Its always frustrating when you spend a few hours of time to write some code to do something, and when you get it working and suddenly months later it stops. Its more frustrating when you can’t figure out whats wrong with the script. When you finally realize that the script is working correctly and the emails are going into limbo its just plain annoying. So you finally about a month later find out that the messages are bouncing with a “550 unknown user” error (not getting the bounce backs was annoying too). Thank goodness Iarwain forwarded me some of them. So here is the issue as I see it…
Dec 6 01:01:21 datasquire sendmail: jB661LXl007965: ruleset=check_rcpt,
Dec 6 01:01:21 datasquire sendmail: jB661J2d007964: email@example.com,
ctladdr=nobody (99/99), delay=00:00:02, xdelay=00:00:02, mailer=relay, pri=31097,
relay=localhost.net. [126.96.36.199], dsn=5.7.1, stat=User unknown
The three hints to me are “relay=localhost.net”, “(may be forged)” and “Relaying denied”. The localhost.net threw me off at first. The answer to this is of course quite simple. “(may be forged)” and “Relaying denied”. Our first hint is that relaying for mail generated on the local machine isnt allowed to be relayed to where it belongs. The “(may be forged)” means that it is having issues resolving the name to an IP address, or the name didnt resolve at all.
Why would this happen all of a sudden? Great question. When the whole thing started i presumed that Perl was upgraded and something didn’t work quite the way it did before. So here after finally sitting down for a few hours today is the reason you all have been waiting for:
Starting with version 8.9, sendmail denies relaying by default.
Well its about bloody time. Rigor is of course running 8.12 Well it seems that we have found the culprit after all.