On Sat, 15 Aug 1998 15:24:36 +0200 (CEST), Walter Haidinger
<walter(_dot_)haidinger(_at_)gmx(_dot_)net> wrote:
AFAIK using load averaging would have the first messages delivered and the
rest queued. However, also not being a sendmail guru, I do not know how to
empty a sendmail queue for incoming mail only.
It would work on emptying the queue whenever the load average was low
enough, which would hopefully be most of the time as long as not too
many Sendmails are running (assuming you set the cutoff at a suitable
level). To me, this sounds like exactly the sort of solution you were
looking for.
Is there some kind of "postprocess" rule/command with is run after all
mails are processed by procmail? I do not remember reading of such at all.
The TRAP= pseudo-variable, but I doubt this is the way to go.
(Sounds like you're begging for trouble. Why not use fetchmail's
multidrop feature? See also the standard gripe about how Procmail is
not really the right tool for this in the FAQ; look for the string
"virtual domain". <http://www.iki.fi/~era/procmail/mini-faq.html>)
All mail in my mailbox is for me. That is, it is a single user mailbox.
Now, what do I really do?
Collecting mail from several university accounts and having some robots
running there too, I distributing all mail to distinct users on my linux
machine by looking at the subject and the sender.
IOW: The mailbox at my ISP is sort of a bottle-neck. Mails from different
locations are collected there and distributed to several user accounts on
a single machine. All those user accounts are owned by me.
Does multidrop apply here too?
I thought it is a mailbox where mails for different users are sent to.
For me, there are multiple senders but just on receiver (me at
haidinger(_at_)pgv(_dot_)at)
I'm not sure if we're speaking the same language. A typical problem is
that a user sets up sue(_at_)home-computer, ed(_at_)home-computer,
joe(_at_)home-computer and then has home-computer set up to have the
upstream deliver all mail to this domain dropped in a single mailbox,
which gets downloaded to home-computer whenever it happens to connect
to the ISP. Something (like for instance Procmail or Fetchmail) on
home-computer is then supposed to figure out which mail was sent to
sue, ed, or joe, typically from looking at the headers, and typically
working well enough that the operator of home-computer doesn't even
realize that there is a problem. And the problem is that the
information about who a message really is for is in the envelope,
which would typically get thrown away at the upstream when the message
was delivered to home-computer's one monolithic mailbox. This matters
e.g. when users on home-computer get BCC:ed mail, or mailing lists
(such as the one you're reading now). Otherwise, you can usually
deduce the envelope recipient with some level of confidence from the
headers, although at least in theory, it's perfectly possible to end
up getting some of ed's mail delivered to joe.
If this is not what you're doing, never mind.
Incidentally, I believe Fetchmail's multidrop essentially does the
"Received: for" header trick (see the FAQ again) augmented with some
additional heuristics. (There's an option to tell Fetchmail something
about what heuristics to use but I still believe its multidrop
delivery mechanism is all essentially based on "educated guesswork"
and not guaranteed to get it right all the time.) (Can somebody who
actually uses Fetchmail confirm or deny this?)
/* era */
--
Bot Bait: It shouldn't even matter whether (`') Just (`') http://www.iki
I am a resident of the State of Washington \/ Married! \/ .fi/~era/