[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: installation difficulties


About two weeks ago, I wrote to the list with some problems doing my first
non-x86 Debian install (on a Sparc 5 I picked up used).  Everyone was
really helpful in pointing out some of the architecture's peculiarities
(max 1Gb root partition, which at first confused me since I assumed this
BIOS bug was similar to that of old PC's where the START of the boot
partition had to be in the first n sectors; it seems on this sparc,
though, that the ENTIRE boot partition has to be under 1 gig).

Anyway, everything's working fine now, but I came across what I thought
might be a bug in the standard debian install, but i'm not confident
enough that I felt to put it in the debian bugtracker, so instead thought
I'd post it here:

To get the install procedure to recognize my internal SCSI CD-ROM drive
(OEM sun) after the reboot portion of the install (it automatically can
boot the cd-rom during start of install), i had to hand-edit /etc/fstab
and change:

/dev/cdrom       to             /dev/scd0

in the first column.  Then the continuation of the install (when it
installs all the selected packages) will automatically find the cd-rom as
it suggests that it should do.

Anyway, again, I'm not sure if this is a bug or not, and I don't really
need any help on it, since I managed to make it work, but I just wanted to
share it with people who are more knowledgeable in case it should be
corrected in the sources/ISO images.

Thanks so much for all the helpful advice while I was installing.

Take care,


On Thu, 19 Oct 2000, Jeremy Rashbrooke wrote:

>  Hi Daniel,
>  Curiously I'm having similar issues with my second Ultra-10 .. I think I
>  fumbled thru the first one by doing the sun-disk-label thing, and taking
>  the cyl/heads/sectors from the Alt-F3 (autodetected) stuff.  It was
>  disturbing that it couldn't detect this itself .. since it had already automatically
>  detected this itself a minute earlier.  [shrug]
>  Anyway .. to your problem.
>  insmod/modprobe/depmod should work the the same way it does under
>  i386.  The scsi cdrom device should be (!) /dev/scd0.  apt-get has a
>  special mode for cdrom's .. so you shouldn't need to mount it yourself.
>  If that's not happy, then you'll have to check /lib/modules/2.x.yy/scsi
>  for something that looks like 'scd.o' -- and insmod that first.
>  If you're using the default kernel image, then I imagine that scsi cdrom
>  support is embedded within the kernel, though.
> >>> Daniel Freedman <freedman@ccmr.cornell.edu> 8:43:41 19/10/00 >>>
> Hi everyone,
> Thanks for all the advice so far.  I've made some progress but I'm still
> not all the way there.  Now that I partition the Sparc 5 with a 500 Meg
> "/", 1GB "/usr" and 512Meg swap, I'm able to get the bootup during
> installation to have SILO find the vmlinuz and continue the installation
> procedure (actually, its not hands-free as it initially boots with the
> dreaded error:
>   Boot device:/iommu/sbus/espdma@5,8400000/esp@5,8800000/sd@3,0 File and
>   args:
>   Bad magic number in disk label
>   Can't open disk label package
>   Can't open boot device
> but then I type "boot disk1" at the "ok" prompt, SILO comes up and finds
> vmlinuz and it continues).
> However, just before starting installation of all the packages, it asks
> for the media on which they're available.  I specify the CD-ROM (Toshiba
> 4x SCSI internal bootable that came with Sparc 5), but the routine is
> unable to find my CD-ROM automatically.  It asks for me to specify the
> cd-rom device (I forget the exact message), so I try both "/cdrom" and
> "/dev/cdrom", but these fail with the error "No block device found".  I
> can't install over the net because the academic environment I'm in won't
> give me an IP drop for this machine.  I'm thinking maybe I have to install
> a loadable SCSI module for the kernel to be able to find the CD-ROM, but
> the boot procedure is able to speak to the SCSI hard drive and recognizes
> the CD-ROM on bootup (to speak nothing of installing the base package from
> it), so I'm not sure if this is it. Further suggestions would be very much
> appreciated (also on how to have the initial power-on boot go directly to
> the appropriate disk1 rather than giving me the error message.  Must I
> tweak silo.conf?).
> Thanks so much and take care,
> Daniel
> On Tue, 17 Oct 2000, Ben Collins wrote:
> > Also try "ls" at the silo prompt and see if it's there. More than likely
> > this is caused by boot being on a seperate partition. Might want to redo
> > that installation without the extra /boot (try a 500Meg "/", 1G "/usr" and
> > 512Meg swap).
> > 
> --  
> To UNSUBSCRIBE, email to debian-sparc-request@lists.debian.org 
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org 

Reply to: