Nigel Kukard wrote:
>You're very right in your analysis, the delegation protocol doesn't 
>really give us enough information to only count successful deliveries.
>
>The mail counter is increased during the RCPT stage, so if 100 
>recipients are received, and the 99th one fails in such a way that 
>Postfix terminates delivery the counter is already increased.
>
>Deferring the counter update to the DATA stage may not be sufficient 
>either, what happens if 100 connections are made for 100 recipients 
>each, the counters are deferred to DATA and all are updated at the same 
>time.
>
>I'm not sure if one can get around this. What about changing the 
>destination concurrency limit?

We don't (in the general case) have control over the servers sending mails 
through us - this is a setup specifically for customers to send/receive thei 
mail.

Deferring the counter updates would work for me, though I can see how it might 
not work for others. With my setup, the theoretical maximum "overshoot" would 
450 recipients - 9 connections*, 50 recipients per message. The result would be 
that the customer would be massively over-limit and their mail would pause for 
a while.
* 3 servers at the moment, each with asmtpd_client_connection_count_limit of 3

As it is, if I increase the limits sufficiently then mail will flow - but they 
are a bit higher than I'd prefer. Both in terms of the message rate, and the 
allowable errors before Postfix drops the connection. The only other way round 
it I can see would be to have high/low marks - so when the high mark is reached 
then deliver stops altogether, and doesn't restart until the low mark is 
reached. The effect would be that mail would come in in spurts. I guess that 
would need extra fields in the tables, and extra logic - beyond my coding 
abilities I think (but I'll have a quick look at the code and see).

Attachment: smime 19.p7s
Description: mailforge

_______________________________________________
Users mailing list
[email protected]
http://lists.policyd.org/mailman/listinfo/users_lists.policyd.org

Reply via email to