LILO booting off second IDE disk 1100

Java Jive wrote a very detailed response to my plea for help, including this crucial snippet:

And that's a critical little bit of information I didn't know; and that was my problem.

By simply replacing:

boot =dev-hdb1

with

boot =dev-hdb

LILO replaced the MBR instead of the PBR, and everything worked exactly as expected (the bios settings for hda-hdb were also swapped, but that's as expected and what I'd done before).

So, thank you, Mr. Jive!!

I guess my Microsoft MBR was booting off hdb, booting the PBR off *hda*, as you say, which also had LILO; and when that lilo ran its second-stage loader it tried to retrieve its mapped sectors off device 0x80 ... which was hdb at the time. Hence the problem.

All working now!!

I wonder how or why the Microsoft MBR does that; it doesn't use the bios and just ask for 'disk 0x80', like the rest of the boot process seems to do?

I found your three booting configurations interesting, and much along the lines of what I've done before, particularly your first:

Interesting! In the bad old days of geometric addressing I've had to have some Linux-Lilo configurations refer to map files and kernels placed in DOS or Windows parbreastions placed at the beginning of the disk, under the 1024 cylinder limit, but I've never placed the maps on drives separate from the rest ofboot.

Thanks for your help; I do appreciate the time you put into your reply; you fixed my problem for me. And my apologies to my other respondents, some of whom I think suggested 'install LILO on the MBR of the second disk', to which I only stubbornly replied 'but the MBR isn't the problem, I know it's executing and running LILO'. Sigh. As it turned out, it was running the *wrong* LILO! I had no bloody idea the MS MBR had that little quirk.

Thanks all,

(booting beautifully) Brad