUNG 200GB SATAII HD using a ext3 fs. everything worked fine till yesterday, when the booting screen showed me that it cant mount the root fs. (dont know why, becouse i did nothing change the day before)
so i booted from a cd and tried to repair the fs. but without success...
e2fsck gives a lot:

Code: Select all

 Inode 2545 has imagic flag set.  Clear<y>? yes

Inode 2546 is in use, but has dtime set.  Fix<y>? yes

Inode 2546 has imagic flag set.  Clear<y>? yes

Inode 2548 was part of the orphaned inode list.  FIXED.
Inode 2548 has imagic flag set.  Clear<y>? yes

Inode 2549 is in use, but has dtime set.  Fix<y>? yes

Inode 2549 has imagic flag set.  Clear<y>? yes

Inode 2550 is in use, but has dtime set.  Fix<y>? yes

Inode 2553 was part of the orphaned inode list.  FIXED.
Inode 2553 has imagic flag set.  Clear<y>? yes

Inode 2554 is in use, but has dtime set.  Fix<y>? yes

and mke2fs shows

Code: Select all

root@slax:~# mke2fs -n /dev/sda
mke2fs 1.35 (28-Feb-2004)
/dev/sda is entire device, not just one partition!
Proceed anyway? (y,n)
root@slax:~# mke2fs -n /dev/sda1
mke2fs 1.35 (28-Feb-2004)
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
12222464 inodes, 24418792 blocks
1220939 blocks (5.00%) reserved for the super user
First data block=0
746 block groups
32768 blocks per group, 32768 fragments per group
16384 inodes per group
Superblock backups stored on blocks:
        32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
        4096000, 7962624, 11239424, 20480000, 23887872

so i tried do run e2fsck with an alternative superblock

Code: Select all

e2fsck -b 23887872 /dev/sda1
e2fsck 1.35 (28-Feb-2004)
Superblock has a bad ext3 journal (inode 8).
Clear<y>?  

but i stoped because i was afraid to loose more superblocks
is there any chance to repair the system? this would be very very great...

[NeddySeagoon]

alhambra,

Allow it to fix the journal. It will do this by deleting the journal, making the filesystem ext2.
You can add it back later with tune2fs.

Be aware that running e2fsck can make things worse rather than better, so if you have the space, make an image of the filesystem first. Thats the only way you will get an undo.

With the journal fixed, you can risk mounting the filesystem read only, so you can look round. Boot with the liveCD to do that.
Now you can back up some of the more important files before you allow extefsck to attempt to repair the filesystesm.

Read

Code: Select all

man e2fsck

to understand what you are about to do.

[alhambra]

ok, i deleted the journal and tried to mount read only but dmsg shows:

Code: Select all

EXT2-fs: corrupt root inode, run e2fsck

e2fsck comes with endless questons...

btw: to creat a disk image i have to use dd right?

[NeddySeagoon]

alhambra,

dd will allow you to create a disk or partition image.

dd if=/dev/hda ... will read all of /dev/hda
dd if=/dev/hda1 .... will read only the first partition of /dev/hda
and so on.

If you run e2fsck -y ... the -y means answer yes to all questions.
This attempts to blindly fix it.
-n answers no to all questions, so e2fsck exits at the first error.
When -n or -y are missing, it asks you what you want to do.

[alhambra]

i did e2fsck -y /dev/sda1 and now have a full lost+found directory.
i found the files in it witch are important to me . so a great thanks to neddyseagoon...
but i think there will be no way around reinstalling the whole system
could a hardware problem (HD) be the reason for this whole thing?

[NeddySeagoon]

alhambra,

If you have all the important files, a reinstall is probably faster than a salvage operation.
Its unlikely to be a HDD failure or you would get back little or nothing.

A crude test is

Code: Select all

dd if=/dev/sda of=/dev/null

f that completes without error, your drive is probably OK.
It reads all of sda and throws away the data it reads.