>
> Hi all,
>
> I'm back and it is all working with no errors, but I am really
> puzzled as to why.
>
[SNIP]
>
> spamclog being a script containing
> #!/bin/bash
> /usr/bin/tee -a /home/spamc-logs/spamc-in.log | /usr/bin/spamc | tee -a
> /home/spamc-logs/spamc-out.log
>
> This meant I could see what was being sent to spamc and what spamc
> was sending back. Next I just had to wait for it to go wrong
> and then examine the log files. But it hasn't gone wrong since
> I started logging all the traffic 5 days ago!! Before I started
> logging it would go wrong maybe 10 time a day.
Don't you sometimes really, really hate computers?
Having worked for 5 days it let me send an email telling you all it
was working as it should. Then it went wrong!! Aaagghh.
So it looks like just logging the traffic hasn't fixed it, just
drastically reduced the error rate.
Monday evening an email from buqtraq got bounced and I managed to get
all traffic to/from spamc. Apart from the headers added by spamc the
input and output are identical. To my uneducated mind this points to
Exim possibly being the culprit and not spamc.
Or is it? I may have just noticed something that helps a lot. In all
the cases of the error that I have just checked back on, 5 minutes has
elapsed since receiving the mail and the error ocurring. I'm no
programmer but a quick grep through the source code and I found:
globals.c:int smtp_receive_timeout = 5*60;
Could this be the source of the problem? Spamc must be either hanging
or dragging its heels in passing the mail back to Exim, which is timeing
out, giving me my error. Shouldn't Exim mention this timeout in the log
though instead of the 'returned 2' error?
Thanks,
Simon Barr
Systems Engineer
Chelsing Assemblies Ltd
Tel: 01992 554-566
Fax: 01992 553-644
E-mail: simon.barr@???
