FC3 Install Problem

Mike McCarty mike.mccarty at sbcglobal.net
Mon Aug 15 15:47:36 UTC 2005


Edward Dekkers wrote:
> Mike McCarty wrote:
> 
>> Celeron 2.1 GHz with two Western Digital 80GB drives.
>>
>> This machine has been running Win NT 4.0 Service Pack 6
>> for a couple of years with no problems. Recently,
>> the first disc was wiped and reformatted as 40GB FAT32
>> and the rest unallocated. Windows 98 was installed in
>> the first (and only) partition.
>>
>> I booted Kanotix (Knoppix variant) and
>>
>> # fdisk -l
>>
>> Disk /dev/hda: 80GB, 80026361856 byte
>> 255 heads, 63 sectors/track, 9727 cylinders
>> Units = cylinders of 16065*512 = 8225280 bytes
>>
>>    Device  Boot   Start  End  Blocks   Id System
>> /dev/hda1   *         1  4865 39078181  c  W95 FAT32 (LBA)
>>
>> Disk /dev/hda: 80GB, 80026361856 byte
>> 255 heads, 63 sectors/track, 9727 cylinders
>> Units = cylinders of 16065*512 = 8225280 bytes
>>
>>    Device  Boot   Start  End  Blocks   Id System
>> /dev/hda1   *         1  9729 78148161  7  HPFS/NTFS
>>
>> I reboot, using the FC3 install disc 1, and test all media.
>> All complete test, and pass.
>> Select English Language, Selece U.S. English keyboard
>>
>> A message box pops up with a blue bar in the top with
>> the word "Bug" in it. In the bottom part is the text:
>>
>> Assertion (heads < 256) at disk_dos.c:486 in function
>> probe_partition_for_geom() failed.
>>
>> There are two clickable buttons, one saying "Ignore" and the
>> other "Fail" (I think, I'm sure about the "Ignore").
>>
>> I click "Ignore" and another comes up. So I click "Ignore"
>> a total of four (4) times, and am asked whether I wish
>> to auto parition or manual. I select manual, so Anaconda
>> will show me what it thinks the disc setup is, and I get then
>> have to "Ignore" the same message six (6) more times.
>>
>> Then I get a description of the disc from Disk Druid's
>> point of view.
>>
>>
>> Drive /dev/hda  (76317 MB) (Model: WDC800BB-00DAA3)
>>
>>   Free
>>   76316 MB
>>
>> Drive /dev/hdb (76317 MB) (Model: WDC8007B-00CRA10
>>
>>   hdb1
>>   76316 MB
>>
>>
>>    Device      type       size    start    end
>> /dev/hda
>>    /dev/hda    vfat       38162       1   4865
>>    Free        Free space 76319       1   9730
>>
>> /dev/hdb
>>    dev/hdb1    ntfs       76317       1   9729
>>
>>
>> So, it appears that Disc Druid thinks that the
>> unallocated and the allocated space overlap. It
>> also seems to think that the unallocated space
>> is larger than the whole drive.
>>
>> As a note, before I started all this, I ripped the MBR
>> from the first disc and stored it in a file on a floppy,
>> using
>>
>> # dd bs=512 count=1 if=/dev/hda of=/mnt/floppy/mbr.bin
>>
>> I then transferred the floppy to a handy Windows machine,
>> and used debug to look at it. It had the 0x55 0xAA boot
>> signature on it, and a cursory look at the disassembly
>> looked like startup code, and the strings present in it
>> looked like reasonable error messages from an MBR
>> about invalid partition tables, etc.
>>
>> I don't have that floppy here, but I can get it, and
>> give a hex dump of it for perusal of the PT if it
>> is deemed necessary.
>>
>> My questions:
>>
>> Is this really a defect in Anaconda/Disk Druid?
>> If so, where/how should I report it?
>> If not (or even if so) what should I do to finish
>> the install? So far, I just backed out, and verified
>> that Win98 still boots with no problem, and shut down.
>>
>> Mike
> 
> 
> Type that error in to google. This is a problem I've had on two occasions.
> 
> There's a recursive partitioning command I can't remember I think based 
> on sfdisk that re-sets the partition geometry.
> 
> I haven't kept it, but each time the search on google pulled up the 
> right advice.
> 
> On each occasion it's fixed the problem for me.

Thanks for the reply. I found the bugzilla report, and it says it
the resolution status is RAWHIDE.

If the bugzilla report is supposed to have a fix or where to find
it in it, then it is buried in the middle of hundreds of lines
of commentary. I read every line of what was written there, and
I didn't see where to get anything from Red Hat to fix, just a
pointer to text that says that there is a fix, somewhere.

I've been partitioning drives since 1984 or thereabouts, and
this should not be a mystery. Reading between the lines, the
tool I used to partition that drive (from Western Digital) did
not write a BR into unallocated space. This seems like the right
thing to do, to me. Unallocated space should *not* have a BR.

Why should unallocated space be such a difficulty for a tool
which is an installer? It ought to be able to start with a
blank disc.

Also, I saw the comment that there is no way for the tool to
find out how many heads are used by the virtual translation
LBA from the BIOS. This also doesn't sound right. Does the
developer mean that asking the BIOS to seek past the number
of heads it is pretending to have doesn't result in a
failure?

Anyway, thanks for the reply.

Mike
-- 
p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);}
This message made from 100% recycled bits.
You have found the bank of Larn.
I can explain it for you, but I can't understand it for you.
I speak only for myself, and I am unanimous in that!




More information about the fedora-list mailing list