Messaging Gateway

We have several production systems using bogus from addresses that are unable to receive replies.  (I'm not involved in that side of the business so I'm unaware of the content or why they do this.)  These systems are sending mail out my gateway and a lot of it bounces.  However, since the bounces use invalid from addresses, they sit in my gateway retrying for our default timeout period.  For most of them, the domain is invalid so the actual error message is "421 4.4.0 [internal] no MXs for this domain could be reached at this time".

How do you guys handle internal bounces to invalid e-mail addresses?

Well, the most obvious answer is to use a routable return address in your emails.

The second option is to try detecting these emails using a content filtering rule and deleting them.

Unfortunately I have no control over option #1.  Also, I know most orgs have non-routable "no-reply" e-mail addresses so was looking into the most common solution people use for those.

Option #2 is the direction I was leaning, but was just checking here to see if that's how others handle it or of there was something I wasn't thinking of.

If you use exchange, you can create a contact that doesn't route mail anywhere and alias all of these undeliverable "no reply" messages to that contact, assuming the no replay addresses are in the domains exchange manages.

I dont' think content filtering (CF) would fit this very well. The messages have already been accepted and scanned, and would be sitting in the delivery queue, so CF wouldn't take further action on them.

Try going into Protocols > Domains and creating non-local domains that you are experiencing problems with and direct these false domains to a catchall SMTP process that deletes all messages. When the message fails delivery, the NDR will be delivered to the catchall rather than waiting for another delivery failure timeout.

This will only work for the domains that are not valid, but you mentioned this is most of them. You will also need the catchall server, but that just requires a system you can open port 25 to which respods to SMTP commands and just junks all messages. Not the best option, but it should do the trick.

As to other orgs, I'm curious how often the address is just a catchall deletion rather than truly non-routable.

Art_P hit the exact issue I was guessing I would run into.  The NDRs are being generated by the gateway and don't seem to go through CF.  I've tried creating a policy, but it fails.  So I assume Cricket17s idea would fail for the same reason.

I honestly don't like the other idea of setting up a dedicated SMTP server for the catchall.

I may have to just beat a linux admin over the head.

Some of these come into the gateway first and then go out for delivery.  For those, does Brightmail have the ability to re-write the From address?  Perhaps I could use Cricket17's idea if I could re-write the address in Brightmail.