Re: question on body_checks


question on body_checks
\ Leonardo Rodrigues Magalhães (14 May 2007)
. \ (Wietse Venema) (14 May 2007)
. . \ Leonardo Rodrigues MagalhÃ£es (14 May 2007)
. . . \ (Wietse Venema) (14 May 2007)
. . . \ Noel Jones (14 May 2007)
. . . \ Blake Hudson (14 May 2007)
. . . . \ mouss (14 May 2007)

7 msg	[Fwd: Re: postfix as spam filtering server]
2 msg	sender restrictions
3 msg	at what stage is message_size_limit applied?
6 msg	pipe and crashing commands
5 msg	postfix as spam filtering server
3 msg	send mail to users on the same domain but a dif...
2 msg	Map several domains to virtual accounts ?
11 msg	postfix and MySQL encrypted passwords (was 554 ...
37 msg	tcp_table query.
2 msg	Last word on epoll and Slackware 11
5 msg	Open relay by forwarding loop?
3 msg	virtual subdomains
3 msg	reject_non_fqdn_sender and reject_non_fqdn_reci...
2 msg	'554 Recipient address rejected: Relay access d...
14 msg	Bounces, stopping them entirely
1 msg	Postfix mailing list administrivia
11 msg	postfix dropping SMTP connection when receiving...
8 msg	verify.db and reject_unverified_recipient
5 msg	proxymap query

Subject:	Re: question on body_checks
Group:	Postfix-users
From:	mouss
Date:	14 May 2007

Blake Hudson wrote:
>> Despite my REJECT on body_checks (soft_bounce is no, have checked
>> that), some servers are still retrying every 5 minutes resending
>> messages that were REJECTED. And as most of my rules are for blocking
>> audio/video content for saving bandwith, i'm having bandwith problems
>> because of that. Instead of receiving once and rejecting and message
>> got bounced, seems some servers are retrying no matter of the REJECT.
>>
>> So, that's why i'm looking for somehow breaking RFC and closing the
>> connection as soon as possible. In that case, i would acchieve the
>> bandwith saving, i think, because i wouldnt need to receive the whole
>> message (with attachments) for sending the REJECT only at the end of
>> the session .....
>>
>>
> Is there a reason that you have not set the max message size to a value
> that accomplishes the same goal? When a remote server connects and
> issues the ehlo command, it will immediately be able to determine the
> maximum allowed message size without sending any other data. - The
> ultimate bandwidth saver.
>

OP is talking about clients that don't understand that 5xx is a REJECT
and is not a temporary failure. hoping that they will check the SIZE
argument in a EHLO response line is extremely optimistic.