Hi Christoph,

I had checked /tmp and it was ok, and I should have mentioned that in my
email.

However, I have now officially slapped myself on the back of the head
because /tmp was not the problem. It was the management user's home
directory where the openpkg*.src.sh was being built from. By the time that
the phase 2 shell script is being built, the two *.rpm's and the *.Z file
have already been built, and when the phase 1 script tries to uuencode the
*.Z file into the phase 2 *.sh file, the total exceeded 60MB, and that is
our quota on home directories, which are mounted file systems here. So I
rebuilt from /tmp and had no problems on this machine.

I guess what threw me off is the lack of error message, and I'm guessing
that the scripts redirect some of those messages to /dev/nul. It does make
problem solving more difficult, though. So it would be helpful if there were
a way to retrieve all error messages in cases like this. A "can't write to
disk" or "disk full" error would have made all the difference. Is there a
way to do this currently (a kind of debug mode)?

Regarding the other Solaris 9 machine where /tmp/rpm was getting "Killed", I
installed gcc from the binary distribution as Ralf suggested, and this now
works fine as well.

Thanks to everyone for their help!

Dennis

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Christoph Schug
Sent: Thursday, September 25, 2003 4:57 AM
To: [EMAIL PROTECTED]
Subject: Re: Strange problem bootstrapping OpenPKG rpm on Solaris 9


On Wed, Sep 24, 2003, Dennis McRitchie wrote:

[...]
> 3) New problem: The Solaris 9 machine that seemed to succeed in phase 1 of
> the bootstrap failed phase 2 of the bootstrap (i.e., executing the *.sh
> script created by phase 1). What happens is that at the end of phase 1,
> uuencode is used to encode the *.Z file into the new *.sh script. This
> appears (according to the output below) to succeed, but in fact, the last
> line in the *.sh script (also shown below) is incomplete (too short and no
> newline at the end). Thus when I try to run the phase 2 *.sh script, it
> fails (output below).
[...]

Can you please ensure that this is not a disk space issue on /tmp.
Setting TMPDIR before bootstrapping to a directory with more than enough
space (whatever this is, guess at least 500 MB ;-) might help.

-cs

______________________________________________________________________
The OpenPKG Project                                    www.openpkg.org
User Communication List                      [EMAIL PROTECTED]

______________________________________________________________________
The OpenPKG Project                                    www.openpkg.org
User Communication List                      [EMAIL PROTECTED]

Reply via email to