[Mailman-Users] mailman problem with local sender
gsutter at pobox.com
Tue Feb 1 00:33:48 CET 2000
On Thu, Jan 27, 2000 at 08:48:04AM -0800, Joseph Scott wrote:
> Gregory Sutter wrote:
> > I have a problem with my freshly-installed Mailman 1.1. When a mail is
> > sent to a list, it goes through properly. Everything works fine.
> > *Unless* the sender is on the list host. In that case, a duplicate
> > is also made and is held in sendmail (8.9.3)'s queue until mailman's "post"
> > script is killed. At that time the duplicate is sent out to all list
> > members and the zombie processes (one sh, one python) that were being
> > held open finally get cleaned up.
> > Python 1.5.2 (#2, Jun 10 1999, 15:33:16) [GCC 188.8.131.52] on freebsd3
> > List of processes being held open:
> > 44829 ?? Is 0:00.01 sendmail -oem -oi -N failure -R full -- test at foo.org
> > 44830 ?? Is 0:00.00 sh -c /usr/local/mailman/mail/wrapper post test
> > 44831 ?? I 0:00.22 /usr/local/bin/python -S /usr/local/mailman/scripts/post test
> > 44832 ?? Z 0:00.00 (python)
> > 44833 ?? Z 0:00.00 (sh)
> > Sendmail logs from the message:
> > Jan 27 00:31:17 <2.6> soivah sendmail: AAA44827: from=gsutter, size=456,
> > class=0, pri=30456, nrcpts=1, msgid=<20000127003117.A42391 at soivah.foo.org>, rel
> > ay=gsutter at localhost
> > Jan 27 00:31:17 <2.6> soivah sendmail: AAB44827: clone AAA44827, owner=te
> > st-admin
> > Jan 27 00:31:18 <2.6> soivah sendmail: AAA44840: from=<test-admin at foo.or
> > g>, size=924, class=-60, pri=138924, nrcpts=1, msgid=<20000127003117.A42391 at azaz
> > el.foo.org>, proto=ESMTP, relay=localhost [127.0.0.1]
> > Jan 27 00:31:20 <2.6> soivah sendmail: AAA44840: to="|IFS=' ' && exec /us
> > r/local/bin/procmail -f- || exit 75 #gsutter", ctladdr=<gsutter at foo.org> (1001/
> > 20), delay=00:00:02, xdelay=00:00:02, mailer=prog, stat=Sent
> From this last log entry here it shows that you are running procmail
> via a .forward, I wonder if procmail may be doing something strange.
> On the machine that I've got we're using procmail for local delivery,
> not via a .forward
This can't be it, because the problem occurs before procmail ever gets
called. In the logs above, the message is cloned and queued before any
type of delivery takes place. It's pretty puzzling.
Gregory S. Sutter Computing is a terminal addiction.
mailto:gsutter at pobox.com
PGP DSS public key 0x40AE3052
More information about the Mailman-Users