Re: d-i on ancient hardware
On Mon, Jul 10, 2006 at 01:56:54AM +0200, Frans Pop wrote:
> Next time please file an installation report instead of sending an
> email to the list.
m'kay.
> > On the other hand, if you insist on keeping loadlin, please provide a
> > .bat file. A beginner user won't know to look in isolinux.cfg to
> > copy the append line; also, there's no GPM in stone-age OSes...
> We should document it better, that is true (and ship a more recent
> version). On my TODO list, but fairly low down.
I guess that it would be enough to mindlessly convert isolinux.cfg
from:
LABEL install
kernel 2.6/vmlinuz
append video=vesa:ywrap,mtrr vga=788 ...
to:
rem LABEL install^M
loadlin 2.6\vmlinuz video=vesa:ywrap,mtrr vga=788 ...^M
(ie, piping it through a simple sed/perl/awk/whatever filter)
A fancy menu would cause portability problems, and is way under my
threshold of caring. Telling people to "uncomment the correct line
if you're unhappy with the default" would be enough for Microsoft
archaeology.
> > 4. [G-I] [Off-topic]: As you already don't support my magnificent 486
> > and similar, why won't you slap any theme on? The current widgets
> > are hideous, and since there's space for games...
> This is still Beta. We are working on that before the release, though as
> always opinions differ:
> http://www.yepthatsme.com/2006/07/08/debian-graphical-installer-excellent-work-guys/
It's an opinion about the installer in general, not the look of
buttons and scrollbars. And the latter doesn't have any _real_
value, yet an amazing amount of people care about pretty widgets.
It makes a helluva difference when trying to sell a program to a
customer... It costs nothing but a bit of disk space to replace the
Win95 look with something better.
> > 6. As proceeding through module selection on lowmem is doomed to
> > lose, the user needs to turn swap on before. Too bad, the initial
> > lowmem message doesn't suggest _when_ swap can be turned on.
> >
> > Various points:
> > * the lowmem message: console 2 is not yet available
> > * "Detecting CD-ROMs": this is when ide drivers are probed
> > * d-i module selection: can't continue past
> >
> > A simple hint in the lowmem message would save people some time.
>
> Not sure about that as IIRC swap is turned off again during partitioning.
Then buh-bye, 24MB. I doubt that it's a good idea to cut away large
pieces of the installer just to handle this case.
> The most important part is not to load more modules than you
> strictly need.
Can't get lower than "no extra modules at all, even fs ones" :p
> > 7. On 24MB+37MB swap, even with no extra d-i modules chosen, anna
> > goes into an OOM loop. Increasing the swap to 100MB doesn't help
> > either.
> We test releases at 32 and 48 MB. I'm not completely surprised that 24 MB
> does not work. We already know that the current lowmem levels need to be
> checked and probably adjusted.
I doubt if getting 24MB to work is really worth your time.
> > 8. I chosen IPv6 support -- incidentally, my network segment was
> > v6-only. Yet, DHCP failed and stateless configuration wasn't used
> > either. The manual config, being given 2002:5033:a761:4::1 started
> > to lecture me how an "IP" address looks like -- saying "IP" where it
> > meant "IPv4".
>
> I don't think IPv6 installations have ever really been tested. It would be
> great if you could do trace the cause of the problems and help us fix
> them. Please file specific bugs against the relevant d-i components if
> you find issues.
There are two issues:
1. DHCP+debconfage
2. wget
For 1., I'm not sure how widespread DHCP6 is. It's pretty new, while
radvd works even with Microsoft crap.
But, at the very least, having the manual config accept IPv6
addresses would be a must.
For 2., it looks like you already found the culprit:
> > 10. And yet, even though ftp.pl.debian.org is reachable over IPv6,
> > mirror-chooser wasn't happy. It turns out that the 'wget' binary
> > provided by busybox is a crippled IPv4-only.
> That should be a matter of enabling the relevant options in the busybox
> configuration.
