One of the recent changes in 7.5 "broke" an incorrect configuration that was previously working, so I thought I would share in my fun.
Someone (not pointing any fingers here...especially since I am the only admin on this particular gateway [:)] managed to leave some static dns entries in a system and then later on added in matching request routing entries to route back to local dns. Probably not a good idea...but easily missed when upgrading from a peer to peer to a client server environment (as if there aren't enough fires to put out).
This seemed to function somehow under 7.4. Upon upgrading to 7.5 everything seemed ok, so out the door I went. About 3 hours later when no incoming mail was coming in I got a callback. Nothing seemed amiss internally...it was email and web surfing as usual.
My first clue came from a bounced test message I send in from an outside domain...RDNS reject? Hmm. Then i started getting the notifications about failure to contact the RBL servers. ???
After perusing the forum a bit it hit me. I had the internal servers with the gateway and the ISP resolvers for forwarding...so internally the astaro dns could fail and everything would still appear to work. Props to BAlf for noting the proper config of pointing the local systems to the gateway and then the gateway to the ISP.
Sure enough, the DNS logs on the astaro reveal BIND with a fatal error since the shiny upgraded version doesn't like two routing entries for the same internal mail server.
The moral of the story is as always...be tidy. 30 seconds saved by not removing those entries just cost me 2 hours!
This thread was automatically locked due to age.