hi all i have a cisco call manager that run on linux os i have had a power failure when i have tried to boot the system i got :
checking filesystems
fsck.ext3: invalid argument while checking ext3 journal for /common
/grub: clean, 26/66528 files, 16820/265072 blocks
/partB: clean, 11/3515680 files, 118535/7018396 blocks

dropping you to a shell, the system will reboot


i have tried to recover from recovery cd and i have got:

checking device /dev/dev6
fsck 1.41.4
group descriptors look bad... trying backup blocks...
errors found (8)


any ideas please need help urgent

Suggestion: Since the journal file on /common seems bad, use tune2fs to remove the journal from it (which will lose anything in the journal, but, hopefully, that's survivable), rerun fsck, and recreate the journal file (now empty, of course) with tune2fs.

but why when i am trying to recover i am having bad group descriptors. does that mean if i remove the /common journal the os will work?.

please write for me the parameters of tune2fs which allow me to delete and create the journal

Look at man tune2fs for the complete description.

Basically, to remove the journal file you would use something like this (where /dev/sda9 is my assumed location of the physical, unmounted, device that contains the /common file system - DO NOT USE THIS LOCATION ON YOUR SYSTEM).

tune2fs -fO ^has_journal /dev/sda9

To create a new journal file (assuming the same physical device, and that you want to use the default settings):

tune2fs -j /dev/sda9

BE SURE TO READ THE WARNINGS UNDER THE -f OPTION IN THE MANUAL FILE. N.B: I would recommend that you use the dd command to create an image of the bad partition before trying any recovery methods. That let's you recover to where you are now if things go wrong. (In fact, if you can keep the image on-line on a different drive, you could mount it using mount -o loop,ro and try the recover on the image instead of the problem device.)

As to the OS working, as you know, removing the journal file converts the file system from ext3 to ext2, so everything should work (but see the WARNING referred to above).

my problem now is i am not able to get to single user mode to enter these commands.
when i press 'e' or 'a' to enter single user mode nothing happen like i didn't press anything the only key which work is the enter key to continue booting

please assist

You mentioned in your first post that you were booting from the rescue CD. If this is what you're doing, why would you need to edit the GRUB menu to get into single user mode? The rescue disk should automatically boot you into init level 3 with you drives unmounted, and let you run fsck (or tun2fs) as you wish.

What file system do you have on /dev/sda6? (I had assumed that it was /common because that was where it first error was reported.)

Are you using Logical Volume Management on your drives? I ask because fsck will completely destroy all logical volumes on a partition used by LVM if it is run on the raw partition. You can only run fsck on the logical volumes listed under /dev/mapper/Volume_group_name/Logical_Volume_Name after the logical volume is activated.

Are you sure that the "Enter" key actually worked? The boot may have continued because the timeout period expired.

I've only had the problem you describe when my keyboard cable was disconnected.

If your keyboard is connected, try pressing the "Escape" key (or Ctrl-[) to see if that gets GRUB's attention.

hey - have EXACT same issue, almost a year later on a cucm 6.1.3.

i boot to the recovery cd, go to user mode

when i run

./tune2fs -fO ^has_journal /dev/sda6

i get..

invalid argument when reading journal inode

did you ever find a solution to this?