Would greatly appreciate some help configuring notifications

[millisa]

I'm assuming you get that message when running

Code: Select all

sudo mailq

Nope, not normal. Sounds like your local mail server (postfix in this case) isn't running. You can quickly verify by telnetting to localhost port 25.

Try stopping postfix, watching your /var/mail/mail.log and starting postfix. It usually will give you a hint why it's not running.

[logic_bomb421]

Yes, I see that message when checking the mail queue.

Alright, so I did

Code: Select all

telnet localhost 25

and everything worked just fine. I got no errors. However, when I stopped postfix and started it while watching the log, I did see something interesting:

Code: Select all

*timestamp* SERVERNAME postfix/master[MISC PORT]: fatal: bind 0.0.0.0 port 25: Address already in use

Where is it getting 0.0.0.0:25 from?

[millisa]

Sounds like you already have something listening on port 25. Stop postfix; I expect you'll still be able to telnet to port 25.
Show what's listening:

Code: Select all

sudo netstat -tulpn

Make it just show you port 25 stuff:

Code: Select all

sudo netstat -tulpn|grep :25

The last column shows the pid/process name. If it isn't obvious which program is listening from that, you can do a

Code: Select all

sudo ls -l /proc/PutPidNumberHere/exe

and it should point you to the binary that is doing the listening.

Quick example (i'm on as root, so I'm leaving out the sudo's)

Code: Select all

[root@abc1]# netstat -tulpn|grep :25
tcp        0      0 127.0.0.1:25                0.0.0.0:*                   LISTEN      27899/master        
tcp        0      0 ::1:25                      :::*                        LISTEN      27899/master        
[root@abc1]# ls -l /proc/27899/exe
lrwxrwxrwx 1 root root 0 Jul 22 13:56 /proc/27899/exe -> /usr/libexec/postfix/master

So I have postfix listening on tcp25.

I'm wagering this next doughnut that you've got sendmail running still or postfix isn't actually getting stopped.

[logic_bomb421]

Alright, managed to figure out that sendmail is still listening on port 25 (it actually says 127.0.0.1:25, why does sendmail have the localhost?).

What I don't understand now, is how sendmail is still present? I did

Code: Select all

apt-get purge sendmail

before I installed postfix because I was worried about conflicts like this. Guess I didn't fully remove it?

Next question, how do I stop sendmail from listening and allow postfix to pick it up?

Thank you for all your help by the way.

[millisa]

I'm not an ubuntu guy, but I saw a few posts where people claimed you needed to do

Code: Select all

sudo apt-get purge sendmail*

Maybe there is a startup script still lurking? For now, you could just kill the process id you found in the netstat above:

Code: Select all

kill putPIDhere

and if it still hangs around throw a -9 in there:

Code: Select all

kill -9 putPIDhere

Recheck that it's gone with the netstat, telnet, etc. Then try starting postfix again. If it looks to be working, you should probably check it on a reboot.

[logic_bomb421]

Awesome! That worked! I still have to test the reboot, but at least I have something to stand on now.

I was able to get mail to come through using the command prompt. Nagios notifications still aren't showing up, but that's because I reconfigured the send command earlier. I'm going to revert back to default and see what that does.

Thank you!

[tmcdonald]

Thanks millisa, as always, for the help!

logic_bomb421, let us know how this works out for you.

[logic_bomb421]

Notifications are 5 by 5! Somehow sendmail was still clogging port 25 as speculated above. I killed the sendmail process and notifications started coming through almost instantly! I rebooted my Nagios machine and postfix was still bound to port 25.

Thank you for all your help! Even though this wasn't necessarily a Nagios issue, I still greatly appreciate you taking the time to troubleshoot this with me.