Send mail to test server, not working

[eXtremeSG]

Hi,

I am running an XServe, and setup my old G4 as a Zimbra test server.
I've mapped my cable IP in the DNS records of one domain on my XServe for testing purposes.

If I 'ping' g4mail.testdomain.com, I get the correct IP address.

If I telnet to g4mail.testdomain.com on Port 25, I can see the server responding.

I've mapped ports 110 (POP), 143 (IMAP) and 25 to the G4 box, but my problem is that from my XServe, I am unable to communicate with the Zimbra mail server to send it some emails.

Am I missing a port here, that the two servers need to talk on..?

TIA,
Steve

[marcmac]

Originally Posted by eXtremeSG

Hi,

I am running an XServe, and setup my old G4 as a Zimbra test server.
I've mapped my cable IP in the DNS records of one domain on my XServe for testing purposes.

If I 'ping' g4mail.testdomain.com, I get the correct IP address.

If I telnet to g4mail.testdomain.com on Port 25, I can see the server responding.

I've mapped ports 110 (POP), 143 (IMAP) and 25 to the G4 box, but my problem is that from my XServe, I am unable to communicate with the Zimbra mail server to send it some emails.

Am I missing a port here, that the two servers need to talk on..?

TIA,
Steve

How are you trying to communicate? SMTP? POP? IMAP? IMAPS? POPS? THe only port you don't have listed is 7025, for LMTP, but you probably won't need that.

When you telnet on the appropriate ports from a remote server, do you connect, or timeout? DId you update the firewall on the G4 to allow connections? Are all the zimbra services running?

[eXtremeSG]

Hi Marc, thanks for the reply...

Originally Posted by marcmac

How are you trying to communicate? SMTP? POP? IMAP? IMAPS? POPS? THe only port you don't have listed is 7025, for LMTP, but you probably won't need that.

I think I didn't explain it well before. If I connect to POP or SMTP, then yes it works (there's no mail in there, yet..), but it does work. The issue is me sending an email TO the test server, not in trying to connect to it.
So, when I send the message, I can see in my XServe postfix log (not Zimbra) that it timed out trying to communicate.

I thought the two servers would talk to each other (i.e XServe try and deliver mail to G4) over port 25..? Whilst I can manually telnet into Zimbra on port 25 (and POP + IMAP), it seems the the XServe can't establish a connection to deliver the message to the G4.

Originally Posted by marcmac

When you telnet on the appropriate ports from a remote server, do you connect, or timeout? DId you update the firewall on the G4 to allow connections? Are all the zimbra services running?

Yes, I connect, and it doesn't time out. Yup, the firewall on the G4 is off, so it's only the mapping of the ports on my home router, which is setup to route #25, POP + IMAP to the G4. This works (manually) via Telnet, so I presume the mapping is fine.

So, server to server (Xserve -> G4), it only uses Port #25 to deliver message from one server to another..?

Cheers for your help, and I hope this makes matters a bit clearer!

Regards,
Steve

[marcmac]

That should work - what happens when you telnet (xserve->g4) on port 25 and feed it smtp commands? Does the mail show up in your inbox? (You tested telnet from the xserve, right? Not just from the g4 to localhost).

WHat are you using on the client side to send the mail (on the xserve?) This could be a DNS issue... Is there an MX record pointing to the G4?

[eXtremeSG]

Originally Posted by marcmac

That should work - what happens when you telnet (xserve->g4) on port 25 and feed it smtp commands? Does the mail show up in your inbox? (You tested telnet from the xserve, right? Not just from the g4 to localhost).

Actually, I fixed the first bug, as in the mail servers can now talk to each other. However looking at zimbra.log, it appears there is an isssue with it not being able to talk to LDAP.

