LinkBack URL
About LinkBacks
Senior Member
Ok, got two spam in the junk folder, and the headers show it seems to be working:
X-Spam-Status: Yes, score=13.668 tagged_above=-10 required=4
tests=[BAYES_60=1, MISSING_DATE=0.001, MISSING_HEADERS=1.292,
MISSING_MID=0.001, MISSING_SUBJECT=1.762, RCVD_IN_PBL=0.905,
RCVD_IN_SORBS_WEB=0.619, RCVD_IN_SPAMRATS_NOPTR=2, RCVD_IN_XBL=3.033,
RDNS_NONE=0.1, SEM_URI=0.5, SEM_URIRED=0.5, URIBL_BLACK=1.955]
autolearn=spam
Moderator
You can add this one as-well to your arsenalCode:################################################################################ # NIX SPAM RBL (http://www.dnsbl.manitu.net/) ################################################################################ header RCVD_IN_NIX_SPAM eval:check_rbl('nix-spam-lastexternal','ix.dnsbl.manitu.net.') describe RCVD_IN_NIX_SPAM Listed in NIX-SPAM DNSBL (heise.de) tflags RCVD_IN_NIX_SPAM net score RCVD_IN_NIX_SPAM 0.5
Senior Member
Here is the link from one that didn't get caught earlier today. This was before adding the extra DNSBL checks:
Email Info - rdn4c1j9
Moderator
Cool .. I score at 8.76 so hopefully your new checks will sort things out for you a bit
Moderator
EDIT: Cross-posted with the last several posts...
Possibly not. You can look in salocal.cf to see if the changes made it through. In my case I went ahead and restarted zimbra entirely because it looked like they didn't. For me this seems to be a change from 5.x.Originally Posted by swrightsls
Good question. There is a settable numeric threshold which if exceeded will result in SA using the message for its Bayesian learning. The default is pretty high, and it's found in /opt/zimbra/conf/spamassassin/10_default_prefs.cfWhat is the significance of the 'autolearn' tag?
It actually has three settings:
The version of SA included in 6.0.5 seems to add some rules (or maybe it's just new reporting). In short there are some exceptions to the simple numeric threshold:Code:bayes_auto_learn_threshold_nonspam 0.1 bayes_auto_learn_threshold_spam 12.0 bayes_auto_learn 1More info: Mail::SpamAssassin::Conf - SpamAssassin configuration file and Mail::SpamAssassin::Plugin::AutoLearnThreshold - threshold-based discriminator for Bayes auto-learning
This probably explains why I'm seeing "autolearn=spam" on some messages and "autolearn=no" on others, even though they have similar scores. However I seem to remember seeing some mail items marked as "autolearn=spam" even though their scores were less than 12. I don't have any samples at the moment to investigate, though.
Senior Member
Here's another very spammy email that just came through:
ATM
Moderator
Update on my previous note, it seems that another reason a message may autolearn (or not) even though it's below/above the default threshold is this:And regarding the actual header message, this is controlled by amavisd-new, not spamassassin, and was implemented in amavisd-new 2.6.2 See http://www.ijs.si/software/amavisd/release-notes.txt[...]the score used to trigger autolearning is somewhat different than the one reported in the final score; therefore a score displayed in the headers that obstensibly should trigger autolearning will not do so. Again, use the "-D" flag to SpamAssassin, and you will see the score that is used to determine whether or not autolearning will be triggered.
Finally, SpamAssassin requires at least 3 points from the header and 3 points from the body, to auto-learn as spam. If either section contributes fewer points, the message will not be auto-learned.ZCS 5.0.20 I believe uses amavisd-new 2.5.4, so that's why I'm seeing the change.- insert autolearn=... information field into an X-Spam-Status header field,
similar to how SpamAssassin does it; suggested by Jonathan Skanes;
Active Member
After upgrading from 6.04 to 6.05 I get more spam to Inbox.
Also the mj doesn't seem to work when used from search results. It does move the email from search result, but the email doesn't get moved to Junk folder.
(There's no "moved 1 email to Junk" AJAX notification when doing the mj command).
Moderator
Why do you believe you are getting more SPAM ? Any examples ? Did you make any custom changes before ?
With respect to the keyboard shortcut have you checked Bugzilla Main Page - Zimbra to see if it has been filed ? and if it has not go ahead and raise one.
Active Member
AJAX notification seems to appear with 6.06.
When marking spam as Junk - either via "mj" or selecting "Junk" - the message gets marked as NOT Spam/Junk. This happens with saved search "is:unread".
Probably related to Bug 45906 – Junk button not working correctly
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
