DragonFly kernel List (threaded) for 2007-03
DragonFly BSD
DragonFly kernel List (threaded) for 2007-03
[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]

Re: first stab at simple mailer

From: Bill Hacker <wbh@xxxxxxxxxxxxx>
Date: Fri, 23 Mar 2007 00:25:17 +0800

Simon 'corecode' Schubert wrote:
Rahul Siddharthan wrote:
But you still need to efficiently queue the mail for forwarding. You
can't just make separate connections to the target for each recipient.
qmail does it...
The claim is that making separate connections is usually faster (fewer
round-trip delays) because of how SMTP works.  It does use more
bandwidth but this is unlikely to be significant on most sites,
most of the time.

These days however, when people send attachments megabytes in size (or
viruses do it for them), I don't know whether that claim is still

1. I think it *is* unfair towards the receiving MX to send the same mail multiple times.

2. I think it is *negligible* for locally generated, private + cron mails.

3. not trying to aggregate mail delivery on one MX makes the code *SO MUCH EASIER*. seriously. think about what you need:

a. central coordination instance
b. MX/A lookups for all domain parts of all mails
c. choose one MX
d. now find all mails which could potentially be sent there
e. weed out rejects
f. start over

point is, you can't say "hey, they have the same MXes", because different domains could share only one MX out of three, for example. you have to check for every MX connect which other mails could be delivered there. definitely not a necessary option for the first round.


Feel free to use these addresses for testing your queing and retry rules on multiple recipients et al:





Typical daily logs are a good test send, as like it or not - some can be perceived to have spammish characteristics.

The postmaster address should be found to accept domain-literals, i.e. addressed to postmaster@, IF there are no other addressees on such a message (illegal for bounces). The others will not.

- it is a server dedicated primarily to MTA R&D, (Exim at present) and I can give you log extracts if the rejection messages you will certainly get are not specific enough.

It should become apparent 'real soon' that building a reliable MTA in the 'spam age' - one that doesn't resemble a zombot - is a seriously non-trivial exercise.



[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]