That is the issue, the receiving server can receive more than what I am sending to it. When I run "postsuper -r ALL; postix flush" I can watch all the queued messages get delivered within a couple minutes no problem. Netstat shows the allotted amount of connections defined in the custom smtp transport open, and messages fly off my server. The bottle neck does not seem to be on the receiver, it seems to be the active queue on the sending host.
On Tue, Feb 16, 2016 at 10:44 AM Travis Dolan <travis.do...@gmail.com> wrote: > Unfortunately there are many MTAs/MUAs that send to my host, which then > sends to the custom smtp transport destination. Building a hash map for all > potential hosts would be problematic. > > I need to allow as many messages as possible IN, and then relay those > messages as fast as possible to my custom smtp transport destination. > > There must be a way to requeue messages in my active queue faster. > > Thanks in advance. > > > On Tue, Feb 16, 2016 at 10:14 AM Wietse Venema <wie...@porcupine.org> > wrote: > >> Wietse Venema: >> > Travis Dolan: >> > > The destination server has no issue with the amount of messages we >> send it. >> > > We have custom a python script accepting port 25, and simply takes >> > > everything and places in RabbitMQ. The incoming queue for this >> destination >> > > is highly active, which is really the root issue. I need to move those >> > > messages out as fast as they come in. >> > >> > Have you considered slowing down the injection rate, so that it >> > better matches the output rate and does not congets the queue. >> >> Corrected example: >> >> /etc/postfix/main.cf: >> smtpd_client_restrictions = check_client_access >> hash:/etc/postfix/access >> >> /etc/postfix/access: >> 1.2.3.4 sleep 2 >> >> This assumes the default setting of smtpd_delay_reject=no. >> >> Wietse >> >>
