On Thu, 2 May 2002, Bill Crawford wrote:
>On Wed, 1 May 2002, Gary Jackson wrote:
>
>> On Wed, 1 May 2002, Bill Crawford wrote:
>>
>> > It will pick up whatever is on the disk; this may be random garbage
>> >or a valid partition table. This looks like a valid table.
>>
>> It's not, I don't know why the redhat install kernel is picking that
>> up. It has a pre-existing valid table on there already, and shouldn't
>> have anything on there that should leave redhat to believe that it
>> should start doing geometry translation.
>
> It will pick up the geometry that was in use when the partition table
>was created. I've used this feature to force use of a particular spec
>sometimes (when moving disks between machines that tend to report the
>BIOS geometry differently).
That's the thing, it isn't.
>> > Use "dd if=/dev/zero of=/dev/hde bs=512 count=2" to wipe the existing
>> >table, then sync and reboot. You should see this effect disappear.
>>
>> Hm, I'm trying to preserve data in the extended partition on hde. I
>> don't think wiping out the partition table will help with that ...
>>
>> Booting a more recent kernel (2.4.18) on one of my 7.1 nodes shows
>> none of the detection problems that the RedHat install kernel is
>> having, so I'm guessing that the problem lies squarely with RedHat.
>
> If you want to preserve data, you'll have to keep the existing table,
>including the geometry used to create it.
Right. The install kernel is probing the partition table incorrectly.
I can't tell what the problem is -- as far as I can tell, after
reading the source, there should be no problem.
Booting the 2.4.9-13 kernel on a 7.1 node with an identical partition
layout shows the appropriate probe (2 primary partitions, 1 extended).
It's something funny and bad that the install is doing. I'll be
opening a bug report with RedHat (yet again ... I'm sick of finding
bugs in the RedHat installer).
--
Gary Jackson
[EMAIL PROTECTED]
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
