Re: [exim] (110): Connection timed out

Páxina inicial
Borrar esta mensaxe
Responder a esta mensaxe
Autor: Rob Gunther
Data:  
Para: Exim Mailing List
Asunto: Re: [exim] (110): Connection timed out
>
> Rob Gunther <redrob@???> (Do 13 Nov 2014 02:02:45 CET):
> …
> > 2014-11-12 15:59:22 1XoJ6d-0006S3-Io exchange.lahore.com [82.71.51.200]
> > Connection timed out
> > 2014-11-12 15:59:22 1XoJ6d-0006S3-Io == shahid@??? R=ik_router
> > T=remote_smtp defer (110): Connection timed out
> >
> > A manual connection to the server via telnet works, I can connect and
> > deliver mail.
>
> Manual connecton from the server in question, I suppose.
>


Yep, a manual connection from the server in question works. Both to the
end users server and another server that we control at a different
datacenter. Both of those connections work fine manually and it is not
like the server can't send any email... it sends tons of mail every day and
this is the only domain we know of that is having an issue.


>
> > In rare situations where there is a routing issue or firewall issue I can
> > clear the queue for a specific domain by routing mail pending for the
> …
> > # move mail to another server if it is stuck on this server for some
> reason
> > clear_queue:
> > driver = manualroute
> > condition = ${if match{$domain}{lahore.com} }
> > route_list = * "122.103.250.10"
> > transport = remote_smtp
> > no_more
> >
> > That normally will just direct any mail to another server and it is
> > processed and delivered from there.
> >
> > This one specific domain, when I try the clear_queue router - it will
> > report connection timed out. I know there is no firewall rules as I
> control
> > both networks, no ISP block and these servers communicate all day long.
>
> Just for my understanding, the clear_queue router and the associated
> remote_smtp transport time out during the connection to the other data
> center?
>
> But the log lines above show the connection attempt from your server to
> the mx in question.
>
> So we got two issues: Why does the connection to the original MX time
> out? AND why does the connection to the other data center time out?
>
> BTW, the MX for lahore.com is not exchange.lahore.com but something
> else. Maybe a special case handled by your ik_router
>
>

We accept mail for the domains, process it and quarantine spam - then
forward the mail on to the users actual server. Normally clean messages
are handled by the remote_smtp transport after we set the destination using
route_data in the router.

So that does not work - we get the timeout for domain in question. Trying
to use my clear_queue router, which is routing to another of our own
servers also gets a timeout when trying to redirect any email for this one
domain.

None of the messages in the queue for the domain are large, nothing over
200k from what I've seen. We have redirected all new mail for this
lahore.com domain to an alternate server and it is being processed there...
So something is funky with this one server or the domain/server/exim
combo... just need to figure out what.