https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6672
Kris Deugau <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[email protected] --- Comment #5 from Kris Deugau <[email protected]> 2011-10-12 20:36:20 UTC --- (In reply to comment #4) > One more point. This message is the ONLY one getting through. It comes > through with no SA report, which my configuration always adds to every > message. I don't see any notes regarding how you're calling SA, but I noticed the complete message you attached is just under 1M. By default, spamc will not pass a message on to spamd if it's over 512K. It will also not add any X-Spam-* headers if the message is oversized. > SA typically has about a 99.999%+ accuracy rate, meaning before this one > started coming through, I would get maybe one or two messages in my inbox > every > 6 months, and 4 or 5 marked spam (not hitting the autolearn threshold) out of > 2500 to 3000 per month. To my account alone. > > As it is, this is the only message coming through. Same subject, same ads, > same everything but the sender and recipient (sometimes it comes to my alias > addresses). And this one *always* comes through, except when it goes to one > of > my honeypot aliases. There's something in the way the headers are set up that > is defeating spamc. When I cut out most of the raw content (just delete most of the lines in the .jpg attachment), spamd on both my desktop and the main spamd cluster returned normal results. This tells me that there's nothing odd enough about the headers to block SA's processing. The fact that spamassassin processed the message, but spamc/spamd didn't, should be another red flag. Both use the same processing code, so if there's something funky in one or more of the headers it should affect both. -- Configure bugmail: https://issues.apache.org/SpamAssassin/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug.
