Re: SX164 installation
Interesting... I too switched from RH 4.2 to Debian this week. I also have
an SX164. I never tried the kernel on the rescue disk; I just overwrote
it with the Redhat install kernel. I downloaded base2_0.tgz to my redhat
partition, mounted and installed it from Debian. So far so good...
My only real problem is that the redhat kernel has paniced twice during
the install. I've been able to fsck and continue, but this never
happened while running redhat. I'm assuming there are bugs in the 2.0.30
kernel which may be triggered by glibc or something else in Debian.
I'm going to build a devel kernel next and attempt to get hwclock
working. I've heard there are problems with dpkg on newer kernels, so I
guess if I need to run dselect I can boot the old kernel and hope it
doesn't panic again :(
Anyway... kudos to the debian developers. The experience has already
been much nicer than with Redhat, and it'll certainly get easier when the
Alpha port is finally released.
Jeff
On Wed, 16 Sep 1998, Ettore Aldrovandi wrote:
> Hi everyone,
>
> I just switched from RH4.2 to Debian (which I was dearly
> missing, I've had 1.3 and now 2.0 on an Intel box) on my
> SX164, and I'd like to share my experience with the list.
>
> So, I first tried a fresh install using Loic's disks from
> http://lhpca.univ-lyon1.fr/axp/debian.html
>
> Well, that didn't work. As noted by somebody else on this
> list, the resc disk boots the kernel just fine, but then
> asks for the root disk and after pressing enter nothing
> happens until one gets fd0 timeout.
>
> Ok, second try. I copied the RH kernel 2.0.30 (sorry!) to
> the resc disk and this time it worked! So I got into the
> installation program. I have a CD-R which I got from LSL to
> take the .debs from, and of course I tried to use that
> one. Problem is, they didn't put the base system and the
> resc, boot and drv disk images on the CD-R so the
> installation program wouldn't go ahead.
>
> Uhmmm... third try. This time I booted into the small RH
> partition, copied a base tarball I had prevously downloaded
> (when Loic disks weren't around) into the Debian root,
> untarred, configured a few network things, rebooted into
> Debian root, made /usr /var/ and /home partitions and ran
> dselect. I should probably mention somebody else described
> this approach on this list a while ago. Also, MILO was
> already on the hard drive from the original RH installation.
>
> Now I hear you asking, why didn't you go ahead in the fresh
> install using, say, the base floppies? Well, I don't know,
> it didn't occur to me. I suppose this things happen at
> 2:00am.
>
> Ok, to go back to what happened, at that point verything
> went more or less smoothly. I had to
>
> dpkg --ignore-depends=libc6
>
> a few things, but this is apparently well known. Other
> things in no apparent order are as follows:
>
> o there is something wrong in the apache installation and it
> dies when I try to start it manually or from the boot
> scripts. I haven't had time to look into __what__ is
> wrong, though.
>
> o compiled a few small test C and C++ programs, and egcs
> seems to be OK. A project of mine is to port the CERNLIB
> to this machine. I'm not a Fortran programmer, but my wife
> does some number crunching, and I'm sure it will be
> interesting to see how this little beast does.
>
> o TeX/LaTeX is fine. Xdvi, on the other hand, dies
> spectacularly when I click in its window to zoom, kills
> the Xserver which doesn't restart, and the machine is
> completely locked: no ctrl-alt-bksp, virtual consoles,
> nothing. I can only reset, at that point. Unfortunately, I
> don't have any other machine lying around, so I can't tell
> whether or not the Alpha is completely dead. Oh, the video
> card is a Matrox Millennium II, with XF86_SVGA.
>
> o gv segfaults, ghostview is fine, instead.
>
> o ppp configuration was a breeze.
>
> o emacs is ok, both 19 and 20. Xemacs, I have two packages
> wich call themselves support binaries and/or libraries (I
> can't remember and right now I'm away from the Alpha) but
> no binary. Can some helpful soul recap how's the situation
> with Xemacs? I did have a look at the list archive, but I
> don't quite get the picture. Do I have to compile it
> myself?
>
> o I'm still going with the RH Kernel. I discovered that the
> Kernel that comes with the resc disk fails to detect the
> PS/2 mouse. So, as you can imagine, no gpm, no X. Might
> this be related for some obscure (to me) reason to the
> failure to load the ramdisk I and others observed?
>
> I'm obviously living in the future: 2018 :-) I guess one
> can look at this as a test for the Y2K problem, but I'd
> like to come back. You see, time-travel means a lot of
> nasty paradoxes... I understand I have to recompile a
> kernel to get proper support for /dev/rtc. The question to
> the experts, is: can I take a source package, say 2.0.33
> or 2.0.34 and go ahead and do it or do I have to patch it?
> I discovered my ideas are quite confused.
>
> Ok, that's it, more or less. I cannot end this without
> praising the efforts of all those who made Debian on the
> Alpha possible. It's so much better than RH (but know that
> already, right?), and I'm so happy to be using Debian
> again.
>
> Ciao,
> Ettore
> --
> Ettore Aldrovandi -- SISSA/International School for Advanced Studies
> Via Beirut 2/4, I-34013 Trieste, Italy
> ph: [+] 39 0403787445 -- fax:[+] 39 0403787528
> ettore@valis.fm.sissa.it aldrovan@sissa.it http://www.sissa.it/~aldrovan
>
>
>
> --
> To UNSUBSCRIBE, email to debian-alpha-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
>
Reply to: