On Sun, Sep 09, 2007 at 04:02:35PM -0400, Victor Munoz wrote: > On Sun, Sep 09, 2007 at 11:00:24AM -0700, Andrew Sackville-West wrote: > > > > One thing that might be helpful is what does the BIOS call those > > drives? In your bios setup screen there will be the usual table of > > harddrives and their positions on the motherboard, as the bios sees > > them. Can you provide that info for us? please include how the bios is > > addressing them (LBA etc). > > Bios says this: > > Primary master: HL-DL-ST GCE-8320B [that's the cdrom] > Primary slave: [Auto] > Secondary master: Fujitsu MPF3102AT [this is the 10G Windows disk] > Secondary slave: ST 3160212A [the 160G Debian disk] > > Details for each drive follow: > > Fujitsu: Cylinders 1024 > Head 255 > Sector 63 > CHS Capacity 8422 Mb > Maximum LBA Capacity 10248 Mb > > ST: Cylinders 1024 > Head 255 > Sector 63 > CHS Capacity 8422 Mb > Maximum LBA Capacity 8455 Mb > > Which is strange, since ST is 160G. Fujitsu data (10G), look ok. I don't have a machine available for reboot at the moment, but there are options that can be set for each harddrive. The language is escaping me at the moment, but it has to do with whether the drive is LBA or not etc. You could fiddle with those settings (take notes!) and see if that helps. if its set for some legacy setting (again, the language escapes me ATM), it may pretend it can't address beyond 1024 cylinders when in fact it can. but i think that's a shot in the dark. > > > > > What motherboard is this and how old is it? It looks like you are > > facing an ancient problem with BIOS that couldn't see beyond the first > > 1024 cylinders. With your large / partition, I'd bet those non-working > > kernels are beyond the range that the bios can see, and I'm willing to > > bet that your bios is configured incorrectly to see those higher > > cylinders. > > > > So your suggestion would be to repartition the disk, leaving a small > boot partition at the beginning of the 160G disk? yup. probably up until now you've always had this problem, but the kernels happened to be written within the first 1024 cylinders and thus caused no problem. Also, the same with menu.lst, it was probably within that boundary as well. Then last week, the menu.lst got rewritten to a part of the disk that the bios can't see and suddenly doesn't work. At least that's the way I understand it. Reality may diverge drastically from my perception. ;) A
Attachment:
signature.asc
Description: Digital signature