>
> A quick grep on the busybox config gives me:
> CONFIG_FEATURE_IPV6 is not set
> CONFIG_FEATURE_WGET_IP6_LITERAL=y
>
> Looks like the first needs to be enabled...
Wonderful. And, udeb-floppy is separate from udeb, so an eventual
slight size increase won't be noticeable for netinst. I'm not that
sure if the diff is bigger than several bytes...
> > 11. It turned out the disk controller/mobo/something on the P1 box
> > was somewhat flaky, causing intermittent faults. If something bad
> > happens during debootstrap, the d-i wrapper over it gives a dialog
> > box with two choices: "Go back" and "Continue". Too bad, whatever
> > you choose, it proceeds with the latter -- yet, failing at the end
> > even if you fix the error manually.
>
> Handling of the two buttons can indeed be improved in some places. It's
> not always as easy as it sounds to define correct behavior though.
"Continue" may be trickier, but for "Go back", the correct behavior
smells like killing away the whole debootstrap's process group.
> > 12. Why does it use debootstrap AT ALL? This is something that could
> > be much better done at d-i build time instead of the installation
> > itself. Unpacking a tarball takes seconds even on this bitty box.
> > Debootstrap takes ages on modern hardware, too -- on the order of
> > 5mins on my desktop when pulling debs from a proxy over a 350KB/s
> > link.
> Please tell us how to get a full tarball on a floppy or inside a netbooted
> initrd...
The very same place these install methods pull debs from. Usually,
the CD.
For floppies/netboot, this would require additional disk space on
network mirriors, and thus it may be not worth it.
For netinst CDs, it would actually _decrease_ the space needed.
For full CD sets, the essentials would have to be put twice -- the
pre-deboostrapped image and separate packages on 10th or so CD in the
case an user wants to --reinstall glibc or bash. A good tradeoff for
taking away several minutes from every single install.
Even if you decide to optimize just netinst CDs/USB sticks, it's
worth it IMO. And, the logic:
if [ -f /cdrom/base.tar.bz2 ]
then
#wrapper blah blah
(cd /target;tar xvfj /cdrom/base.tar.bz2)
else
#wrapper blah blah
cdebootstrap etch /target
fi
is pretty trivial to implement.
> > 13. When booting, I get "Begin: Waiting for root file system... ..."
> > and a >5min wait before getting dropped to a shell. It then needs a
> > "modprobe ide-disk". Curiously, you need to wait several seconds
> > before closing the shell or it will fail again -- this gives a good
> > hint towards the cause.
> > (This happens both on the 486 and the P1, with the same disk).
>
> Please make sure you have very latest udev and initramfs-tools installed.
> If it still happens then, please file a bug report against
> initramfs-tools.
Sid's udev and initramfs-tools weren't enough, upgrading 2.6.15 (d-i) to
2.6.17 (sid) made it work. Cool.
> > 15. Once the installation finished, the keyboard mapping got screwed
> > on the next boot with regards to Caps Lock:
> > aBcDeFGHIJKlMnoPQRsTUVWxYz
This is not d-i's fault, dpkg-reconfigure console-data broke my sid
desktop, too. I'll look into the bug and report it there, at least
it's not your concern.
OOT:
I thought a bit about the whole CC-ing off-list people thing, and it
appears that it's both counterproductive and even a bit rude. For an
one-time poster, the cost is just
"mail -s subscribe debian-boot-request@l.d.o", adding a stance to
.procmailrc and then unsubscribing a few days later. Much cleaner
than asking people to CC me. Costs a tiny bit of network resources
instead of human time.
It may be a good idea for the lists' guildelines...
And, even though I'm one of spammiest posters on -devel (and
_certainly_ way too spammy compared with my level of clue), I
get CCed nearly on every reply...
Cheers and schtuff,
--
1KB // Microsoft corollary to Hanlon's razor:
// Never attribute to stupidity what can be
// adequately explained by malice.
Reply to: