Problems running Red Hat Linux 6.1 and 6.2

[mikebw at colossus.bilow.com (Mike Bilow)]

This is a grave misstatement in the Lilo docs, then.  The "linear" mode
enables 24-bit LBA disk access, which is needed to load the kernel if the
boot area is on a partition that exceeds the 1024-cylinder boundary. 
Since only SCSI drives had more than 1024 cylinders historically, this
was, in fact, mostly used for SCSI drives.

Unfortunately, "linear" is also needed for IDE drives if the boot area is
on a partition larger than 512 MB.  So this is why it is the default.  In
general, "linear" should work if your motherboard BIOS supports LBA as
nearly all modern releases do.

As of Lilo 21.4, there is a new keyword "lba32" which enables the newer
32-bit LBA access.  This is necessary if your boot area is on a partition
that exceeds 8.4 GB.

What is causing your problem is that Lilo must be run in userland and with
whatever configuration will actually be used to boot.  If you upgrade your
kernel, for example, then you must run Lilo from the console.  When Lilo
reads the /etc/lilo.conf file, it will figure out the physical location of
the boot kernel according to the rules stated.  If, for example, you use
the "linear" keyword and run Lilo, then you will not be able to boot if
your machine does not support LBA mode.

-- Mike


On Tue, 4 Apr 2000, Ken Gosier wrote:

> One possibility, if it is something to do with the hard
> drive: I just upgraded from 6.0 to 6.2 also.
> 
> I noticed at the LILO config screen, it had the option
> marked as default to read the hard drive in linear mode,
> and it said that this mode was mostly used for SCSI drives.
> 
> Even though I'm IDE all the way. I don't remember it trying
> to misread my hard drive like this on previous
> installlations.
> 
> --KenG


-
Subcription/unsubscription/info requests: send e-mail with
"subscribe", "unsubscribe", or "info" on the first line of the
message body to discuss-request at blu.org (Subject line is ignored).