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

Re: Status of Debian on Sun Blade 150



I don't think I can necessarily help with your parted issue. I just
wanted to say that I was amazed with your creativity and patience in
creating this special installation, and that I was impressed with your
diagram of the disk. Also, the black magic you described in your e-mail
to get this all to work was well thought out and very hilarious. It
reminded me of what I did at work today:

A few weeks ago, I spent half of a workday fighting with the Xerces-C++
XML processing library which we need for validating some integration
data files. I was trying to get it to compile on Solaris 2.9 in 64 bit
mode with GCC/G++ instead of the officially supported (but expensive
and not installed by default at work) Sun Forte. Their
pre-GNU-configure build script performs some disturbing mungefest on
some environment variables, inserting some secret values that
GNU-configure uses to configure the build. Of course, the secret values
that work right for Forte don't work well at all for GCC/G++. Not
surprising.

So, today, I spent half of a day reverse-engineering the secret values,
then making some educated guesses at what GCC/G++ and GNU-configure
might prefer that they be. Amazingly, it built right on the first try!
However, ironically, when I tried to redo the build with optimization,
it failed.

I believe I figured out why. Firstly, GNU-configure makes some changes
to the tree that make distclean does not fully reverse. Secondly, I
moved the directory to a different volume between compilations. I
believe that this broke some of the directory paths that GNU-configure
hardcoded into the source tree someplace. Tomorrow I'll fix it all (at
least everything always seems to work right on Fridays at my job) and
make it happy again.

One philosophical question. Does the fact that I am demented enough to
figure out how this sick and twisted build system works lead to the
need to question my own mental state? I wondered about this on the way
home this afternoon.

Anyhow, thanks for motivating me to write this funny story down.

--- Wiktor Wandachowicz <wiktorw@ics.p.lodz.pl> wrote:

