On Wed, 24 Mar 2004, Niall Mansfield wrote:
> Our Exim queue has jumped from a usual length of 1 or 2,
> to approx 100, consistently. This has happened in the last day or so.
> (Also, the no. of Exim processes jumped to over 200.
> and so we changed exim.conf:
> queue_only
> queue_run_max = 40
> for now, at least.)
>
> * Why would the queue suddenly expand?
> * How do you go about working out why a message spends so long in the
> queue? (Even a message from me to another internal user?)
There should be enough information in the Exim log files to get you
started. If you know the message id of the problem message, you can use
exigrep to fish out all the log lines that pertain to that particular
message.
When you set queue_only, you caused any incoming messages to wait for
the next queue runner process. The delay will depend on how often your
queue runners are started.
Setting queue_run_max=40 is unlikely to have any effect. It controls the
maximum number of queue-runner processes that can be simultaneously
active. Each queue-runner delivers just one message at a time. Unless
you have messages that really do take forever to deliver, I doubt very
much whether you will ever get more than one or two queue-runners active
at once. Only if one hasn't finished before the next one starts does the
number exceed one. Thus, if you start them every 15 minutes (a typical
kind of time), the first one will have to take over 10 hours to process
the queue to still be alive when the 40th one is started.
Philip
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
Get the Exim 4 book: http://www.uit.co.uk/exim-book
