[exim-dev] [Bug 2665] SMTP timeout after sending data block

Top Page
Delete this message
Reply to this message
Author: admin
Date:  
To: exim-dev
Old-Topics: [exim-dev] [Bug 2665] New: SMTP timeout after end of data
Subject: [exim-dev] [Bug 2665] SMTP timeout after sending data block
https://bugs.exim.org/show_bug.cgi?id=2665

Phil Pennock <pdp@???> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pdp@???


--- Comment #10 from Phil Pennock <pdp@???> ---
Decoding RTF files and puzzling a lot, is this a correct summarisation?

1. You are running Exim 4.94
2. For the domain gloucestershire.pnn.police.uk the MX is Message Labs, for
filtering, and they then forward onto you
3. You are running the service mailgate.gsi.gov.uk
4. For some mails, predominantly those from Amazon SES originally, when Message
Labs sends onto your server those mails, you see long timeouts
5. These GSI servers are in Vodafone hosting, so not on your premises

In addition:

6. Is it fair to say that these messages from SES are a bit larger than many of
the other external-sender inbound messages received through this filter?
7. Please confirm that the firewalls in your hosting environment are configured
to allow ICMP 3/4 ("Destination Unreachable", "fragmentation needed and DF
set") to pass through.

An unfortunate pattern is for firewalls to block all ICMP, claiming that this
improves security, while ignoring that Path MTU Discovery requires that certain
ICMP pass through. For "large packets seen first from TCP connection
initiator" (as opposed to "small request, big response" patterns such as most
HTTP), this causes connections to hang, but only when the packet sizes go up
and the sender and recipient are trying to figure out the correct MTU size to
use for packets.

--
You are receiving this mail because:
You are on the CC list for the bug.