SpamAssassin spam filtering problem

[cmaxwell]

I'll ask and get back to you shortly.

- Chris

 

[rommer]

Thanks Chris!

 

[rommer]

I went and ran mchk on one of my servers and the user_prefs file was recreated in the proper directory. My only concern at this point is having a user_prefs file in the .spamassassin directory going to cause any problems or will it just be ignored?????

 

[cmaxwell]

I'm not sure... I have moved the user_prefs file to the right location on one of our servers and the spam filtering has reverted to the default hits score of 7.0 even though the score is set to 4.0 in Plesk.

I have some questions still open with SWsoft regarding this so I am looking forward to hearing from them.

Will update this when there is more information.

- Chris

 

[cmaxwell]

The problem with SpamAssassin defaulting to 7.0 hits for every user is fixed by running:

/usr/local/psa/admin/sbin/mchk --spam-only

This also puts all 'user_prefs' files in the right location which is rather handy, so I will edit my step-by-step guide above to include this step.

In the past 5 hours SpamAssassin has caught 50 spams - which is about 50 more than it was catching yesterday!

And rommer, the 'user_prefs' file in the '.spamassassin' sub-directory is redundant now and can be deleted or left safely. I'd leave it because it's not exactly using up gigs of space!

- Chris

 

[rommer]

The above fixes have helped 1000% Spam is now being tagged or deleted as the customer sets up.

Two last points that I have found;

1) I ran a sightly different command to recreate the user_prefs

/usr/local/psa/admin/bin/mchk --with-spam

2) The Bayes stuff will ONLY take effect if you have trained or the system has autolearned at least 200 of each ham and spam. Until then the bayes test are NOT done. That is the Spamassassin default which they feel is conservative in order to reduce false positives. You can over ride this and in fact I have done it for several customers that are closely monitoring their email by the following.

In user_prefs add

bayes_min_ham_num 10
bayes_min_spam_num 100

(Use what ever number your comfortable with)

Keep in mind thou that the next time you run mchk the custom settings will be lost but by then they should have enough trained so that the deefaults are OK.

BTW - I also added the RulesDuJour stuff but I'm wondering if there is a test message I can send to see if the new rules are working?

 

[cmaxwell]

Hi rommer,

The mchk command that you ran will do the same job, except it does all of the mail settings and not just SpamAssassin's.

Unless you were having other problems there is not much benefit in running it with the '--with-spam' option which will take quite a bit longer to run.

No harm done though!

- Chris

 

[rommer]

Originally posted by cmaxwell

No harm done though!

What a relief!!!!!

 

[SecondPhase]

Training SA

Hi There, wanted to ask a little about SA training.. when working with sa-learn from the command line it wants to only work with the system user home directory files and cannot deal with the virtual users that qmail uses..

What i did which seemed to work ok, although a lot of work which i haven't finished yet was to run sa-learn as root on a few archives of spam and ham.. then did a sync... and copied the bayes and auto-whitelist files to each qmail users home dir.. I am now getting bayes test results in the spam subject for all users.

Does anybody know if running a system wide bayes deal would be better than the default way this works? or should i not worry about it now that i'm seeded.

Thanks

 

[cmaxwell]

I think it would be best to start this question in a new thread as the original problem has been resolved... and unfortunately i haven't used sa-learn manually so i'm not the man to answer your question

- Chris