Hi folks.
For some reason my server started tossing "Relay access denied" at me. The server is on my LAN and it's been working sporadically. I've restarted the server and my mail.app client and I have no clue why it's giving me this.
Any advice appreciated. I have been using this account for years.
Cheers
Mac Pro, OS X Mountain Lion (10.8.2), Hypercard UG!
A relay access denial usually means you're not authenticating to the server or you're using port 25 for your mail client (and avoid that), or that the server isn't authenticating correctly to the relay server, if you're using one of those.
Is this the same mail server that has had occasional issues before? (Here and here and here are some earlier discussions.)
Launch Terminal.app from Applications > Utilities and post the output of the following
sudo changeip -checkhostname
That should report no changes required.
Might also want to post the output of your postconf configuration. (I think postconf -n will still work on 10.8; the configuration files moved in 10.9, and the default postconf -n doesn't show the local configuration there.)
if you're still at that same domain, then there's also no public reverse DNS for the mail server from the ISP, which means mail that isn't being relayed can end up getting dropped by receiving mail servers:
$ dig +short MX pints.com.
0 mail.pints.com.
$ dig +short mail.pints.com.
69.196.158.234
$ dig +short -x 69.196.158.234
$
A relay access denial usually means you're not authenticating to the server or you're using port 25 for your mail client (and avoid that), or that the server isn't authenticating correctly to the relay server, if you're using one of those.
Is this the same mail server that has had occasional issues before? (Here and here and here are some earlier discussions.)
Launch Terminal.app from Applications > Utilities and post the output of the following
sudo changeip -checkhostname
That should report no changes required.
Might also want to post the output of your postconf configuration. (I think postconf -n will still work on 10.8; the configuration files moved in 10.9, and the default postconf -n doesn't show the local configuration there.)
if you're still at that same domain, then there's also no public reverse DNS for the mail server from the ISP, which means mail that isn't being relayed can end up getting dropped by receiving mail servers:
$ dig +short MX pints.com.
0 mail.pints.com.
$ dig +short mail.pints.com.
69.196.158.234
$ dig +short -x 69.196.158.234
$
Hi there.
I killed a domain recently that has an associated SSL attached to it. Server.app won't delete it as it's being used for mail (how, I don't know). Ya the whole connection from my client to the server is screwy. Fixed internal IPs. Server.app is the most horrible piece of software Apple puts out. It would be hugely beneficial if it read from config files.
OK, killed the certs, reset mail.app, now she's working.
So all of this is because a domain was killed and the server.app was pointing to a cert that it shouldn't be using.
Thanks for your attention. I keep coming in here and it's you that answers. I appreciate it.
Cheers
SMTP mail server definitely gets wonky when its associated DNS is incorrect. Do check that.
The SSL/TLS SMTP submission ports do use certificates.
Server.app does read from config files — plists — and those are accessed via serveradmin commands.
Certificate handling has been somewhat wonky for many years. Sometimes better, sometimes worse. It's been decent lately, but I first ran into related weird with 10.5 or 10.6.
FWIW, I've had reasonable luck with Server.app 3.0.2. It works. Haven't had serious problems with it in a while, though there are cases where I've had to go 'round it, such as with setting up SMTP relay through TCP 587 via serveradmin commands.
If things get crazy enough with Server.app, I'll load up one of the BSD variants on a spare x86-64 box, and go have a test drive; see how well that configures.
DNS was fine, but because of vhosts the cert was messing things up and other domains defaulted to using that most recent cert. Again, server.app. The config files I would expect would be like httpd.conf syntax, not plists. plists can reside post-config, so that it's easy for in-house to get this done, especially since the GUI's dropdowns (or whatever) keep messing up associations.
I now have it working without SSL. The server is 10.8.5, running server.app 2.2.2. I'll go have a look at 3.0.2. I'm running Mavericks on this box, but from what I remember, I can't seem to talk to the server for some reason.
Cheers
Hi there.
So I'm still having issues. This is what I'm getting for multiple accounts:
The server response was: Service currently unavailable
The account has default ports with SSL. The only SSL is the self-signed cert, which isn't really anything. Authentication is "none" and I've previously opened this up some time ago. TLS certificate is None as well.
I thought I had it going without SSL, but it won't allow me to disable it.
Any advice?
Cheers
New information, which worked. On the server, in the terminal, I did the following under the superuser:
postfix reload
It reset something.
Solved.
Cheers
Relay access denied
