Author: Peter Bowyer Date: To: Exim Users Mailing List Subject: Re: [exim] Freezing non relay_to_domains
On Wed, 23 Feb 2005 10:00:32 -0800, Matt Hemingway <matt@???> wrote: > Hi. I didn't see anything in the archives about this.
>
> We have exim setup as a smart host which relays incoming mail for 3
> of our domains to an internal mail server. "domainlist relay_to_domains
> = " is set to these domains and "hostlist relay_from_hosts = " is set
> to our internal IP addresses so we can send mail out. In the router
> section we have the appropriate routers set up for these 3 domains to
> forward mail to our internal mail server.
>
> When I look in /var/spool/exim4/msglog/, I see a grip of frozen
> messages to addresses having nothing to do with us. Here's an example
> of one (with the domain and IP address changed):
>
> 2005-02-20 19:25:41 Received from <> R=1D34CW-0008Lg-Ce U=mail P=local
> S=4399
> 2005-02-20 19:25:41 sciyzzoecqig@??? R=dnslookup
> T=remote_smtp defer (-5 3): retry time not reached for any host
>
> 2005-02-21 10:14:14 mail.notourdomain.com [123.123.123.123]: Connection
> timed out
> 2005-02-21 10:14:14 sciyzzoecqig@??? R=dnslookup
> T=remote_smtp defer (11 0): Connection timed out
These are frozen bounces, due to attempts to deliver to nonexistant
addresses at your domains. The sender is also nonexistant, so the
bounce freezes.
To avoid this, implement recipient checking at SMTP time. You can do
this in your RCPT ACLs. It's trivial for locally-delivered domains,
and for your relay domains, you can do it with recipient verify
callouts, provided you can make the destination server do SMTP-time
recipient checking itself.
