8.0-Beta installation problem -- Unable to find /dev/ad0s1b

[Jin Guojun]

Did not find anything from current, but found the same problem has been reported in earlier releases in those archives, and the latest was May 2009:

    * Rambler: FreeBSD mail archives search
    * MarkMail: FreeBSD mail archives search

Two directly related to 8.0 was on Mar 25, 2009. One failure was due to the FreeBSD is on the second slice, which sound liek a bug; but the other is not clear.

In my case, FreeBSD is on the first slice, so it is not the same problem.

>From 8.0-Beta{3, 4} cannot recognize FreeBSD 7.2 partition tables, it looks like that there is a disklable and/or partition related problem in 8.0-Beta.

--- On Mon, 9/14/09, Robert Huff <roberthuff at rcn.com> wrote:

> From: Robert Huff <roberthuff at rcn.com>
> Subject: 8.0-Beta installation problem -- Unable to find /dev/ad0s1b
> To: "Jin Guojun" <jguojun at sbcglobal.net>
> Cc: questions at freebsd.org, freebsd-stable at freebsd.org
> Date: Monday, September 14, 2009, 2:13 AM
> 
> Jin Guojun writes:
> 
> >  Tried to install from both 8.0-BETA{3, 4}-i386-dvd1.iso on a
> >  Phenom 9600 system and having some disk problems.

The system is running 6.x and/or 7.2 FreeBSD (on different drives).

Disconnected rest drives and only left 7.2 drive in.
ad0: 76319MB <Seagate ST3802110A 3.AAJ> at ata0-master UDMA100

Boot 8.0-BETA3-i386-dvd1.iso and sees no partition at all on the 7.2 disk.
Boot back to 7.2 and everything works fine, so system has no hardware
problem.

Boot 8.0-BETA3-i386-dvd1.iso again, still sees no partition, so
reconfigured partition with
autoconfigure - A

512M for / on /dev/ad0s1a
about 4G for swap on /dev/ad0s1b
about 2.2G on /var
512M for /tmp
rest spaces are for /usr 

> >  After commit, installation says --
> >  
> >  Unable to find device node for /dev/ad0s1b in /dev
> >  The Creation of file system will abort.
> >  OK
> >  [Press enter or space]
> 
>     Look at the archives of current@ within
> the last three weeks.
> I (and I believe at least one other person) had what seems
> like the
> same problem and found a work-around.
> 
> 
>            
>     Robert Huff
>