On Fri, 14 Jun 2002, Cory Daehn wrote:
> Actually, I posted what the log says in a previous message...
Sorry. I get snowed under by this list at times.
> I still
> have no idea why the hell exim is running with uid=8 euid=8 ... that's
> UID=adm GID=adm it's not *SUPPOSED* to be. the exim.conf file and the
> own/mod info on the program itself say it shouldn't be doing that at
> all... even if it was running mail.mail, it would be UID=12 GID=12 ...
> But here we go again...
I think you need to learn about exim debugging. It produces voluminous
output, but should help (either you or me) in determining exactly what
is going on.
> /root/ # exigrep 17IhZ0-0005mD-00 /var/log/exim/exim_mainlog
> 2002-06-13 22:15:50 17IhZ0-0005mD-00 <= emailbot@???
> H=vader.labnet.com (vader) [205.238.148.11] P=smtp S=1287
> 2002-06-13 22:15:50 17IhZ0-0005mD-00 Unable to get root to set uid 60003
> and gid 12 for local delivery to patricia: uid=8 euid=8
OK, having got a message in that state, what you need to do (as root) is
exim -d -M 17IhZ0-0005mD-0 2>/some/file
That will force a delivery, and collect the debugging information in the
file.
However, as you previously said that -qff did do the deliveries, I
rather suspect the delivery will work, and so we won't be any further
forward.
Analysis:
If it really is the case that deliveries fail as above when the message
arrives, but succeed when you subsequently use -qff, the problem would
appear to be that the delivery process started automatically by the
daemon is not getting the right permissions.
A thought: ARE you using a daemon? Or is this via inetd? If the
latter, what is the command in your /etc/inetd.conf file?
I think we need to know that answer to those questions before spending
time pursuing this further.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
