Local Loop issue...

[KDios]

Hello all, recently we set up an Exchange 2010 system to co-exist with users on Zimbra, and set up (we think) the system as a split domain. Routing internally and externally, all mail does get to where it is supposed to go.

The strange part is, we're getting local loops. The loops are caused by something that doesn't appear to exist. Mail that goes from exchange accounts to any other accounts work fine. Mail from outside in is fine. Mail from inside out is fine. Mail from zimbra to exchange is not fine.

Users can send mail internally from zimbra to exchange, and they receive it, but many emails sent from zimbra also generate a new email that bounces back and forth between both boxes. The email is sent to ''''@mydomain.com. Exchange only shows the quotes as the user. There obviously is no """" user. In zimbra, the log looks like this when the email finally stops and NDRs:

Jul 15 13:57:25 mail postfix/smtpd[14909]: connect from mail.myexchangeserver.com[10.42.*.**]
Jul 15 13:57:25 mail postfix/smtpd[14909]: setting up TLS connection from mail.myexchangeserver.com[10.42.*.**]
Jul 15 13:57:25 mail postfix/smtpd[14909]: Anonymous TLS connection established from mail.myexchangeserver.com[10.42.*.**]: TLSv1 with cipher AES128-SHA (128/128 bits)
Jul 15 13:57:25 mail postfix/smtpd[14909]: A7543200D177: client=mail.myexchangeserver.com[10.42.*.**]
Jul 15 13:57:25 mail postfix/cleanup[14916]: A7543200D177: message-id=<20110715195517.25B58200D17C@mail.***.net>
Jul 15 13:57:25 mail postfix/qmgr[2366]: A7543200D177: from=<>, size=7045, nrcpt=1 (queue active)
Jul 15 13:57:25 mail postfix/smtpd[14909]: disconnect from mail.myexchangeserver.com[10.42.*.**]
Jul 15 13:57:30 mail postfix/smtp[14917]: A7543200D177: to=<@mydomain.org>, relay=10.42.*.**[10.42.*.**]:25, delay=5.1, delays=0.1/0.01/0/5,
dsn=5.4.6, status=bounced (host 10.42.*.**[10.42.*.**] said: 554 5.4.6 Hop count exceeded - possible mail loop (in reply to end of DATA command))
Jul 15 13:57:30 mail postfix/qmgr[2366]: A7543200D177: removed

The bold part is where it gets me. It generates a blank or blanket user @ my domain. I'm not sure if that's because I mistyped something in the catchall, or something else. Has anyone seen anything like this? Does anyone know where all the zmprov md information is stored?

Thanks guys, I'll be happy to post any more info if needed.

[phoenix]

Originally Posted by KDios

The strange part is, we're getting local loops. The loops are caused by something that doesn't appear to exist. Mail that goes from exchange accounts to any other accounts work fine. Mail from outside in is fine. Mail from inside out is fine. Mail from zimbra to exchange is not fine.

What sort of email from Zimbra cause this problem? IS Zimbra the primary of secondary for the split domain? Di you follow the instructions in the wiki article for this configuration? Is you DNS correct fro both systems?

[KDios]

Thanks for the response,

Zimbra is set up as per the Split Domain wiki as a secondary, I used the Catchall commands

"$ zmprov md example.com zimbraMailCatchAllAddress @example.com
$ zmprov md example.com zimbraMailCatchAllForwardingAddress @example.com
$ zmprov md example.com zimbraMailTransport smtp:mail.example.com
"
As well as the commands suggest afterwards. External mail doesn't loop at all. Internal mail from exchange doesn't loop, only mail originating from Zimbra and going to exchange loops, although we only have a few accounts at the minute set up in exchange. I checked the DNS a few times, and modified the hosts file on Zimbra to make sure exchange was showing up correctly. We use dyndns for other routing. Is there any special way I should modify the hosts file or is there any other location to check?

What about the "relayhost = " field in the main.cf file? Is that only for external routing or does it need to be filled in for internal as well?

Last thing...do you know where to find the aforementioned Catchall configuration when using zmprov? I can't figure out where that information is stored.

Thanks again!

Kevin