Re: Disk Corruption: SOLVED
On Tue, 04 Feb 2003 03:20:07 +0100, "Jonathan Brandmeyer"
>What if I am going about this the wrong way? Suppose that MBR(s)
>damaged? Will fsck(8) detect this type of damage? My reading of the
>page says that it doesn't.
ASFAIK .. that is correct. The first sectors of your boot are
proprietry to data for MBR and as such are not a 'filesystem' in
as promised my submission follows.
you need to consider that because I read parts of this thread
@ deb.user ML and (mostly) some in UseNet @ l.d.u its possible to have
missed some of the thread. thus, if i repeat- forgive :-)
I am new to MaiLists and -so far- not enthused as to their
it appears you have read much on this so I am going to assume a few
HOW you set your system up is determined by intended use.
mine boots Win95b(OSR2) by default and a 2.4.18-k6 kernel by choice.
#H0megrown lilo Jan30-2003
#Boot windows by default
#Boot Linux GNU by choice
label = Linux
#end verbose boot script
I had shitloads of trouble getting this preference up and running.
windows rules the MBR so it has to be installed first.
u have to have a method to rewrite the MBR after finding the latest
tweak to LILO is a non-goer. the D0S fdisk /mbr handles this for 95b.
[ its a while back but IIRC i used to use Disk 1 off'n the D0S 6.2 set
to perform the same function on an NT4 drive. maybe that would remain
the same for XP?]
as you can get XP to boot by choice you should be able to organise an
MBR rewriter from here:
the procedure is to config lilo by HAND, run lilo, reboot and at the
LILO prompt type in "linux" for linux to boot or hit return to boot to
WinX or just wait for timeout and WinX will boot by default.
For ANY other result, reboot using the WInX boot disk and rewrite the
MBR, reboot using the linux boot disk and twiddle lilo.conf again.
repeat the process till it works.
dont use any part of the script written by lilo - edit your own.
dont use sym links - point the boot directly at the file/s
dont use "lba32" or "compact" or "delay"
(the first two auto-config anyway.)
dont use an editor in windows to write this - I use/d vi
always run lilo at root before shutdown -r now
I am certain your troubles result from MBR and LILO status/config
misnomers - hence all of the above.
to address the drive partitions/layout you have I can only comment
that on my journey it resolved that D0S demanded it be first in line on
my single drive for the kernel that I was loading at that time.
the 1024 thing is no problem any more so there was no trouble
satisfying the whinge.
in all cases i used PQM (D0S v 6) to manipulate the partitions.
i used PQDI (D0S v4) to move images around
(yes, i *did* make bakups b4 i ventured out)
but a word of warning here.
my troubles began (and what Anita sorted out for me) when I wrote my
backup to a different position on the drive. This totally confused the
table read in linux and thereby fstab.
i couldnt find a way out of this other than to start again - format.
I will be setting up a 2xHDD machine shortly so I will watch with
interest your result. I am not convinced the location of <windows.exe>
or <bootmenu.b> is of any consequence in the mapping of HDD's.
0T0H the equivalent to <msdos.sys> for a winX machine does appear to
like sunning itself on the beach in a dual unix/winX system.
"unstable yet breathing"
$$posted to both l.d.u and firstname.lastname@example.org $$
http://movies.yahoo.com.au - Yahoo! Movies
- What's on at your local cinema?