[Mailman-Users] Why some messages aren't distributed?
mary at bgumail.bgu.ac.il
Mon Mar 17 13:18:18 CET 2003
Yes, the defaults are exactly as you write except for SMTP_MAX_RCPTS
set to 400.
I'm reluctant to try "DELIVERY_MODULE = 'Sendmail'" because of
so many warnings of possible problems - is it worth trying
On 16 Mar 2003, Jon Carnes wrote:
> What are your "Delivery defaults" set to in ~mailman/Mailman/mm_cfg.py
> (or Defaults.py)?
> #DELIVERY_MODULE = 'Sendmail'
> DELIVERY_MODULE = 'SMTPDirect'
> # MTA -- but then also see POSTFIX_STYLE_VIRTUAL_DOMAINS.
> MTA = 'Manual'
> POSTFIX_STYLE_VIRTUAL_DOMAINS = 
> POSTFIX_ALIAS_CMD = '/usr/sbin/postalias'
> POSTFIX_MAP_CMD = '/usr/sbin/postmap'
> SMTP_MAX_RCPTS = 500
> SMTP_MAX_SESSIONS_PER_CONNECTION = 0
> MAX_DELIVERY_THREADS = 0
> SMTPHOST = 'localhost'
> SMTPPORT = 0 # default from smtplib
> Have you tried running with
> DELIVERY_MODULE = 'Sendmail'
> On Sun, 2003-03-16 at 10:50, Marina Markus wrote:
> > Hello and thank you,
> > But I checked both places and found nothing:
> > > Check to see if the messages have been shunted:
> > > ~mailman/qfiles/shunt/..
> > I didn't find those specific undistributed messages
> > in this directory, although found some older ones.
> > May I ask what this "shunt" directory serves for?
> There are some archive messages from Barry that explain it.
> I believe they are for messages that take too long for Mailman to
> process. There might be a problem with their headers or with the
> system. In any case, rather than let the message clog up the queue,
> Mailman gives up on the message and shunts it.
> Check out the ~mailman/bin/unshunt command
> > >
> > > Also check on your MTA and see if it is using your fully qualified
> > > domain name or merely a subset of your domain name (like fred instead of
> > > fred.mynet.com).
> > >
> > Yes, it uses a fully qualified domain name "lists.bgu.ac.il".
> > Can there be a situation when "mailmanctl" dies by itself
> > and then starts anew by itself? This could be an explanation,
> > although I can't see how this could happen.
> > Regards,
> > -Marina
> Something similar can happen.
> One of the features of version 2.1 is that it times it's processes as a
> last resort to figure out if a process is hung somehow. As an example,
> the Archiver qrunner has a specified number of minutes to handle each
> message, if it exceeds that time limit then it is killed and restarted
> after all the other qrunners have been given another go.
> (at least that is how I read the code...)
> HtH - Jon Carnes
> BTW: are you running any Spam/virus scanning in conjunction with your
More information about the Mailman-Users