System Doesn't Boot Part 2

Ken Moffat ken at linuxfromscratch.org
Sat Jan 26 17:57:48 PST 2008


On Sat, Jan 26, 2008 at 05:09:07PM -0800, Mark Olbert wrote:
> Ken,
> 
> One additional thing I noticed: if I do a ls -l /dev at the end of the LFS udev script, neither console nor null appear; they're not listed as devices.
> 
> I don't know if this is because they're not in the ramfs that's mounted on top of /dev when udev runs, or if it's related to the problem I'm experiencing, but I thought I'd pass it along because it looked odd to me.
> 
> - Mark
> 
[ insert standard "please don't top-post message" ]

 Hmm, I was going to say that the nodes looked normal, I suppose
their disappearance explains the error message.

 My understanding of udev is that it creates the devices based on
what it finds.  Your original post was for missing disk partitions,
I think ?  I don't know what we used to do in the 6.0 days in the
udev script to start it.  Perhaps we ran udevstart - if that is
anything like what we do, and whichever program we run creates
_some_ devices, you seem to have a _very_ broken system.  I guess
the vast number of pt* and tty* devices in a normal system mean you
won't be able to identify which sort of devices are present, and which
sort are missing.

 After you've identified the program, do you have a backup you can
restore from ?  I guess not, otherwise you wouldn't be asking the
question.

  Actually, I suppose the interesting thing might be the output from
running the udev initscript (perhaps add something after it which
sleeps for a few seconds, so you can see the messages - unless you
can scroll back to them when it halts).  Does the script claim to
end normally ?  I can imagine it might be segfaulting (e.g. corrupted
libc), but not a situation where it creates some of the devices and
then ends normally.

 My gut feeling is that something catastrophic has happened, either
some sort of disk failure, or perhaps a program maliciously replaced
by something else.  Maybe it's only a more common "hardware wears
out" problem (in particular, memory can fail, and power supplies can
go bad), but in that case I would expect you to have problems using
the slackware installation.  A mis-typed 'dd' command could probably
overwrite part of a program or library, but that's the sort of thing
where people usually know they're doing something dangerous.

 I Hope I'm being unnecessarily pessimistic in this post!

 ĸen
-- 
das eine Mal als Tragödie, das andere Mal als Farce



More information about the lfs-support mailing list