On 01 Sep 2020, at 07:37, Stephan Seitz <stse+post...@rootsland.net> wrote:
> On Di, Sep 01, 2020 at 07:18:44 -0600, @lbutlr wrote:
>> It could be as simple as the remote severing taking very slightly longer to
>> process for reasons on its end (slower lookup into its tables from the
>> secure ports, for example).
>
> I used port 25 for both tests, so that I could hope to avoid different
> options for smtpd and submission.
>
> So the only difference in the SMTP session should be the AUTH command.
Do you control the target server? If not, then you have no idea what mechanisms
exist on that server.
Is this consistent across many servers? Does the delay scale based on the
number of messages (is 1000 message 94 seconds versus 60? How about 5000
messages, 470 seconds versus 300?
Obsessing over a delay of 3 tenths of a second on a remote server you have no
control over is going to lead to, probably, nothing but frustration. That said,
sure, it COULD be something on your end, but you first would have to find what
that delay in your logs, and unless it's really obvious (4 second pause every
100 messages, for example), that's going to be very very hard.
--
"One of the great tragedies of life is the murder of a beautiful
theory by a gang of brutal facts." - Benjamin Franklin
