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

Re: Only m68k and i386 in hamm [Fwd: from RH list]

For you folks using Alphas with Debian, here is a note from RH describing what to do
to work with boot disks.

Hope that helps you get your version working. I've found listening to both RH
and Debian is useful.

I've been patiently waiting for the Debian Alpha port so I can take a look at it.
I'd hate to see you not put it out just because of boot floppy problems.


Wes
--- Begin Message --- 
> 
> I have some serious problems with my NCR875 clone (Diamond Fireport 20)
> and IBM DCAS-34330 (4.3G) Ultra-SCSI-drive. I want to install RH5.0 from
> the scratch but that's quite impossible, because the NCR-driver which
> comes with milo-sx164-980324 (v2.4a) is way too old.

Check ftp://gatekeeper.dec.com/pub/DEC/Linux-Alpha/ for alternate
milo images.  I do not know about SX but we were using Diamond Fireport
20 and IBM DCAS on LX without any problems.

> Milo is unable to find more than the first two partitions of the drive,
> if it finds any partitions at all.

If you say so.

> The partitions made with fdisk doesn't get written to the hard drive.

And what this has to do with milo?  At this moment you are likely
running from a ramdrive which you loaded from a floppy; milo long
time forgotten.  BTW - if you have another Linux machine, not
even necessary Alpha, then changing (updating) a contents of three
floppies you are using for booting is not a very big deal.

A floppy with linload.exe and milo is just a FAT (MS-DOS) diskette.
The second one, from which kernel is loaded is plain ext2fs floppy
which can be mounted on any Linux box.  The third one contains
compressed, with gzip, image of a file system which gets loaded
into memory.  To get to this one you need Linux with either
ramdrive support or loop device support (or both).
In the first case you do
  zcat /dev/fd0 > /dev/ram ; mount /dev/ram my_mount_point
In the second case
  zcat /dev/fd0 > something ; mount -o loop something my_mount_point

After you performed all changes you wanted to get back a contents
back on a floppy you first 'umount' (important!) and, for the first
case:
  dd bs=1k count=4096 if=/dev/ram | gzip -v9 > /dev/fd0
In the second case
  gzip -cv9 something > /dev/fd0

4096 blocks above is a default.  This should be an actual size of your
ramdrive.

And that is about it.

> If I run fdisk second time during install, just after I've made the
> partitioning, I do _not_ the partitions which I have made, instead I
> get pure crap.

All instructions about a safe use of fdisk say "after you wrote a new
partition table - reboot; you do not want caching to screw you up".

> I'm really getting desperate, could someone help me

Calm down and think!  It helps.

I realize that to modify installation floppies you need some other
source of modified software - which can be a chicken and egg problem -
but it is often possible to find required pieces ready when you
will look around.

  Michal

-- 
To unsubscribe: send e-mail to axp-list-request@redhat.com with
'unsubscribe' as the subject.  Do not send it to axp-list@redhat.com

--- End Message ---
Reply to: