How to bounce mails immediately?

[dipeshmehta]

Hello,

Everytime when I go to check mail queues on Administration Console, there are few mails found in diferred queue, which are by mistake sent by our users to wrong users or wrong domains with typo mistakes. Is there anyway by which I can bounce the mail back to sender immediately? There are 'Hold', 'Requeue' and 'Delete' buttons only.

Dipesh

[thunder04]

I have a similar question. Is there a way to adjust the length of time Zimbra re-tried sending an e-mail before it bounces?

For instance, a user sent an e-mail to gmail.co (an obvious typo) which took almost a week to bounce.

The bounce error was:

"<johndoe@gmail.co>: Host or domain name not found. Name service error for
    name=gmail.co type=MX: Host not found, try again"

I suspect part of the problem is that we use OpenDNS, and OpenDNS returns A records for any DNS lookup.

In any case, how would I go about adjusting the retry limits? I'm thinking I want it no longer than 24 hours.

Thanks!

[mmorse]

postconf -e maximal_queue_lifetime=5d (I'd still keep it 1-2d minimum. If someone you're sending to has a service outage, dns issues, throws a temp error etc it gives them time to get their server back in working order.)

There's also bounce_queue_lifetime which sounds relevant but according to my reading of the manual actually control how long a mailer-deamon message sits ("The maximal time a bounce message is queued before it is considered undeliverable. By default, this is the same as the queue life time for regular mail.")

The queue lifetimes are also a double edge sword, sometimes I wish there we're more variables: If you use postfix as an edge that accepts mail from external and delivers to a next level internal, you want to accept from outside and keep retrying should the internal go down - as you want to see what the incoming email was all about - so it might suit your needs to have two separate parameters controlling the outgoing (sent) vs incoming (received). Of course the counter argument revolves around it's courteous to let the external senders (incoming) know their mail didn't go through either so it's probably more prudent to use a temp error from the edge right away when internal mta isn't reachable...

OP's question re triggering a mailer-daemon on demand when the admin can see it won't work sounds like a good RFE

[thunder04]

Awesome, thanks! I'm thinking 1-2d as well.