Postfix Issue

Hello all, I finally had to register rather than reading the threads because I came across an issue that I cannot figure out.

Here is what I have...

Fedora Core 5, running ispconfig with postfix and spamassassin (spamassassin is not very well configured due to lack of time so it really isnt doing anything as of yet). This has been running for several weeks without a single hitch. Until this morning. It seems as though incoming email is getting lost somewhere, but I cannot find where it is going. I have another mail server and I can read in the logs where mail is being delivered to the postfix server, but when users check their mail, there is nothing there. There is no problems with logging into their mail accounts, everything seems fine. Just no mail.

Anyone encounter a similar issue? And have some information on where I should take it from here?

 

Did you enable Maildir in ISPConfig (Management -> Server -> Settings -> Email)?

 

Yes, I did. I believe that is one of the steps in your FC5 perfect setup.. Which is the tutorial I used for the initial setup. (Thank you very much for your work in helping!)

 

Did you double-check?
What's in your mail log?

 

Yes, I just checked it.. The mail log is pretty long.. Is there anything in particular I should look for? I do notice a line saying "Cannot Load Certificate Authority Data" could this be a certificate issue?

There is also this line..... comes up every so often in the log.
warning: TLS library problem: 3143:error:2006D080:BIO routines:BIO_new_file:no such file:bss_file.c:125:

 

Send a mail to your server, and then post the mail log excerpt from that time.

 

Ok, here it is.. I sent a test mail at 8:32 AM.. Here is the log for that time

Code:

Oct 16 08:32:35 XXXXXX dovecot: pop3-login: Login: user=<XXXXX>, method=PLAIN, rip=::ffff:XXX.XXX.XXX.XXX, lip=::ffff:XXX.XXX.XXX.XXX
Oct 16 08:32:36 XXXXXX dovecot: POP3(XXXXX): Disconnected: Logged out top=0/0, retr=0/0, del=0/0, size=0
Oct 16 08:34:15 XXXXXX postfix/anvil[23793]: statistics: max connection rate 1/60s for (smtp:85.140.115.27) at Oct 16 08:30:54
Oct 16 08:34:15 XXXXXX postfix/anvil[23793]: statistics: max connection count 1 for (smtp:85.140.115.27) at Oct 16 08:30:54
Oct 16 08:34:15 XXXXXX postfix/anvil[23793]: statistics: max cache size 1 at Oct 16 08:30:54
Oct 16 08:37:37 XXXXXX dovecot: pop3-login: Login: user=<XXXXXX>, method=PLAIN, rip=::ffff:XXX.XXX.XXX.XXX, lip=::ffff:XXX.XXX.XXX.XXX
Oct 16 08:37:37 XXXXXX dovecot: POP3(XXXXX): Disconnected: Logged out top=0/0, retr=0/0, del=0/0, size=0

 

I don't see any mails arriving in that log. Are you sure that the MX record for your domain is pointing to the correct server?

 

It is possible that is the problem, I am not doing the DNS forwarding for this domain. It is done at the isp's DNS server. However, I would think that the mail would come back as undeliverable. But, it doesn't.

The website that is running on this domain is still working just fine. So, I am inclined to think that it is not a forwarding issue. But, it certainly could be. Just to be sure I will call the isp and see if they are forwarding port 25 to the proper public address.

 

Can you post the domain name here that you're sending to?

 

I would be more comfortable pm'ing it to you if that is ok with you.

Again, thanks for all your help, it is very much appreciated.

 

Is 199.xxx.xxx.114 your server's public IP address? Is it connected directly to the internet, or is it in a local network?

 

Yes, that is the public IP. It is in a local network, using port forwarding to translate to the local ip address.

 

Maybe your ISP decided to block port 25? You should ask them.

 

Good thought, but not the issue, I have another mail server running on a windows machine that is using the same connection (different public IP) and is still working properly. I put this server together to learn more about linux and thought I had done pretty good until this popped up..

Would it be a good idea to uninstall and reinstall postfix?

 

I don't think it's a Postfix problem. Did you check the MX records of your domains? Are they correct? You can check with

Code:

dig mx mydomain.com

 

here is the output of that command... looks good to me.

Code:

; <<>> DiG 9.3.2 <<>> mx XXX.net
;; global options:  printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 17765
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 2, ADDITIONAL: 3

;; QUESTION SECTION:
;XXX.net.                   IN      MX

;; ANSWER SECTION:
XXX.net.            1200    IN      MX      0 hostmail.XXX.net.

;; AUTHORITY SECTION:
XXX.net.            1200    IN      NS      teton.srv.net.
XXX.net.            1200    IN      NS      dns1.srv.net.

;; ADDITIONAL SECTION:
hostmail.XXX.net.   1200    IN      A       199.XXX.XXX.114
dns1.srv.net.           1200    IN      A       199.104.81.3
teton.srv.net.          1200    IN      A       199.104.81.6

;; Query time: 44 msec
;; SERVER: 199.104.81.6#53(199.104.81.6)
;; WHEN: Mon Oct 23 08:59:26 2006
;; MSG SIZE  rcvd: 145

 

Looks ok. Can you run

Code:

tail -f /var/log/maillog

and at the same time send a mail to your server? What happens in your mail log then?

 

Well, I did this.. what I saw was "unknown local recipient" error and an error saying "TLS Library problem".. So, I removed the "test" domain in ISPconfig and sent another mail to the server. It came through. So, something was messed up. No idea what or why but it seems to be working properly now. Thank you very much for your suggestions.