problem in Linux Startup

[Muhammad Shamroze]

Hello, Im using RedHat AS 4 on DL380 PRoliant server and Cluster Application is running on it as the server was abnormal shutdown when i startup the server it is showing following problem that is

"Uncompressing Linux....ok, booting the kernel.
audit(1154959027.963:0):initialized
Redhat nash version 4.1.18 starting
WARNING: can't access (null)
exec of init ((null)) failed!!!: 14
unmount /initrd/dev failed:2
Kernel panic-not syncing: Attempted to kill init!"

kindly help me out in this regard so ill be thankful..waiting for ur reply....

[g33k]

it seem as problem with ramdisk, initrd.

unfortunatly I'm not such expert...

here is some documentation which may help

http://www.tldp.org/LDP/Linux-Filesystem-Hierarchy/html/initrd.html
http://www.faqs.org/docs/evms/x3834.html

anyway if it's not necesery to have this distribution I would try anotherone

[Muhammad Shamroze]

well thanks for replying as I boot from rescue CD and when shell prompt occurs what i did i first check all the partitions and label there were all fine but when i mount my / partition and list the filesystem by typing ls -al and i found that in etc and root it was showing as follows

"? ?? ?? ?? etc"
"? ?? ?? ?? root"

and all the rest fs are fine

kindly help me in this regards...thankyou

[Manuel Wolfshant · Accepted solution]

boot from a rescue disk (RH first disk, pass rescue as argument on the first start line -- use F1/F2/F3 for help details) and run a file system check. it looks like your root partition has errors.

[JBR]

Hi, you can try mount de initrd image after booting from RH CD # 1 (linux rescue as option) and mount the boot partition, after this :

# cp /boot/initrd-.img /tmp/initrd-.gz

#mkdir initrd
#cd initrd
#cpio -cid -I ../initrd-

#cat init (this script execute many instructions to boot) and verify the instructions in this script.

Best Regards...

[Muhammad Shamroze]

Hello, thank for your reply as i have already replied in my question that i have already check through rescue disk and when i mounted the main partition that is / then i take the listing of / partition and found that two directories one etc and another one root both were corrupted i mean to say when i take a listing by typing "ls -al" for the / partition i found these two directories as follows

?? ?? ?? ? ? ? ? etc

?? ?? ?? ? ? ? ? root

so now my question is i want that these two files in their orignal position as in etc the are devices and while etc is crrupted it is not mounted those fs through fstab and there is in initrd it is showing errors so kindly help me out in this regards as i also dont have any backup .

[Manuel Wolfshant]

usually the /root directory does not contain much ... unless you have explicitely placed something over there. so removing that folder and recreating it should not be painful.

/etc on the other hand IS painful. ALL configuration data is over there and without backup.... The next step depends very much on the file system you have chosen for the root partition. If it is ext3 (RH default option), I suggest to run fsck using the "alternate superblock" (-b) option. For proper usage of this tool it is mandatory to read the fsck.ext3 man page.
In my opinion however, I think it would take less time to copy the /etc directory from another computer and to carefullly adjust it to your needs. I for one would go this way. The reason is that, even if you recover the lost info from the original /etc, you will still need to verify each and every recovered file.