> Jurij Smakov wrote:
> 
> > Everyone is working towards the same goal here, so let's not get
> > too picky about the choice of words.
> 
> Thanks for bringing me back to my senses. Let's focus on the topic.
> 
> # fdisk -l /dev/hda
> 
> Disk /dev/hda (Sun disk label): 16 heads, 255 sectors, 19158
> cylinders
> Units = cylinders of 4080 * 512 bytes
> 
>     Device Flag    Start       End    Blocks   Id  System
> /dev/hda1           258       319    124440   83  Linux native
> /dev/hda2             0       258    526320    3  SunOS swap
> /dev/hda3             0     19158  39082320    5  Whole disk
> /dev/hda4         15391     19156   7680600    2  SunOS root
> /dev/hda5          3660      6744   6291360   83  Linux native
> /dev/hda6          6744     10574   7813200   83  Linux native
> /dev/hda7         10574     15391   9826680   83  Linux native
> /dev/hda8           319      3660   6815640    8  SunOS home
> 
> I'd like to reiterate: I was doing tests on a live system. It came
> to life this way:
> 
> 1) Install Solaris 9 on a clean disk:
> 
> +----------------------------------+  \
> | hda2   SunOS swap         514 MB |   \
> |                                  |    |
> +----------------------------------+    |
> |                                  |    |
> |                                  |    |
> +----------------------------------+    |
> | hda8   SunOS home           6 GB |    |
> |        UFS   / (Solaris)         |    |
> |                                  |    |
> +----------------------------------+    |
> |                                  |    |
> |                                  |     \
> |                                  |      >  hda3  Whole disk
> |                                  |     /
> |                                  |    |
> |                                  |    |
> |                                  |    |
> |                                  |    |
> |                                  |    |
> |                                  |    |
> |                                  |    |
> +----------------------------------+    |
> | hda4   SunOS root         6,5 GB |    |
> |        UFS   / (Solaris)         |    |
> |                                  |   /
> +----------------------------------+  /
> 
> Solaris installer proposed this in different way:
> (look where hda1 and hda2 were going to be!)
> 
> +----------------------------------+  \
> | hda2   SunOS swap         514 MB |   \
> |                                  |    |
> +----------------------------------+    |
> | hda8   SunOS home         ~23 GB |    |
> |                                  |    |
> |                                  |     \
> |                                  |      >  hda3  Whole disk
> |                                  |     /
> |                                  |    |
> +----------------------------------+    |
> | hda1   SunOS root           6 GB |    |
> |                                  |    |
> |                                  |   /
> +----------------------------------+  /
> 
> ... but I persuaded it to the previous layout.
> 
> 2) Use a mixture of Debian/sparc64 and Gentoo/sparc64
> install CDs to create additional partitions and ext2/ext3
> filesystems:
> 
> +----------------------------------+  \
> | hda2   SunOS swap         514 MB |   \
> |                                  |    |
> +----------------------------------+    |
> | hda1   Linux native       126 MB |    |
> |        ext2  /boot  (separate)   |    |
> +----------------------------------+    |
> | hda8   SunOS home           6 GB |    |
> |        UFS   /      (Solaris)    |    |
> |                                  |    |
> +----------------------------------+    |
> | hda5   Linux native         6 GB |    |
> |        ext3  /      (Debian)     |     \
> |                                  |      >  hda3  Whole disk
> +----------------------------------+     /
> | hda6   Linux native         6 GB |    |
> |        ext3  /home  (separate)   |    |
> |                                  |    |
> +----------------------------------+    |
> | hda7   Linux native       ~10 GB |    |
> |        ext3  empty  (spare)      |    |
> |                                  |    |
> |                                  |    |
> +----------------------------------+    |
> | hda4   SunOS root         6,5 GB |    |
> |        UFS   /home  (Solaris)    |    |
> |                                  |   /
> +----------------------------------+  /
> 
> 3) Install Debian sarge around February 14th 2005.
> 
> I didn't let parted format any partitions apart from maybe
> hda5 and hda6 (I can't remember really). Instead, I ordered
> it to use existing partitions without deleting contents,
> because they were empty. And it worked.
> 
> At that point it was a real achievement, b/c current d-i was
> unusable on Sun Blade 150 at the time. So I used an old version,
> dated 20040511, with kernel 2.4.26. It was a long process, but
> eventually I've finished that with only minor problems.
> 
> After that I've had a working multiboot machine. Debian sarge
> became the default, whereas specifying 'other=4' in silo.conf
> let the Solaris start. A one-liner into boot scripts let the
> Debian reuse Solaris swap (an mkswap before swapon, really).
> Overall success, if you ask me.
> 
> 4) Then me and another fellow cloned the 'mother copy' to the
> remaining 9 machines, using a mixture od 'dd', 'scp', and
> several other programs. Don't ask. It just worked.
> 
> 5) Recently I tested the unofficial netboot installer, just as
> Jurij asked. On the first run parted could see all the partitions
> correctly. So I think this proves that I did everything right.
> 
> I let the partitioner format hda1 (/boot), doing a backup first
> onto hda5, of course. A fresh Debian has been installed on the
> spare hda7 partition. In the end SILO 1.4.8 has been replaced
> by SILO 1.4.9, the fresh Debian became the default, and the only
> bootable system on the machine. I added several lines to silo.conf
> (just the like I had in backup), this way I can boot three OS-es
> now  (Solaris9, Debian sarge and fresh Debian sarge). After that
> I've found a problem in SILO 1.4.9 which I reported as bug #306012
> (it doesn't display a message file before boot prompt).
> 
> On every next run parted became unable to list existing partitions.
> The machine still boots, I checked the partitions using fdisk
> (the dump at the beginning of my post is from this fdisk, actually).
> Everything works apart parted. Something is strange.
> 
> I hope that the above clears my position and shows that my intentions
> are pure :-)
> 
> Hoping to find a reasonable resolution and friendly,
> Wiktor Wandachowicz
> 
> 
> -- 
> To UNSUBSCRIBE, email to debian-sparc-REQUEST@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact
> listmaster@lists.debian.org
> 
> 



Reply to: