[exim-dev] [Bug 1885] Mail heisenbounces. Probable cause: IP…

Top Page
Delete this message
Reply to this message
Author: admin
Date:  
To: exim-dev
Subject: [exim-dev] [Bug 1885] Mail heisenbounces. Probable cause: IPv6 and/or lack of encryption.
https://bugs.exim.org/show_bug.cgi?id=1885

--- Comment #7 from Thaddeus H. Black <thb@???> ---
I wrote:

> Please allow two weeks.


The two weeks have passed. I am not yet ready to report, but will reply again
within another two weeks.

--- Comment #8 from Thaddeus H. Black <thb@???> ---
This is an update. No action is requested.

I wrote:

> Please allow two weeks.


Then, I wrote:

> The two weeks have passed. I am not yet ready to report, but will reply
> again within another two weeks.


For the record, let me explain what is going on. (You are not required to be
interested.)

1. For me, it has been tricky to set up Exim to debug-log as requested on both
ends -- without otherwise modifying the configuration being tested and without
seriously interrupting normal mail handling. This has taken numerous attempts
and has made for an interesting learning experience! At any rate,
debug-logging is now underway.

2. Unfortunately, since debug-logging has been underway, no heisenbounces have
occurred.

3. I have tried shortening DNS TTL from 24 hours to 5 minutes, waiting a
couple of days to allow the shortened TTL to propagate. I am not sure, but
shortening the TTL had no effect I could discern. (The TTL is now again 24
hours.)

4. As a trial, I switched off debug-logging. An unlogged bounce did indeed
then occur.

5. Now, I am trying debug-logging on one end at a time.

Once I have isolated an instance of a heisenbounce and have debug-logged it on
both ends, I will post the logs; but this is a patient business. I have not
yet learned to produce bounces at will, but must wait for bounces to occur on
their own, perhaps twice a week.

I will report again as soon as I have something to report, or in any case by
October's end.

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