Feb 14 13:40:21 g4mail postfix/trivial-rewrite[1680]: warning: dict_ldap_lookup: Search error -1: Can't contact LDAP server
Feb 14 13:41:22 g4mail postfix/trivial-rewrite[1680]: warning: dict_ldap_lookup: Search error -1: Can't contact LDAP server
Feb 14 13:42:22 g4mail postfix/trivial-rewrite[1680]: warning: dict_ldap_lookup: Search error -1: Can't contact LDAP server
Feb 14 13:42:22 g4mail postfix/trivial-rewrite[1680]: fatal: ldap://opt/zimbra/conf/ldap-vad.cf(0,100): table lookup problem
Feb 14 13:42:23 g4mail postfix/smtpd[1578]: warning: premature end-of-input on private/rewrite socket while reading input attribute name
Feb 14 13:42:23 g4mail postfix/master[1543]: warning: process /opt/zimbra/postfix-2.2.5/libexec/trivial-rewrite pid 1680 exit status 1
Feb 14 13:42:23 g4mail postfix/smtpd[1578]: warning: problem talking to service rewrite: Unknown error: 0

Am not really sure where to look for this.... but when I checked in the main.cf file for Postfix, I found this unusual entry:

virtual_transport = error

All the rest show something like: ldap:/opt/zimbra/conf/ldap-vxx.cf

Finally, entering "zmcontrol status" shows everything is running (even LDAP!), *EXCEPT* for mailbox, which is shown as stopped.

Next steps...?

Cheers,
Steve

[marcmac]

what do you get from:
host `hostname`
those are backticks, not single quotes.

Also try:

su - zimbra
zmmtainit localhost
postfix stop
postfix start

That may clear up your ldap/postfix problems.

With the mailbox problem:
su - zimbra
sh -x bin/zmmailboxctl status

[eXtremeSG]

Originally Posted by marcmac

what do you get from:
host `hostname`
those are backticks, not single quotes.

I get: Host hostname not found: 3(NXDOMAIN)

Originally Posted by marcmac

Also try:

su - zimbra
zmmtainit localhost
postfix stop
postfix start

That may clear up your ldap/postfix problems.

So, what to do with this bit..?

Originally Posted by marcmac

With the mailbox problem:
su - zimbra
sh -x bin/zmmailboxctl status

g4mail:~ zimbra$ sh -x bin/zmmailboxctl status
+ BASE=/opt/zimbra
+ SCRIPTS=zmmtaconfigctl tomcat mysql.server
+ STATUS=0
+ '[' zmmtaconfigctl = mysql.server ']'
+ /opt/zimbra/bin/zmmtaconfigctl status
+ R=0
+ '[' 0 -ne 0 ']'
+ '[' tomcat = mysql.server ']'
+ /opt/zimbra/bin/tomcat status
+ R=1
+ '[' 1 -ne 0 ']'
+ STATUS=1
+ '[' mysql.server = mysql.server ']'
+ /opt/zimbra/bin/mysqladmin status
+ R=0
+ '[' 0 -ne 0 ']'
+ exit 1

I also see this at the beginning of the startup sequence:

Feb 14 14:00:44 g4mail /usr/sbin/serveradmin: servermgr_ipfilter:ipfw config:Noticeisabled firewall
Feb 14 14:01:49 g4mail slapd[275]: @(#) $OpenLDAP: slapd 2.2.28 (Nov 9 2005 14:23:20) $\n root@xserve1.liquidsys.com:/Users/build/p4/main/ThirdParty/openldap/openldap-2.2.28/servers/slapd\n
Feb 14 14:01:49 g4mail slapd[275]: /opt/zimbra/openldap/etc/openldap/schema/zimbra.schema: line 2889: bad config line (ignored)\n

[marcmac]

Originally Posted by eXtremeSG

I get: Host hostname not found: 3(NXDOMAIN)

Then you did it wrong. host `hostname` will run the hostname command, then do a host lookup on the output. If you got "host hostname not found" then you used single quotes, not backticks. backticks, on my keyboard, are to the left of the "1" key, under the "~".

