Yes, sure seems like that there is 2 running for some weird reason, unsure if the default Red Hat Enterprise Linux 5 comes with spamassasin and that somehow have taken control over the zimbra version of spamassasin.
Doubt the traffic goes trough both, but what it seems like is that traffic actually goes trough the one outside /opt/zimbra , but training of the spamfilter goes to the spamassasin in /opt/zimbra. Which makes the whole solution quite "useless" atm, set in ""s cause it catches spam by all means, but when users use the Junk button the filter in use aint actually in use. These are speculations, but it sure seems like it.
What would you suggest I do to make absolutly sure that it uses only the one in /opt/zimbra (I expect thats whats it supposed to do right?) so I could remove any reference to spamassasin I can find outside /opt/zimbra (is that safe to do?).
A question for later testing, if I send an e-mail, with some text, lets say just some random spam message grabbed from a website posting spam they have recived, but make sure its an e-mail not going above my minimum score. When it reaches my inbox, hit the junk button and run /opt/zimbra/bin/zmtrainsa as zimbra user. Should that e-mail get caught as spam the next time I send it, or after the 2nd time I do that procedure or? Cause I did this test, I marked it as spam about 7 times and it still didnt get marked as spam which was the first thing to make me wonder that something isnt working as it should with spamassasin.
Appricate all the help to resolve this matter.
__________________
Release 5.0.8_GA_2462.RHEL5_20080709170039 RHEL5 FOSS edition
|