9.10 Server i386: Installation step failed .. failing step is: Select and install software

This did not appear in our pre-release testing (we would not have
released 9.10 with such a bug), which means that we need information
from you to track down the failure.

Could you please attach the /var/log/installer/syslog file from your
completed installation? (I'm assuming that you didn't reboot between the
failure and correcting it.)

 status incomplete

I seem to be in the same situation - installation fails in "select and install software" step.

Boot from CDROM, tried in expert mode and in normal mode.

It says that "something" has failed - it only showd the failed command ... the whole dpkg output in a text field would be great in such an unexpected case.

Anyways, created debug logs and entered the shell.

The last entry I saw failing was the linux-headers-2.6.31-pae package so I ran the command it showed me in the error window during installation: "chroot /target --install --force-...something /media/cdrom0/[...]/linux-headers-2.6.31-[...].deb" and ran that command there in the shell.

To my astonishment, it gave up with the following:

~ # df
Filesystem 1K-blocks Used Available Use% Mounted on
tmpfs 222312 144 222168 0% /dev
/dev/sr0 655944 655944 0 100% /cdrom
/dev/sda1 19528876 605744 17946384 3% /target
/dev/sda1 19528876 605744 17946384 3% /dev/.static/dev
tmpfs 222312 144 222168 0% /target/dev
/dev/sr0 655944 655944 0 100% /target/media/cdrom0
/usr/bin # chroot /target/ dpkg --force-all --install /media/cdrom0/pool/main/l/linux/linux-headers-2.6.31-14-generic-pae_2.6.3
1-14.48_i386.deb
(Reading database ... 14458 files and directories currently installed.)
Unpacking linux-headers-2.6.31-14-generic-pae (from .../linux-headers-2.6.31-14-generic-pae_2.6.31-14.48_i386.deb) ...
dpkg: error processing /media/cdrom0/pool/main/l/linux/linux-headers-2.6.31-14-generic-pae_2.6.31-14.48_i386.deb (--install):
 unable to create `/usr/src/linux-headers-2.6.31-14-generic-pae/include/config/nf/nat/proto/udplite.h.dpkg-new' (while processing `./usr/src/linux-headers-2.6.31-14-generic-pae/include/config/nf/nat/proto/udplite.h'): No space left on device
dpkg-deb: subprocess paste killed by signal (Broken pipe)
Errors were encountered while processing:
 /media/cdrom0/pool/main/l/linux/linux-headers-2.6.31-14-generic-pae_2.6.31-14.48_i386.deb
/usr/bin #

... which in my opinion makes no sense.

Only thing I could think of are too strong or unlucky compilation flags for dpkg or the running kernel or file system code, something I had sometimes in a custom-compiled Gentoo system.

Will try to attach my full syslog output I transfered out via USB storage.

Promised installer syslog.

I am facing the same problem with the amd64 alternate install CD. The only way I managed to get past this issue was by constantly modifying /target/etc/apt/sources.list to a valid ubuntu http mirror. I am not sure if the problem is that it is trying to look for _Karmic Koala_ instead of "Karmic Koala" when it tries configuring APT.

Ernst, if it says you're out of disk space, then it means it. Compiler flags have nothing to do with it. I'll grant you that our error handling in this case isn't great ...

Well, notice the line from my extract from "df":

/dev/sda1 19528876 605744 17946384 3% /target

... and it says "no free space". This should not be! There are 17G free space. I assume a problem in the installer logic.

I am also certain from experience that compiler flags can create misbehavior in code from "optimizations".

I have the same problem as the first comentor, just hoping for an update on this issue.

[Expired for pkgsel (Ubuntu) because there has been no activity for 60 days.]