Well, shit finally happened ... and I've been running Linux since 1.0.9
1- Background: Kernel 2.0.0, Slackware 3.3
I had the system all tuned up and running smoothly on a Pentium
100 MHz (Award BIOS). I have three IDE disks installed:
- hda (most recent) Fujistu. 1.75 Gig. CHS: 847/64/63 (LBA). Made the
/dev/hda1 partition for Win95 (using fdisk in DOS, cylinders
1-491). Installed Win95. No problems. I left cylinders 492-847
available for a future /dev/hda2 (ext2 file system).
- hdb (oldest). Small Maxtor. /dev/hdb1 is my system disk. /dev/hdb2
is 16M swap
- hdc. Large Maxtor. /dev/hdc1 and /dev/hdc2 are both ext2 file systems.
These are my Linux "work" file systems. I also have hc3 and hc4 as 16
Meg swap partitions.
- hdb is my CD-ROM drive
I had these three disks recognized by LILO w/o a hitch. LILO would boot
the system like a charm. The only one specifically mentioned in lilo.conf
is /dev/hdc. I have this entry there:
disk = /dev/hdc
My Linux boot disk is /dev/hda (boot=/dev/hda). So far so good.
2- The Shit That Happened
I decided to upgrade my internal modem. It seems (I'm not sure), that when
I tried to remove my old modem (which is sitting very very close to the BIOS
battery) or when I installed the new one, I may have accidentally touched the
battery, or at least something happened ... When I tried to re-boot the system
with the new modem in place I got a * message from the BIOS saying that it
could not find anything and that it was loading the default values. Uhg!! Of
course, I failed to keep a hard copy of all my BIOS settings, so I spent an
entire afternoon re-building the BIOS. I THINK I got it to the way it was.
But I'm note 100% sure. Read on ...
LILO now won't work. I've tried every possible combination for the disks
parameters. I did keep a hard copy of my Linux boot-up messages (using "dmesg",
re-directing the output to a file), so I'm pretty sure the disks settings are
correct. One disturbing things is, however, that /dev/hda shows "0kB Cache"
(the other two disks have 128kB). When I boot up, I get:
L 80 80 80
I looked into LILO docs and it says that when LILO only gets to the first
stage it means there is a disk geometry mismatch, or a media failure. The
80 means "Disk Timeout": Drive isn't ready. Either the disk is bad, or is
3- Additional Details:
- I'm able to boot up Linux from a floppy and when I do, all disks are
recognized by the kernel (that's when I saw the "0kB Cache" message).
- The /dev/hda seems ok. I removed LILO and I can work under Windows 95
(apparently) w/o any problems. Win95 boots up normally.
- More Shit: I booted up Linux with a floppy. Run "fdisk" to verify
my partition table for /dev/hda: 1-491 shows as a DOS system. OK. I then
created a new partition, /dev/hda2, as a ext2 file system. No problem.
hda1 1-491 (cyls) 989M (DOS)
hda2 492-846 (cyls) 717M (ext2)
Total: 847 cylinders of 4032 x 512 bytes
AH! But I then tried to format this file system:
mke2fs would not work if I didn't include the amount of blocksQuote:> mke2fs -c /dev/hda2
So then I tried specifying the exact number of blocks:
It took about 4 hours to check all the blocks. But at the endQuote:> mke2fs -c /dev/hda2 717696
I got this:
Bad block0 ouf of range; ignored
Block1 in primary superblock/group descriptor area bad
Block1 through 5 must be good in order to build a file system
So, after all this, I'm at a loss. The system worked fine and
now LILO doesn't boot, the disk works OK under W95, but there
seem to be somehting wrong. ANY help on this matter would be
appreciated. Even if it's a pointer to even further documentation.
Thanks you for "listening" and your help.