On Mon, 9 Sep 2002, Philip Hazel wrote:
> On Mon, 9 Sep 2002, Alan Thew wrote:
> > 2002-09-07 06:57:53 17nTq0-00056z-00 failed to get IP address of stdin \
> > when \
> > socket passed from previous delivery: Transport endpoint is not connected
>
> This is now understood, and I've made some changes for 4.11 which should
> improve things.
>
> Executive summary: Exim delivers a message successfully, notices that
> there is another message waiting for the same host, so starts a new
> delivery process, passing it the SMTP connection. The new process routes
> the addresses in the new message - but there's a delay on one or more of
> them (DNS timing out, typically). By the time Exim comes to try to use
> the SMTP connection, the server has timed out.
I have seen this error a lot recently - I route email to an ISP smarthost,
and it seems that that host is not very keen to keep the SMTP connection
open for very long. Thus full queues have tended to hang around for a
long time (we are on an ISDN line and only dial up every hour).
I have since upgraded from 3.36 to 4.10, which gets rid of the paniclog
messages, but I don't know what is happening to the email. If an Exim
process finds that the SMTP connection it has been passed has been
dropped, will it merely die gracefully (without delivering the mail) or
will it make a new SMTP connection and send the mail?
Should I just wait for 4.11 and look at the changelog?
John
--
It would have been jolly to talk like this, and really, it wasn't much
use having anything exciting like floods, if you couldn't share them
with somebody.
