Author: Drav Sloan Date: To: Edward Jakosalem CC: Exim-users Subject: Re: [exim] MailScanner/Exim Long Queue Delay
Edward Jakosalem wrote: > Hi there,
> This maybe a bit off topic but our of desperation, I am sending it anyway.
> I have been searching in the MailScanner and Exim archives for a
> solution to my problem. I have found some suggestions but none of them
> worked. So here it is.
> I have 3 incoming mail gateways hosting about 60K users in multiple
> domains. Each process a max of 200K messages a day. For the last
> months, these pesky spams have been pestering our servers causing long
> delays in the delivery of emails. I have no choice but to free up the
> queue using exim -q.
If the messages are arriving in groups of recipients >10 then you may
want to set smtp_accept_queue_per_connection to a higher value to allow
more messages to be processed directly rather than queued on connections.
Also do you run the exim process from inetd or in daemon mode? If from
inetd, then make sure you also start seperate exim -q<time val> processes
(such exim -q30m). If you run in daemon mode make sure you add the -q<time>
option or again run seperate processes to handle queued mails in a timely
fashion. The time you use for the q option you will have to tweak, too low
and more queue runners will be started before the previous ones have completed
and too high and the queue will continue to grow and not be processed.
