Port Problem

[Haden-Orbital]

Hello Kyle,

the exim version has failed on our system again, 80.88.198.51
username darren/root, password [removed by moderators]

the exim executable was not present so we I presumed another update was
run ? anyway the MTA is failing with the following error in the logs;


2004-08-30 14:16:08 1C1m10-0002yJ-R1 == enquiries@bearhotel.com R=central_filter defer (-1): internal problem in central_filter router (recipient is enquiries@bearhotel.com): failure to transfer data from subprocess: status=0100 readerror='No such file or directory'


ps when running the fix ./scripts/installpkg exim I get the following error towards the
end of the installtion;

*** Existing file called exim has the same version and compile number ***
*** Exim installation failed ***
*** Error code 1

Stop in /usr/ports/mail/exim/work/exim-4.42.
*** Error code 1

Stop in /usr/ports/mail/exim.
*** Error code 1

Stop in /usr/ports/mail/exim.
** Command failed [exit code 1]: /usr/bin/script -qa /tmp/portupgrade11812.0 make reinstall
** Fix the installation problem and try again.
---> Installation of mail/exim ended at: Mon, 30 Aug 2004 14:20:23 +0100 (consumed 00:00:01)
---> Fresh installation of mail/exim ended at: Mon, 30 Aug 2004 14:20:23 +0100 (consumed 00:00:22)
---> Reporting the results (+:done / -:ignored / *:skipped / !:failed)
! mail/exim (install error)
---> Session ended at: Mon, 30 Aug 2004 14:20:23 +0100 (consumed 00:00:23)



please advise ASAP,


Warm Regards,


Haden.

[eisman]

You may want to change your root password, right away. This wound up posting to a public thread.

[Holbrookau]

Exim failed again for us too after the automatic upcp yesterday (Australian time).
Running /scripts/exim4 and then a reboot got it working again but one customer is still reporting intermitent errors.
Why is this happening? - I'm getting tired of having to manualy fix things that upcp breaks. I don't bother putting in tickets any more - it takes 2 days to get a reply..

[neraath]

I would whole-heartedly have to agree that having to fix whatever is broken by the UPCP script is totally bogus. We might be good sys-admins (and there are some out there who probably aren't as good, and are at a complete loss), but good programming skills would mandate that stuff like this ISN'T supposed to happen.

But anyways, I think I fixed the problem for all FREEBSD users! *crosses fingers* In the above posts I did not see in my /etc/make.conf file the "WITH_OPENSSL_BASE" line. But, I did see the one package that was being compiled with exim (OPENSSL_OVERWRITE). If you look, there might be a COUPLE (yes, somehow mine had more than one) "OPENSSL_OVERWRITE_BASE=YES" lines in there. Comment those out. Then I commented out the "WITH_OPENSSL_PORT=YES" line, but somehow, one of the scripts ended up adding it back in there. That was okay, because when I ran /scripts/exim4, it rebuilt it and the build process was COMPLETELY different than the other because it was using the actual OPENSSL port rather than the OPENSSL_OVERWRITE_BASE port or whatever it was. In any case, I had a successful build 2 times, ran /scripts/upcp and it didn't bomb on me and ran /scripts/exim4 again and it seems to be working. I'll post tommorrow whether or not it fails tonight.

[neraath]

Okay, my fix above officially DID work. I recieved mail and the server is still live. The only problem (and I should have realized this) about it is that SSL on the outgoing way does not work. I plan on playing around with this some more when I have time, but officially, the above fix does stop the server from trying to rebuild itself and die every night.