For the second time in three days I now get the following error message when booting (this is from failsafe boot, but the others produce similar errors, and then stall so that I can't scroll up anymore):

JBD: Failed to read block at offset 65
EXT3-fs: error loading journal
mount: error 22 mounting ext3 flags Kernel panic: No init found. Try passing init=option to kernel

The other error messages also refer to a missing journal and that block 65. But the advice in the last line isn't exactly helpful - what am I supposed to do?

The last time I low-level formatted the disc after nothing else worked (booting CD rescue mode etc.) and installed everything anew - after all I am still learning linux and not doing anything sensible which I could lose. But now, getting the same problem again after two days, I'd like to know whether there is a less time-consuming option.

Thanks in advance.

Robin

cause i've seen this error in this particular instance, did you by chance recompile your kernel?

Not at all, at least not knowingly. In this new installation I just installed mc and a couple of other packages and updates (but no updates the first time). Both times the system was shut down cleanly before the fatal boot.

This time around for some reason the mount points were no longer working in KDE before I rebooted the system (to see whether they would come up ok again); meaning when I clicked the CD ROM icon for one of my drives instead of showing me the disk content the dialogue asking for which software to use to open the file in question showed up. Not so the first time though, there were no problems at all.

Robin

just skimming through the post, forgive me if i mention something already discussed....

i have had a similar "kernal panic" problem, and solved it by changing the symbolic links in my grub.conf (it is probably lilo.something for u) to actually point to my vmlinuz and initrd files

if u can boot up in rescue mode through mandrake cd, cd to /boot and take a look at the entries for vmlinuz and initrd, write down the most recent kernel entry for those (or kernel, u should only have 1 version)

edit your grub.conf or lilo.conf in the entries that say just vmlinuz and initrd and replace them with the full filenames u wrote down

hopefully someone can help u with lilo, cuz i hear u have to run something after editing lilo

hope everything works

also, if all else fails, do a completely brand new install but install only the default components, u can then install other things later

Thanks, I've tried that - but there is no (whatever).conf file in /boot - the other files are there though. But when I go into /boot (which I can only do after mounting the drive using the rescue menu, and that doesn't always work) after entering a couple of commands like ls -a etc. I get a whole screen of error messages, after which the system does not respond to commands anymore - I can still type them but all it does is a line feed.

The error (typos possible, I wrote it down by hand):

cd.Assertion failure in __journal_remove_journal_head() at journal.c:1863 : "buffer_jbd(bh)"
kernel BUG at journal.c:1863!
invalid operand: 0000
(here follows a list of what I take to be drivers)
CPU: 0
(and a list of unintelligible codes, ending in a line of two digit hex codes)

Ctrl+Alt+Del still shuts down the system after that, but actual commands are not accepted (though scrolling with Shift+PgUp still works)

Regarding a re-install: is it possible to install just the basic system without hurting the previously installed programs? I'm afraid if I walk through the installation process again and don't check any software presets it will uninstall what is there... Which would be a nuisance rather than a problem, but nevertheless I'd rather solve this problem more elegantly, especially as it has cropped up twice this week, and who know how often it will in future... I can't very well use the system if I have to reinstall every other day.

Robin

I agree with TheOneAndOnlySM. When you see "try passing XXX to init", it is probably due to some mistake in your grub.conf (or LILO). But also, I use Grub instead of LILO.

/**** the following assume you installed grub****/
If you installed Grub, you can find grub.conf in /boot/grub There are three lines important. For example, mine is:
##############
root(hd0,6)
kernel /vmlinuz-2.4.20-8 ro root=LABEL=/
initrd /initrd-2.4.20-8.img
##############
The first root(hd0,6) tells linux booting device is the 6th partition of my harddrive. Second kernel /XXXX is where your kernel resides, then root=LABEL=/ is the root directory of my linux The third initrd is also important. Write the right name and position of your initrd file, usually it is in / If you are not sure where are those files, try to press "c" when grub pops up, one the command line grub>find /vmlinuz (also grub>find /initrd).
In any case you should be able to boot from the boot floppy.

I tried booting from the boot floppy, but that failed as well. In any case neithter in /boot nor in /grub there was any *.conf file, so I guess important parts of the system have been destroyed.

I have reformatted the disc and installed everything yet again, but this time, after various unsuccessful attempts at installing to a newly low-level formatted harddrive without any errors during install, I switched my two ATAPI drives so that the DVD ROM was the boot device rather than the CD recorder (I've also got a normal Plextor drive, but it is SCSI and I can't boot from it).

Setup ran without any problems on that DVD drive; it had done so a couple of times on the recorder, too, but I suspect the system might not tell me about every error. In any case setup ran smoother and even the packages I installed later did so without any problems; I have rebooted a couple of times since and so far everything is up and running. So if this installation survives the weekend I'll be optimistic

But thanks anyway for the help. I have copied it in case everything crashes again anyway.

Robin

hi.
i've been facing this problem ever since i installed
a local distribution (actually, enhanced redhat 8.0)
my kernel version is 2.4.18_24.8.0
initially, i blamed my aging hardware, now after hving
found several users reporting similar problem, i feel
it cud be a bug in ext3 journaling / a problem with my old hard disk
or both. what is happening to me is even if i cleanly
shutdown my system, sometimes on subsequent boot up
i get messages like "/ not cleanly unmounted, check forced.."
etc. and in the worst case it doesn't boot and halts with
"kernel panic, no init found, try passing " kind of message.
if i boot my system at this point using knoppix ( a CD based
linux distribution) and try to mount my linux partition, it fails.
if i run fsck.ext3 kind of tool to verify the file system's integrity,
i find a hell lot of error messages, missing/misplaced/corrupted
inodes etc. once after fsck, the system started booting okay.
but, yesterday it didn't and when i checked, there was no copy
of "init" left anywhere on my harddisk.
wud like to know what u guys think of this.
thanks.

Well, for me it works now; probably the original boot drive did not read the data cleanly and there were errors right after installation already. Linux has been working perfectly since I switched those drives with no more such problems.

Robin