On 2014-07-08 at 11:46 +0200, David Saez wrote:
> Yes, i can also switch off the server or not run exim at all, both
> of them will work well
What sort of solution are you looking for, with Exim?
You have a diagnostic message, warning you about something that's
potentially worthy of attention if you're looking at the logs to see
what happened for an email, and the message is arising from your use of
a particular option in combination with current configured resource
limits.
Diagnostic messages explaining why Exim is not doing a particular
delivery are logged. There are a wide variety of causes and it's not
completely trivial to just not log one particular cause.
You can probably reduce the occurrence rate by setting
`queue_smtp_domains` to a hostlist matching the remote hosts, or `*` for
all hosts (or start Exim using `-qq` instead of `-q`); this will cause
Exim to not immediately try to deliver remote email, but instead do the
routing and drop the message on the queue, so that a subsequent
queue-runner will deliver it, and this increases the chances of
connections being reused to deliver multiple emails down one SMTP
connection. You'll probably want to tune the rate at which you start
queue-runners; eg, every minute.
If you want cleaner logs, the normal solution is to ship your logs into
a log aggregation system which can process the log lines and provide
the ability to drill down; you can also just drop lines completely when
you do that.
There are a number of such systems available, commercial, open source and
cloud services. Cloud: Sumo Logic, Loggly, Papertrail. Commercial: Splunk.
Open source: logstash, or roll your own.
--
My employer, Apcera Inc, is hiring sysadmin; primarily San Francisco:
http://www.apcera.com/jobs/#operations-engineer
(but all the mistakes in this email are made in my personal capacity)
