Re: Computer unusable -- can't get lilo to work
Just wanted to say that I've solved my lilo problems. I hooked up an old
disk with debian on it and used the lilo on that. Now my new IBM hard
drive boots properly (when made the third drive).
I am still curious about why it the partition table of the IBM drive is
not recognized when made the first drive. But this is more a matter of
curiosity than anything --- the main thing is that I now have a working
debian installation again!
On Mon, 16 Oct 2000, Mark Phillips wrote:
> My computer is currently in an unusable state. I am borrowing my Dad's
> computer to write this!
> The story in summary form:
> 1. Purchased a new 15Gig IBM ATA/100 hard drive. Hooked it up as
> 1st drive. BIOS detects okay.
> 2. fdisk complains that there is no known partition table on it. I get
> fdisk to create one, as well as partitions, and write to disk. All is
> well again.
> 3. Many trials and tribulations later I get Debian Linux (potato)
> successfully installed on new drive. Using custom kernel
> 2.4.0-test9 --- upgraded a few packages to woody in line with this.
> (I needed 2.4 for LVM and ReiserFS.) Created boot floppy for it which
> successfully booted it several times.
> 4. I actually wanted the new drive as the third disk, so I swapped disks
> and cables around. Got BIOS to re-detect the drives, which it did okay.
> 5. Used Debian rescue disk to boot. I hoped to use the rescue disk to
> run lilo from in order to make the new drive bootable again. But lilo
> failed. The rescue disk seems to use the lilo executable on the IBM
> drive, and this failed. It gave the error message:
> /target/sbin/lilo: /lib/ld-linux.so.2: version `GLIBC_2.2'
> not found (required by /target/lib/libc.so.6)
> I'm not sure what the problem is here. It could be because on the IBM
> drive installation (/target) I upgraded a few packages to woody. I could
> be because /target did not have mounted /usr or /var (and I can't mount
> them without the 2.4 kernel). Anyway, whatever the reason, I can't get
> lilo to work so I can't boot the partition. I can't even use a
> commandline option to boot from the rescue disk with /dev/hdc1 as root,
> because again, I need my custom 2.4 kernel so I can get LVM and ReiserFS
> stuff so that /usr and /var will be mounted.
> 6. I swapped cables around so that my new IBM drive became the first drive
> again, and got the BIOS to re-detect the disks. This way I hoped to use
> my custom boot disk for booting the new drive, which I created when it was
> /dev/hda. The kernel started booting okay with this, but it couldn't find
> a partition table on the IBM drive! "Unknown partition table" the kernel
> complained. This is the same problem it had at the very beginning of this
> story. I am not about to use fdisk to recreate a partition table and wipe
> my disk.
> 7. I swapped the cables over again a few times and it seems the partition
> table is readable when I make the new drive /dev/hdc but unreadable when I
> make it /dev/hda.
> So that it the current state of play. I can't think what the problem is
> with the partition table being unreadable. I suspect it might be some
> problem with the motherboard. In which case the best thing to do is to
> stick with /dev/hdc. But how can I get it bootable as this? How can I
> get a lilo which will work and set up the boot sector properly?
> Please cc any replies directly to me.
> "They told me I was gullible ... and I believed them!"
"They told me I was gullible ... and I believed them!"