--- Tony Earnshaw <ton...@billy.demon.nl> wrote:
lør, 13.08.2005 kl. 14.14 skrev email builder:
[...]
Yes. Use the current version of Courier, whose current version of
maildrop
reports an out-of-quota condition as a temporary failure condition,
Oh. Well... we are using Postfix, and I don't think we'll be changing
that
any time soon. :(
Anyone done something like this with postfix/maildrop?? Do people just
end
up eating this? I'm surprised no one has tried to do something about
this
before now (under other MTAs)...?
Postfix 2.2.5 -> maildrop 1.8.1 via a long chain on this LDAP-based rig:
postfix smtpd -> amavisd-new 2.3.2/Sophos/pre-queue proxy -> postfix smtpd
listener/lmtp -> dspam 3.5 daemon/smtp -> postfix smtpd listener/virtual
transport -> maildrop.
The first Postfix smtpd listener will not give a 250 or 4/5xx to the
connecting client until maildrop at the end of the chain has accepted or
refused the message, and will return to the connecting client whatever
failure message maildrop gives, permanent or temporary.
No backscatter, ever, for whatever reason. Using amavisd-new and dspam
as transparent proxies in this manner is NOT recommended, simply because
they are NOT transparent proxies (read the relevant docs), but they work
well enough in practice on this low-volume rig.
The point being, that maildrop can be configured as anti-backscatterdly
as you would wish, under Postfix (and the two other extras).
maildrop can? i think you mean postfix can, or is there some special config
options you are using? we run pretty high volume here, so waiting for the
very last application in our whole email server to give a delivery response
is not feasible - that scares the hell out of me, especially during our peak
;)
looks like we're stuck creating a little backscatter :(
5 messages in net.sourceforge.lists.courier-maildropRe: [maildropl] Quota bounces using m...
