Hi, registered just to post this, in hope that it'll be of help for some other users. This pertains boxes with Plesk + SpamAssassin.

Recently my server has been hit hard with job offer spam, and upon investigation, I realized that the messages in question (which are always self-addressed in some form or other) are actually being delivered without any SA headers after all. However, the logs (the "maillog file", which is present in /usr/local/psa/var/log) showed that the messages in question *did* in fact go through SpamAssassin, so something's up with SA, Postfix, or Plesk's glue scripts. No whitelisting or similar things are in effect.

After much hair-pulling, I contacted my server ISP and they're aware of this attack and that it seems to be affecting a lot of people everywhere, though so far appears to be limited to Plesk boxes. This seems to point to a Plesk-specific bug or exploit.

All this has likely zero to do with SpamAssassin (esp. since Plesk is still on 3.2), but I posted this here at the suggestion of a user in #spamassassin @ Freenode. A temporary very hacky fix is to filter the messages out at the iptables level (see http://twitter.com/#!/rackspot/status/96193246165860353), but this can obviously go wrong in very interesting ways.

-- Bruno Ferreira

Reply via email to