Re: [exim] syscall: Connection reset by peer (outlook.com)

On 2021-02-27 21:31, Heiko Schlittermann via Exim-users wrote:
> MRob via Exim-users <exim-users@???> (Sa 27 Feb 2021 21:38:09
> CET):
>> Hello,
>> I'm getting lots of these in exim log:
>> 2021-02-25 04:31:15.790 [19189] SSL_write: (from
>> mail-pu1apc01hn2214.outbound.protection.outlook.com (APC01-PU1-
>> obe.outbound.protection.outlook.com) [52.100.183.214]:42879) syscall:
>> Connection reset by peer
>
> I assume, that a connection coming *from* outlook to you.

Yes I should have been clearly

> They (outlook)
> seem to have the bad habit to forcefully tear down the connection as
> soon as they received your "250 OK". They don't send a QUIT, nor do
> they
> perform a regular SSH "good bye" handshake. They just behave rude.
>
> Your server tries to bring the connection down, but all your side's
> politeness is *rejected* by the other (outlook) side.

that is a relief, thank you and Jeremy. But see below, as it seemed to
create a real delivery problem

>> I also get plenty of successful mail from outlook.com so its
>> confusing. But
>> some remote sender using microsoft reported getting mail returned due
>> to
>> expired retries/unable to connect.
>
> That can be related, but doesn't have to be related.

I found this situation because single user forwarded microsoft bounce
message to their user said:

Generating server: AM6PR05MB6309.eurprd05.prod.outlook.com
Total retry attempts: 26

user@???
Remote Server returned '550 5.4.300 Message expired -> 451 Temporary
local problem - please try later'

So checking my log I found the last attempt to deliver this message
resulted in this situation where outlook.com immediately dissconnected
(its how I learned about this "connection reset"). Reading Victor, is it
important to say this disconnect seems to happen before mail
transmission, yes, probably after "250 OK". In this case is "451
Temporary local problem" a problem on the microsoft outbound server?

So could I conclude some microsoft servers have strange connection
problem and it is coincidence a message like this is retried every time
from a server having the problem? I only assume Microsoft have enough
functional servers it would never happen