Partitioning fails due to drive geometry inconsistency

[Peters Micheal A Contr GSI/SCBN]

Ok, which OS was installed on it before trying to install BSD?

-----Original Message-----
From: owner-freebsd-questions at freebsd.org
[mailto:owner-freebsd-questions at freebsd.org] On Behalf Of Carsten Zimmermann
Sent: Monday, May 03, 2004 1:03 PM
To: Malcolm Kay
Cc: freebsd-questions at freebsd.org
Subject: Re: Partitioning fails due to drive geometry inconsistency

Thanks for the answers so far.

I tried to use the autocorrected values, of course. But I can't create a
partition with them (as said: I was able to delete a NTFS partition w/o
problems). 

When I try to write the partition table, sysinstall / fdisk catches signal
11. It seems fdisk is not able to write the partition data since the (now
autocorrected) drive geometry is (still) not valid.

Carsten

Am Mon May 03, 2004 at 11:0325PM +0930 schrieb Malcolm Kay:
> On Monday 03 May 2004 18:33, Carsten Zimmermann wrote:
> > Hello list.
> >
> > I know this had been discussed quite often, but none of the posting 
> > google offered worked out for me. Please (B)Cc: me as I am not on the
list.
> >
> > I have a Western Digital WDC WD400AB-00CMB0 40 GB IDE drive hooked 
> > in as pri slave in my system and wish to install FreeBSD on it. This 
> > is not my first install, but it won't work this time (using 5.2.1-R):
> >
> > I try to partition the drive with fdisk during the installation process.
> > Everytime I hit "w" to write the changes, the program catches ``sig 11''
> > (segfault?). Curious thing is: I was able to delete partions but I 
> > cannot create new ones. I tried every option concerning bootsector 
> > (boot mngr, std mbr, none) and the error occures every time.
> >
> > When I enter the fdisk tool it tells me that the drive geometry 
> > "778545/16/63" is wrong. I don't have a clue where it gets that value.
> 
> This set of figures would correspond to a disk of 401GB (k=1000) or 
> 374GB
> (k=1024) -- Clearly not a 40GB drive.
> 
> > The BIOS detects the geometry as 19158/16/255. The auto-corrected 
> > value
> 
> This set corresponds to 40.0Gb (k=1000) so seems to be an acceptable 
> set of numbers except that I understand that the seectors per track 
> can't exceed 63 so the this set is unsuitable. In LBA mode it is not 
> necessary that the specified geometry is the same as the physical 
> reality. It is common to set the heads and sectors to maximum values
255/63.
> 
> > is 4865/255/63 which equals the BIOS-setting when access mode is set 
> > to LBA manually. When I enter the geometry manually with "g", it 
> > still
> 
> Which corresponds to this auto-corrected value and again calculates to 
> 40.0GB (k=1000). This looks like the best geometry to use.
> 
> > mourns the value 19158/16/255 is not correct and again falls back to 
> > those LBA values.
> >
> 
> Again the 255 value is too large. Don't try to set your own geometry; 
> go with the auto-corrected value.
> 
> Malcolm
> 
> 

--
Carsten Zimmermann
mailto:cz at aegisnet.biz
Tel.: +49-(0)40-98760543

Aegis:Net IT-Dienstleistungen
Postfach 620371
22403 Hamburg
http://www.aegisnet.biz
Info: +49-(0)700-AEGISNET
Fax: +49-(0)40-98760547