[Mailman-Users] Mailman under Fedora Core 2

[Rich West]

I migrated our main server from Fedora Core 1 (running our own built 
version of mailman 2.1.5) to another server running a fresh install of 
Fedora Core 2 (running our own built version of mailman 2.1.5) about two 
weeks ago.

Since that time, we have had continual problems with our mailman 
installation.

Our installation is pretty vanilla, and it handles about 30+ lists with 
some lists having low membership and others having hundreds of addresses.

Anyhow, on the FC1 machine, everything ran fine since we rolled that in 
to production back in January of this year, and previously, under RedHat 
9, things ran seamlessly as well.

Since we have upgraded to Fedora Core 2, we're getting some odd problems 
that are just frustrating the living daylights out of me.

The problem shows itself when we get a report of no activity on any of 
the lists.  Quick investigation shows that /var/mailman/qfiles/out is 
full of pending messages.  Unsure as to why they are pending, we quickly 
restart the mailman daemon.  Unfortunately, two of the processes don't 
die ("mailmanctl -s start" and "OutgoingRunner"), so they have to be 
killed manually.  Bringing mailman back on-line begins the processing 
again, but I am not certain that it is even processing correctly.

On our test list, it does seem to send to *some* of the test users, but 
not all.

The smtp-error log periodically shows a gazillion "Low level SMTP error: 
(111, 'Connection refused')" messages followed by a gazillion "Low level 
smtp error: please run connect()first" and finally a gazillion "delivery 
to <emailaddress> failed with code -1: please run connect() first"

I've gone through and insured that my mm_cfg.py contained all of the 
right information, and I have even gone through the effort to compile it 
up (not sure if this is even necessary, but I am running out of ideas) 
via : python -c "import compileall; 
compileall.compile_dir('/var/mailman/Mailman', ddir='/var/mailman/Mailman')"

The mail server is working fine for all other types of delivery 
(confirmed easily).  I'm at a complete loss at this point.. Is there any 
way to turn up the debugging information to see what host it is getting 
the "Connection refused" message from?

Also, I'm monitoring the qfiles/out directory.. I've been watching it 
for about a half hour and have watched the directory's contents 
continually grow after a fresh restart of mailman.  It just doesn't seem 
to send. :(

Argh!  Totally frustrating as we never had any problems with mailman 
before..  Any help, ideas, comments, etc, would be wonderful!!

-Rich