postfix & named constantly restarting

[RonBraithwaite]

I’m pulling my hair out on this one (fortunately, I still have lots of hair). Basically, I am getting a repeating set of errors, about every 10 seconds. I have studied the DNS settings, but everything looks correct there. And postfix is supposed to be turned off, dammit!

Anybody got any ideas on this?

Apr 1 08:58:36 goliath named[75319]: starting BIND 9.4.2-P2 -f
Apr 1 08:58:36 goliath named[75319]: /etc/dns/publicView.conf.apple:39: zone ‘86.206.66.in-addr.arpa’: already exists previous definition: /etc/dns/publicView.conf.apple:15
Apr 1 08:58:36 goliath named[75319]: loading configuration: failure
Apr 1 08:58:36 goliath named[75319]: exiting (due to fatal error)
Apr 1 08:58:36 goliath com.apple.launchd[1] (org.isc.named[75319]): Exited with exit code: 1
Apr 1 08:58:36 goliath com.apple.launchd[1] (org.isc.named): Throttling respawn: Will start in 10 seconds
Apr 1 08:58:39 goliath postfix/master[75320]: fatal: bind 0.0.0.0 port 25: Address already in use
Apr 1 08:58:40 goliath com.apple.launchd[1] (org.postfix.master[75320]): Exited with exit code: 1
Apr 1 08:58:40 goliath com.apple.launchd[1] (org.postfix.master): Throttling respawn: Will start in 9 seconds
Apr 1 08:58:46 goliath named[75322]: starting BIND 9.4.2-P2 -f
Apr 1 08:58:46 goliath named[75322]: /etc/dns/publicView.conf.apple:39: zone ‘86.206.66.in-addr.arpa’: already exists previous definition: /etc/dns/publicView.conf.apple:15
Apr 1 08:58:46 goliath named[75322]: loading configuration: failure
Apr 1 08:58:46 goliath named[75322]: exiting (due to fatal error)
Apr 1 08:58:46 goliath com.apple.launchd[1] (org.isc.named[75322]): Exited with exit code: 1
Apr 1 08:58:46 goliath com.apple.launchd[1] (org.isc.named): Throttling respawn: Will start in 10 seconds
Apr 1 08:58:49 goliath postfix/master[75323]: fatal: bind 0.0.0.0 port 25: Address already in use
Apr 1 08:58:50 goliath com.apple.launchd[1] (org.postfix.master[75323]): Exited with exit code: 1
Apr 1 08:58:50 goliath com.apple.launchd[1] (org.postfix.master): Throttling respawn: Will start in 9 seconds

[RonBraithwaite]

Well, I figured out the DNS problem and I think I have fixed it – at least changeip -checkhostname doesn’t complain and there aren’t any more complaints about named in the log. What I did was stop DNS, delete everything in there, recreated it all, saved it, and then restarted DNS. Now that appears to be OK.

What is still getting spit up into the log, however, is the complaint about postfix:

Apr 1 13:02:51 goliath com.apple.launchd[1] (org.postfix.master[25001]): Exited with exit code: 1
Apr 1 13:02:51 goliath com.apple.launchd[1] (org.postfix.master): Throttling respawn: Will start in 9 seconds
Apr 1 13:03:00 goliath postfix/master[25066]: fatal: bind 0.0.0.0 port 25: Address already in use
Apr 1 13:03:01 goliath com.apple.launchd[1] (org.postfix.master[25066]): Exited with exit code: 1
Apr 1 13:03:01 goliath com.apple.launchd[1] (org.postfix.master): Throttling respawn: Will start in 9 seconds

So *something* is trying to start postfix every ten seconds. But what?

When I do a postfix stop or abort, I get told that it isn’t running. OK, I understand that, but WHAT is causing launchd to try to fire it up every 10 seconds?

Sigh.

-Ron

[RonBraithwaite]

That’s what I thought, too, but I have stumbled onto something that fixed it, but I still don’t know what caused the problem in the first place.

I used Lingon (a wonderful launchd configuration editor) to look at the settings for org.postfix.master and noticed that there was a setting to fire off postfix every time a file was dropped in /var/spool/postfix/maildrop/ so I deleted the folder and – low and behold – everything settled down. So I turned it back on and went to watch the folder, but there wasn’t anything there, so I turned it back off and checked again, but still nothing.

So, was something touching (in the unix sense of the word) the folder and that was what was triggering it?

It fixed it, but I’m not comfortable not knowing how this little slice of my consensual reality works.

Sigh.

Thanks for your help, folks.

-Ron

[RonBraithwaite]

Oops, in my last post, I said that I deleted the /var/spool/postfix/maildrop/ folder. I did NOT delete the folder, I deleted the reference to the folder in the launchd configuration file using Lingon. Sorry for any misunderstanding.

-Ron

[RonBraithwaite]

Yeah, if it was just one server it would be too much trouble, but I’m responsible for about 30 different servers and I spent a lot of time chasing this down. If I know what the actual problem is I can do something to make sure it doesn’t happen again.

But I’ve run out of ideas on where to look for what might be triggering this. Do you have any ideas on how I might be able to catch whatever it is in the act or what I might grep for or whatever? I don’t really want to give up on this until I have a culprit. The hunt becomes a lower-priority task, but I want to keep it running in the background.

Thanks for your help, Joel (and anyone else who wants to chip in).

-Ron

[Author]

Posts