My question revolves around Spam Training and checking.
I have two servers (both 4.5.5 FOSS).
Mac OS X
LDAP master, MTA, mailbox server (with POP3/IMAP enabled)
MTA, mailbox server (with POP3/IMAP enabled)
So both of my machines are setup as MTA's.. currently I only have MX records pointed at Server 1 since I just deployed Server 2 not too long ago.
Now to the root of the question:
Does each server has its own spam/ham training accounts? I am only seeing one place for a spam/ham training address, and it looks like server 1 yank's email out of that on its cronjobs and that's it. Therefore Server 2 never gets any "updates" if it falsely classified something. Did I miss something?
I realize I will probably need to copy my dspam/SA training data from Server 1 to Server 2, but what about after that? anything? nothing? That's where I'm not sure of how its supposed to work.
Any insight would be useful, or tips from what others are doing in a multi-MTA environment using Zimbra AntiSpam.