Hi All, I am trying to configure spam assassin. By default it is set to auto learn (AWL). But i found that many proper mails are being sent to spam folder and some spam mails are going undetected to Inbox..
I tried checking whitelist using spamassassin-tools that's when i realised that it couldnt autolearn. Let me explain the mail server setup for better understanding. The mail server is database driven. All user auth and mail boxes are in separate database servers. Incoming mails are first scanned by anti spam anti virus server (single server, 4 in total). After compleltion of scan it is forwarded to a windows server where user is looked up and the mails are diverted to its correct database server. Now, spamassassin is trying to look for the mailbox path on the same server (scanning server), which it cannot locate. It fails. And so, it doesn't auto learn. The company had documented AWL (auto whitelist) method for its scanning servers. So, prime focus is to use the same setup.. if it cant be done that way, then i need to manually configure spamassassin. In the first case of AWL... i read somewhere that an improvement had been made and that spamassassin would use berkley db file to store/maintain whitelist instead of using $user mail box path. But, the documentation ended with "ToDo". I found no relative information thereafter. If some one can throw more light on this aspect, then it will be really useful for many on the internet. Plan B: If AWL with db file setup isn't possible, then how do I setup spamassassin to block unwanted mails. Is anyone aware of how google/yahoo does it? Thanks and regards... DoOrsOfpErcEpTiON
