From: SoloCDM (deedsmis at domain aculink.net)
Date: Sun 07 Oct 2001 - 11:14:10 IST
\\x2fBrian O'D stated the following:
>
> It sounds like when fsck comes along to fsck you're root partition it's
> already mouted right?
>
> hrmmmmm......., If you go into /etc/init.d there should be some kind of
> script in that mandrakee area yes? Ok I guess you could try grep -l -r
> -s fsck * and try to find the explicit mount before the fsck and negate
> it ie cut it out of where it is in the script and move it to after the
> fsck, though that may be really dumb. Yeah if memory serves then about
> ohhh wait a moment just before you're fsck add this to the script that
> is running fsck yeah?
>
> umount /dev/hdb1
>
> if Mandrake remounts the partition then all well and good yeah? If not
> then right after the fsck add
> mount /dev/hdb1 /
> presumably it's root or /usr or something right?
> yeah surely if you just find the fsck somewhere in etc all you will have
> to do is add the line
> umount /dev/hdb1 to the same script as is running the fsck right?
>
> Che,lives
>
> >I haven't had any problem with Kernel 2.2.19-10mdk on LM8.0 booting,
> >until now, after adding the most recent updates.
> >
> >Every time I boot the system, the following error occurs:
> >
> > Checking root filesystem /dev/hdb1 is mounted. Cannot
> > continue, aborting.
> >
> > Failed to check filesystem.
> > Do you want to repair the errors?
> > (beware, you can loose data)
> >
> >
> >How can this issue be repaired?
I found the culprit, but it still doesn't explain the lockout, unless
(unknown to me) Linux has a special safety feature or this was a freak
accident. (Note: because the system showed every sign that it locked
up during the crisis . . . I used the cold reboot reset button on the
system hardware unit.)
I have a script that runs a monthly scan over the entire system that
requires the use of the command find, but it unleashed over 200
find-s. Sometime back, I made a minor change in the script without
testing it, because it seemed at the time that it was in check with
every known parameters. Obviously not! The script is fixed, but not
the reasons for Linux's behavior.
I checked into umount and found that it is capable of umount-ing any
item without writing the results to /etc/mtab by invoking -n. No part
of the before mentioned script included this execution!?!?
I am not entirely confident that if I had deleted /etc/mtab . . . the
system would have returned to normal. I believe something in /proc or
/var contributed to this issue, but what I don't know.
--
Note: When you reply to this message, please include the mailing
list/newsgroup address and my email address in To:.
*********************************************************************
Signed,
SoloCDM
This archive was generated by hypermail 2.1.6 : Thu 06 Feb 2003 - 13:12:35 GMT