Re: [Exim] SIZE param for callbacks?

Pàgina inicial
Delete this message
Reply to this message
Autor: Alan J. Flavell
Data:  
A: Exim users list
Assumpte: Re: [Exim] SIZE param for callbacks?
On Mon, 6 Oct 2003, Kevin P. Fleming wrote:

> Chris Edwards wrote:
>
> > So it doesn't seem too much of a stretch to attempt to check we can send a
> > bounce of a plausible SIZE ?
>
> I believe this has been discussed before, and it's a bit risky.
> Over-quota conditions tend to be transient (at least for normal
> mailboxes anyway), and the callout caching would interfere because it
> wouldn't retry the callout soon enough after the over-quote had been
> cleared.


Just so we understand each other: there'd be no automatic retry (the
mail would be refused with a 5xx). The issue is what happens when the
would-be sender clears their inbox and then tries to send the mail
again, agreed?

Yes, as you say, our exim would have cached the response, which could
result in a bona-fide sender *still* getting refused after fixing
their problem.

> If you lower the callout cache expire time, then you generate
> many more callouts and that's a waste of resources (on both ends).


You have a very reasonable point, I don't deny that. On the other
hand, as Chris says, it's a not unreasonable policy to want to be able
to send subsequent status reports, so it's not entirely unfair to
insist that senders keep space available in their inbox if they hope
to send outgoing mails.

I must say, I was quite intrigued when Chris brought this observation
to my attention. It had never occurred to me to try the callback with
SIZE. I can't help being curious as to what proportion of MTAs would
make that distinction in their response.