RT: 3.6.5
MySQL: 5.0.27
OS: Fedora Core 5
Mail agent: Postfix with a procmail filter
Without going into too much detail one of our queues is set up with a
procmail filter that blocks email from one address while allowing
email from only two domains. The reason for this has to do with the
one address being a mailing list which was wreaking havoc with the
ticketing systems mail processing. I don’t recall the details but this
was the solution. That said, I believe it is working proplerly as the
.procmail/log file is showing all kinds of happy goodness going on.
Every now and then an email that is sent won’t be added to the ticket
which it is meant for. Procmail indicates that it has been passed on
to rt-mailgate but nothing relevant shows up in the RT log:
Postfix log:
May 12 08:12:25 rt postfix/smtpd[28085]: connect from unknown[192.168.7.52]
May 12 08:12:25 rt postfix/smtpd[28085]: F3B512E004C:
client=unknown[192.168.7.52]
May 12 08:12:26 rt postfix/cleanup[28083]: F3B512E004C:
message-id=A8BB894EBE311F479369F18EDCC1748F1856D84A@EX2K7VS02.4emm.local
May 12 08:12:26 rt postfix/qmgr[4562]: F3B512E004C:
from=xxxx@xxxxxxxxxxxxx.com, size=74002, nrcpt=1 (queue active)
May 12 08:12:26 rt postfix/smtpd[28085]: disconnect from unknown[192.168.7.52]
May 12 08:12:26 rt postfix/local[28089]: F3B512E004C:
to=faclog@rt.ops.xxxxxxxxxxx.com, relay=local, delay=1, status=sent
(deliver
ed to command: /usr/bin/procmail)
May 12 08:12:26 rt postfix/qmgr[4562]: F3B512E004C: removed
May 12 08:12:27 rt postfix/pickup[27969]: 223402E004D: uid=48 from=<>
May 12 08:12:27 rt postfix/cleanup[28083]: 223402E004D:
message-id=20100512121227.223402E004D@rt.ops
May 12 08:12:27 rt postfix/qmgr[4562]: 223402E004D: from=<>,
size=75587, nrcpt=1 (queue active)
May 12 08:12:27 rt postfix/smtp[28084]: 223402E004D:
to=xxxx@xxxxxxxxxxxxx.com, relay=216.12.128.78[216.12.128.78],
delay=1, statu
s=sent (250 Ok: queued as BEAD1A7EE)
May 12 08:12:27 rt postfix/qmgr[4562]: 223402E004D: removed
Procmail log:
From xxxx@xxxxxxxxxxxxx.com Wed May 12 08:12:26 2010
Folder: /usr/local/rt3/bin/rt-mailgate --queue facilities-logistics 73030
lockfile: Try praying, giving up on “/var/mail/faclog.lock”
RT Log:
[Wed May 12 08:11:42 2010] [warning]: Encode::Guess failed: ; fallback to
iso-8859-1 (/usr/local/rt3/lib/RT/I18N.pm:412)
[Wed May 12 08:14:40 2010] [crit]: Apache2::RequestIO::rflush: (103) Software
caused connection abort at /usr/lib/perl5/vendor_perl/
5.8.7/HTML/Mason/ApacheHandler.pm line 1029 (/usr/local/rt3/bin/webmux.pl:127)
As you can see, RT didn’t do anything between 08:11 and 08:14. Looking
at the Apache error log doesn’t show anything happening either. Just a
repeat of the Apache2::RequestIO::rflush error above and another one
hour prior. Unfortunately, we didn’t have debugging on so I don’t have
anything more helpful to look at.
Until this happens again with debug logging on I don’t have anything
else to provide or look at. However until then, if anyone else has
experienced similar events and could give me a pointer to look at I’d
appreciate it.
-Mathew