[H-GEN] degrees of filesystem sickness?

[Allen Grace]

On Sun, 7 Jun 1998, Brenton Bills wrote:

> On Thu, 4 Jun 1998, Allen wrote:
> 
> > - At boot time, VFS trying to mount the root filesystem readonly causes
> > multiple kernel panics,
> > - mount run from an emergency disk reports "can't find an ext2
> > filesystem on dev 03:42", 
> > - e2fsck run with any option reports that the superblock is corrupted,
> > try another... so I did ("-b 8193", as it suggested) and that block
> > appeared to be also.

In addition, fdisk is "unable to open /dev/hdb2"... partition table
damaged?

> 
> Hmmm same thing happened to me, well.. I was running zgv a graphics file
> viewer, I had just got a new graphics card an (et6100 2megs), I looked at
> the first file, seemed fine, looked at the second file stuffed up the
> display, slowed down did wierd things. I swapped to another vc (could just
> make out a vc window from all the garble) and exectued /sbin/halt a few
> times, I could here the hd clicking but it did not seem to do anything...
> Anyway I ended up pressing reset, it would not boot.... I eventually
> worked out somehow it(zgv or halt, dunno) deled the bootblocks. It has
> also done something similar to the other drive ( a mounted drive which I
> was reading the files from).
> 

What horrors.

> I dont know how the graphics card could have caused so many problems, I
> have been using it since withthatother opsys with no worries. (excuse =
> exams no time for a re-install).

Yes that's my explanation too. 

> Any ideas? 

Not so sure about your graphics card, but I would try to contact the
manufacturers and get some kind of opinion as to whether it's worth your
while trying again. I think most people have some kind of bad experience
when they realise their new nifty bit of hardware is enslaving them to the
evil empire.

As for running real OSes again: seeing as we both have partitions which
are not usable anyway, maybe it's possible to write some kind of protected
mode program to give us an idea of what is actually in the first block,
and then if some kind person gave references to information telling us
what is supposed to be there (An index of some kind, yes, and
presumably a partition table... but how exactly are these set up?) then
we could assemble something and manually (sic) put it in place. Having
a look at the code for mke2fs and fdisk should be revealing. The worst
that could happen would be that there would be even more scrambled egg on
each respective disk, but there's not much practical difference between an
unreadable partition and a more unreadable partition.

But such fun things are going to have to wait for a couple of weeks, no?
Tell me what you think.

allen

--
PD: pardon of my horror syntax english. you must be patient with me.


----------------------- HUMBUG General List --------------------------------
echo "unsubscribe general" | mail majordomo at humbug.org.au # To Unsubscribe