Originally Posted by eXtremeSG

So, what to do with this bit..?




g4mail:~ zimbra$ sh -x bin/zmmailboxctl status
+ BASE=/opt/zimbra
+ SCRIPTS=zmmtaconfigctl tomcat mysql.server
+ STATUS=0
+ '[' zmmtaconfigctl = mysql.server ']'
+ /opt/zimbra/bin/zmmtaconfigctl status
+ R=0
+ '[' 0 -ne 0 ']'
+ '[' tomcat = mysql.server ']'
+ /opt/zimbra/bin/tomcat status
+ R=1
+ '[' 1 -ne 0 ']'
+ STATUS=1
+ '[' mysql.server = mysql.server ']'
+ /opt/zimbra/bin/mysqladmin status
+ R=0
+ '[' 0 -ne 0 ']'
+ exit 1

I also see this at the beginning of the startup sequence:

Feb 14 14:00:44 g4mail /usr/sbin/serveradmin: servermgr_ipfilter:ipfw config:Noticeisabled firewall
Feb 14 14:01:49 g4mail slapd[275]: @(#) $OpenLDAP: slapd 2.2.28 (Nov 9 2005 14:23:20) $\n root@xserve1.liquidsys.com:/Users/build/p4/main/ThirdParty/openldap/openldap-2.2.28/servers/slapd\n
Feb 14 14:01:49 g4mail slapd[275]: /opt/zimbra/openldap/etc/openldap/schema/zimbra.schema: line 2889: bad config line (ignored)\n

Don't worry about that. THe first command verifies that it's tomcat that's not starting.

[hootjr29]

marcmac,

You've helped me with some other issues today. I think I'm almost setup. I'm having a similar issue to eXtremeSG except I think I know what my issue is and I don't remember how to resolve this with postfix and maybe you can help.

when I run host `hostname` it returns with my hostname as:
myserver.domain.com has address 69.31.211.5

of course, myserver.domain and the address are replaced with my real name and address. The problem is that this server is running in a vm in my network that doesn't have a static IP address nat'd to it. INstead, I'm port forwarding on my home router to this internal vm that's on 192.168.200.11/24. I remember in the past, when I've setup postfix on an internal server, I had to tell it somehow to not do a dns lookup on my domain name when delivering mail and instead, use a virtuals table or something like that-- which basically told postfix that anything bound for @somedomain.com should go to some IP (namely 192.168.200.11, instead of the outside IP).

Currently I have a ton of incoming mail stuck in postfix, when I do postqueue -p from zimbra's user. When I look in the /var/log/maillog and postqueue -p it looks like all these messages are trying to be delivered to to my outside IP address of my router instead of the internal IP of 192.168.200.11.

Here is one such result of postqueue -p:

AEFA1B5B 1656 Wed Feb 15 15:19:56 somesender@someotherdomain.com
(delivery temporarily suspended: connect to myserver.domain.com[69.31.211.5]: Connection timed out)
myrcpt@domain.com

When I tail maillog, here is one of the results:

Feb 13 16:37:52 myserver sm-msp-queue[32422]: k1C2VVTn002294: to=postmaster, delay=1+18:59:59, xdelay=00:00:00, mailer=relay, pri=3913803, relay=[127.0.0.1], dsn=4.0.0, stat=Deferred: Connection timed out with [127.0.0.1]

Mail internally can go out, mail coming into postfix comes in and seems to get stuck (I'm thinking trying to go to amavis or clam or spamassasin or somewhere that should be internal to this box) trying to leave postfix for another internal destination.

Might you have any suggestions?

UPDATE: actually, here is a post that seems to be exactly my issue:
Mail not getting to mail boxes

and

Installed OK - Can send but not recieve email

but it looks I feel the same way about KevinH's resolution to put up an internal dns server. I know that there is some postfix .db file that you can create to adjust the behavior of postfix only for that domain... I'll keep searching on postfix sites and here, but if anyone knows what this file is, please let me know.

Thanks, Joe

[hootjr29]

I don't know what the file is, but I just opened 7025 on my firewall and now everything is working (although it may not be working the most effecient, since I imagine things are going out to my firewall and then coming back in on port 7025).