Baslinux mailing list

[baslinux AT lists.ibiblio.org]

On Mon, 28 Apr 2008, baslinux AT lists.ibiblio.org wrote:

> Slackware 7.1 uses lilo 21.4 so I do not need to compile it.

I was unable to boot with LILO when linux was not in the first 1024 
cylinders, from a suspect drive and suspect computer.  LI and it stopped.

To fix the 1024 cylinder limit problems, I used Slackware 7.1 lilo and 
typed at the top of lilo.conf

lba32

then copied all of /boot to /hda1/bl2 and changed /boot to /hda1/bl2 in 
lilo.conf, so these would be in the first 1024 cylinders.

Using lilo 21.4, I typed lilo -v on the 100GB drive (which appears to have 
a bad FAT table) and it made me a lilo boot floppy, which when I tried to 
boot with it using the 100MB drive gave me a FATAL error.  I tried this in 
the suspect computer and in one other computer that seems perfect. 
Without lba32 it gave me LI and stopped.

I did the same thing on a 60GB drive which has no obvious problems other 
than some bad sectors, and it booted directly to BL2 (my default) without 
giving me a prompt to type in DOS or BL35 or BL34 on the second computer, 
but booting stopped at LI on the suspect one.

So LI appears to mean I cannot boot with BL3 lilo from beyond cylinder 
1024 and/or the computer (ide controller?) is bad, and the FATAL error 
means the 100GB drive is bad.

PQMagic 5 for DOS crashes with this drive in any computer.  Linux has no 
problems with it - can I format it to FAT32 with linux after partitioning, 
or is the partition table worn out and if so, why only for DOS partitions 
(three of them, 500MB each)?

It is thus possible to boot linux when linux is not in the first 1024 
cylinders, by putting the boot files there and changing /boot to /hda1/bl2 
in lilo.conf, and using SW71 or later lilo, with lba32 in lilo.conf.

I tried repartioning the 100GB drive and linux is happy with it, but I get 
all sorts of problems with the DOS partitions - won't format, or formats 
but I get gibberish, or files reappear after I reformat then disappear. 
The drive occasionally booted in another computer.  So I suspect one
'drive geometry' problem here is in the hard disk.

When I copied a large file to it from flash drive, it kept crashing badly 
in the faster computer, with ide0 errors, so the ide controller is 
probably also bad.  It had no such errors in two other computers. 
Another drive also had similar errors in this computer.  Whatever killed 
the cpu maybe also affected the controller.  It may work with live CD. 
How else might I test the controller in linux other than by copying large 
files?

Another 'fast' computer works with HP 3p but not newer scanners, something 
about synchronous vs asynchronous that causes the new ones to always scan 
at 100 dpi.  And yet another (Micron 500MHz) works with Xvesa but not 
svgalib (any card).  Maybe some laptops just won't work with some wireless 
cards.

Sindi