Ok, now I am starting to get seriously confused with zimbra..
Been playing a bit with OpenProtect's SpamAssassin sa-update channel to test, and from I can see there those things are based on the old Rules Du Jour or SARE or what to call it ;p But the guide pasted above works with dirs outside /opt/zimbra, and then it suddenly (after creating various dirs etc that sa-update needed...) takes those (seems also to be rather old..) rules in to consideration. I thaught whole zimbra setup worked inside /opt/zimbra, at least most of it?
Which brings me to ask, which spamassasin setup is supposed to be the running one? The conf and such inside /opt/zimbra/conf/spamassassin or the ones in /var/lib/spamassassin/3.001009/ ? Which also makes me wonder, is maybe the autolearning of spam/ham not working properly cause it trains the spamassasin rules in /opt/zimbra/conf/spamassassin but spamassasin actually uses /var/lib/spamassassin/3.001009/ and so forth.
Anyone can clearify a bit for me?
To keep me even more confused, after making the dirs sa-update (ran as root) told me where missing, and importing default rules + SARE rules there (as explained above) the spamassasin filter works better. Suddenly spam gets much higher score then before and the % of the caught spam went up. and the list of scores is much longer then before (see paste below).. Obviously I have triggered something here, but, its not supposed to be this I way I take, something gotta be wrong somewhere here, just afraid I sit here now with 2 spamassasins, 1 new one now outside(!) /opt/zimbra that actually checks the e-mails that comes in, and one inside /opt/zimbra that gets trained with Junk / Not Junk button of users, but thats actually aint in use at all....
X-Spam-Status: Yes, score=13.362 tagged_above=-10 required=5
tests=[BAYES_50=0.001, FB_ADD_INCHES=2.131, HTML_MESSAGE=0.001,
RCVD_IN_SORBS_WEB=0.619, RDNS_NONE=0.1, SARE_SUB_INCHES=0.221,
URIBL_AB_SURBL=1.86, URIBL_BLACK=1.955, URIBL_JP_SURBL=1.501,
URIBL_OB_SURBL=1.5, URIBL_SBL=1.499, URIBL_SC_SURBL=0.474,