Joe Laffey writes:
On Fri, 22 Apr 2005, Sam Varshavchik wrote:
Joe Laffey writes:
Here is what I see in the log:
Apr 22 00:01:23 raw courierd: started,id=002DCE56.4261BA37.00004CD1,from=<>,module=dsn,host=,addr=<[EMAIL PROTECTED]
mputer.com>
Apr 22 00:01:23 raw courierd: Waiting. shutdown time=none, wakeup time=Fri Apr 22 00:01:32 2005, queuedelivering=
37, inprogress=1
Apr 22 00:01:23 raw courierdsn: Problems injecting bounce - submit failed.
Apr 22 00:01:23 raw courierd: completed,id=002DCE56.4261BA37.00004CD1
Apr 22 00:01:23 raw courierd: Waiting. shutdown time=Fri Apr 22 00:06:23 2005, wakeup time=Fri Apr 22 00:01:32 20
05, queuedelivering=37, inprogress=0
s_trap is a spamtrap setup via an alias map pointing it to spamtrap and a bofh entry pointing that address to be a spamtrap.
I think that might be a problem.
You should still be able to cancelmsg this message, and the doublebounce would go to postmaster, which should not be spamtrapped.
Well, I ran cancelmsg on it. The first time it said is was scheduled for cancelling. But repeated calls to malq show it is still there (even a day later). BUT, If I try cancelmsg again I get "Message not found".
Any thoughts on that?
When you sa above that the spamtrap setup might be a problem am I better off putting the individual spamtrap addresses into bofh ? If so, will this work for domains other than the mail server's domain (virtual addresses)?
I can certainly move them into bofh if that will solve the problem and/or is the correct way to do it.
I thought you were already referring to spamtrap addresses in bofh. If not, you'll have to explain what makes your "spamtrap" a spamtrap.
If you've jury-rigged your own spamtrap mechanism, it's your onus to make sure that it accepts bounces, else you'll be causing these kinds of problems. Courier doesn't know, of course, that you've rigged an address to refuse all mail, so it will attempt to deliver a bounce there just like any other message, and if it doesn't go in the original message remains in the queue.
Still, you have pointed out a flaw where Courier would still check bofh spamtraps on bounces, which will interfere with bounce processing. I'll have to fix that.
pgpXz4dJVwoXe.pgp
Description: PGP signature
