grub error 22. Bad partition table? [solved]

John DeDourek dedourek at unb.ca
Fri Nov 26 13:22:27 UTC 2004



Paul Howarth wrote:

> Duncan Lithgow wrote:
> 
>>> If Paul did say that, he's wrong. The _Master_ boot record is on hda 
>>> - the first sector.
>>
>>
>>
>> My mistake, I misunderstood by not seeing the difference between hda 
>> and hda1. If I've been able to follow this thread since I started it 
>> then I get the idea that while there is a _Master_ BR at the start of 
>> hda, there is also a space at the start of hda1 and each partition - 
>> for the purpose of each partition being able to be bootable and hold a 
>> boot manager.
>>
>> I am basically on track?
> 
> 
> Yes. The Windows bootloader that it puts in the MBR just looks for the 
> first "bootable" partition and loads the bootloader from the start of 
> that partition. On the other hand, if grub is installed in the MBR, it 
> uses its configuration file to decide what to do and completely ignores 
> the "bootable" flags on the partition table entries.
> 
> At least that's my understanding.
> 
> Paul.
> 
This would agree with my investigations.  In my case, I was working
with an IBM T42 and attempting to make the machine dual bootable,
including leaving the IBM diagnostic partition bootable using
Fedora Core 2.  (I do know about the BIOS/IDE geometry issue and
used the appropriate workarounds for that.)

My first attempt was to use Partition Magic to reduce the size of
the Windows XP partition, install Linux partitions in the free space
created, install grub in the linux boot partition and make the
linux partition active.

This did not work: two possible hypotheses.  I had something
configured wrong in grub.conf.  OR.  Either the BIOS or grub
didn't like the fact that grub and/or its stage1.5 were beyond
the 2 GIG boundary.

I finally gave up, and installed grub in the MBR.  Have a
successful dual boot machine, but can not currently get the
IBM diagnostic partition to boot.

John




More information about the fedora-list mailing list