[gihrig]

Here's an update:

We contacted our ISP to see what their mail server was seeing. They reported:

2007-06-25 15:56:58 fixed_plain authenticator failed for
( zimbra.marketshreservices.com) [72.25.121.211]:36914 
I=[72.34.38.4]: 25: 535 Incorrect authentication data
( set_id=zcollector+marketshareservices.com)

The 'zcollector' account was setup in the original Zimbra install in October 2006 to retrieve mail from the ISP's mail server for all accounts in a single session. That scheme did not work for us, so the zcollector account was abandoned, but never deleted from the Zimbra server, fetchmail, or the ISP's server.

The zcollector account was deleted from all three locations, Friday afternoon while doing maintenance on the Zimbra system. Discussions with users seem to confirm that the probable time of deleting the zcollector account is when out bound mail stopped working.

It appears that the zcollector account was possibly involved in an smtp receive-before-send protocol and that I caused this problem by deleting that account.

I recreated the zcollector account, in all three locations and this seemed to result in a mail loop. Any message sent from the Zimbra server was bounced to the user with a '554 Error: too many hops' message in the mail header.

After deleting the zcollector account followed by a bunch other unsuccessful attempts to glean a little insight, I recreated the zcollector account and the mail loop does not exist. But outbound messages still end up being deferred, just as before.

I'm guessing that I have created a problem between SASL on the Zimbra server and the ISP's SMTP server. Beyond that I'm lost.

Any suggestions of even guesses are sincerely appreciated.

Best,

-Glen