[milters] Archive

Lists Index Date Thread Search

Article: 305
From: Anthony Howe
Date: 2005-01-11 05:42:16 -0500
Subject: Re: performance of milter-ahead?

Removal...........: milters-request@milter.info?subject=remove
More information..: http://www.milter.info/#Support
--------------------------------------------------------

Anke Breeuwsma wrote:

>>If you're just talking about the regular creation and deletion of 
>>accounts as customers come and go, then I wouldn't bother using a low 
>>TTL. Just let them expire normally so your system can benefit from 
>>longer TTL for the regular users.
> 
> 
> I know, deletion is not my problem. My problem is that customers know their email 
> addresses before the system makes those addresses active. So, it's possible that 
> they're trying to use their address before it's active and then they become
blacklisted for 
> x days. 
> 
> Is it possible, like milter-gris, to have a different ttl for temporary fail entries
and for 
> accepted entries stored in the cache?

It would be possible to add something like that, but that assumes that 
you return a temporary failure. If an address doesn't exist, Sendmail 
returns a permanent failure.

Hmmm. An off the top of my head idea for milter-ahead (and maybe other 
cache situations with my other milters). Maybe an exponentially 
incrementing TTL might be the trick. The milter could start with a short 
TTL, like 60s. After the short TTL elapses, the next message would 
trigger a retest: failure would double the short TTL, success would 
change its status, this would keep going on until max. TTL is reached.

Might this be if interest to yourself and other users?

-- 
Anthony C Howe                                 +33 6 11 89 73 78
http://www.snert.com/       ICQ:
7116561         AIM: Sir Wumpus

            "Once...we were here."  - Last of The Mohicans


Lists Index Date Thread Search