1. Please take a little time for this simple survey! Thank you for participating!
    Dismiss Notice
  2. Dear Pleskians, please read this carefully! New attachments and other rules Thank you!
    Dismiss Notice
  3. Dear Pleskians, I really hope that you will share your opinion in this Special topic for chatter about Plesk in the Clouds. Thank you!
    Dismiss Notice

Spamassassin user prefs template

Discussion in 'Plesk for Linux - 8.x and Older' started by liquorvicar, Sep 9, 2005.

  1. liquorvicar

    liquorvicar Guest

    does anyone know where Plesk reads the default user_prefs from when you create a new mail name?

    i'd like to amend them so the default is not to include the original message as an attachment.

  2. jamesyeeoc

    jamesyeeoc Guest

    I believe Plesk keeps it in mySQL.

    Database: psa
    Table: sa_conf

    each field has a default value.

    Field: spam_action
    Default: save
    Type: enum('delete', 'save')

    So I guess you would change it to delete.
  3. liquorvicar

    liquorvicar Guest

    not quite what i was after. that table controls the SA settings as visible from the CP. what i'm after is the user_prefs file saved as
    my default looks like this:
    required_hits 7
    rewrite_subject 1
    subject_tag *****SPAM*****

    i want to add:
    report_safe 0
    use_terse_report 1
    report_header 1

    to each one so that SA just changes the subject line and headers of the original message.
  4. ProfiTiger

    ProfiTiger Guest

  5. jamesyeeoc

    jamesyeeoc Guest

    Hmmm, I am not sure what table you are looking at, but on mine, the sa_conf table shows the following fields (default value):

    flt_enabled (both)
    rw_subject_tag (*****SPAM*****)
    hits_required (7)

    Looks like the table to me....
  6. ProfiTiger

    ProfiTiger Guest

    Hi jamesyeeoc,
    I think we are talking about SA-configuration per-user, not server-wide-settings. SA can be configured server-wide with the file /etc/mail/spamassassin/local.cf. Here is the behavoir of SA (like scoring, use of razor, pyzor, dcc and so on) defined. This can be done also on a per-user-configuration with the file user_prefs in /var/qmail/mailnames/domain.tld/mailname.
    With this file (equal to local.cf), you can specify user-specific settings for SA, so each user can use his own bayes-database. Some lines of this file are equal to data in sa_conf (required_hits 7, rewrite_subject 1, subject_tag *****SPAM*****). So I think plesk stores these settings in sa_conf and generates with this data the user_prefs-file for SA.
    The problem of liquorvicar and me is, that we want to modify the "default" user_prefs, which is generated (by script?) or copied on mailbox-creation so that we don't have to edit or copy this file manualy for each new mailbox. I asked the same question in http://forum.plesk.com/showthread.php?s=&threadid=27091 (Thought, "Installation and 'How Do I' Questions" is the right place for this question).
    Here is a sample of my user_prefs:
    required_hits 4
    rewrite_subject 1
    subject_tag ***SPAM***:
    check_mx_attempts 2
    check_mx_delay 5
    dns_available yes
    report_safe 1
    use_terse_report 1
    use_bayes 1
    score RCVD_IN_BL_SPAMCOP_NET 0 2.25 0 1.50
    score RCVD_IN_MAPS_RBL 1
    score FAKED_UNDISC_RECIPS 2.899 2.694 2.800 2.700
    score GAPPY_SUBJECT 2.326 1.316 2.696 2.270
    score FROM_OFFERS 4.300 3.932 4.095 4.100
    score ADVERT_CODE 2.899 1.578 2.633 1.817
    score BAYES_60 0 0 1.250 1.250
    score BAYES_70 0 0 2.000 2.000
    score BAYES_80 0 0 4.250 4.250
    score BAYES_90 0 0 5.500 5.500
    score BAYES_99 0 0 7.750 7.750
    bayes_auto_learn 1
    bayes_auto_learn_threshold_nonspam 0.3
    bayes_auto_learn_threshold_spam 4.0
    auto_learn 1
    auto_learn_threshold_spam 5
    auto_learn_threshold_nonspam 0
    bayes_expiry_max_db_size 10240000
    skip_rbl_checks 0
    use_razor2 1
    razor_timeout 10
    use_dcc 1
    use_pyzor 1
    ok_languages all
    ok_locales all
    As you can see, only the 3 first lines are similar to SA-settings in CP. And it is very bad to check the server from time to time for new mailboxes and than copy /edit the user_prefs-file.
    I'm searching a way to make this automaticly on mailbox-creation.

    Greets from Germany
  7. jamesyeeoc

    jamesyeeoc Guest

    Hi, yes I know you were talking about per user. As far as I know/knew, it took the basic per user settings from that table. I have been unable to find any other file where it matched on those 3 items exactly, so unless they have it hard coded into their encrypted source scripts, the database table is the only place I know of that those entries exist exactly as Plesk uses on new user creation.

    Your sample user_prefs looks more like mine on the systems were I have replaced the Plesk version with the full SA. On servers which are running the standard Plesk SA, I never see it create nearly that many options by default.
  8. ProfiTiger

    ProfiTiger Guest

    Hi, this is "my sample", not the one by plesk ;-) I took loong time to build this file and to configure Plesk-SA to work fine with rules_du_jour, pyzor and so on... But now, nearly 99% of spam are detected. Everything is fine now, only this file has to be copied to each new mailbox. I wish, my plesk could create nearly that many options by default ;-)
  9. Griffith

    Griffith Guest

    Could you please explain how you have configured spamassassin to work with pyzor ++?
  10. ProfiTiger

    ProfiTiger Guest

    Not a big thing. I'm not a shell-freak, but I found a webpage with instruction bout it (sorry, forgott the url, can't find it quick), followed them did an install of pyzor and razor (http://razor.sourceforge.net/docs/install.php). After that, I modified local.cf and user_prefs. That's it...
  11. Griffith

    Griffith Guest


    I installed DCC and pyzor from ART.. worked ;)
  12. Jllynch

    Jllynch Regular Pleskian

    Nov 11, 2003
    Likes Received:
    Hi profitiger,
    I implemented your per user settings exactly as above but the mesages come up with the following for example;

    X-Spam-Status: Yes, hits=5.2 required=4.0 tests=BAYES_40,DCC_CHECK, HTML_MESSAGE,RAZOR2_CF_RANGE_51_100,RAZOR2_CHECK autolearn=no version=2.63

    Is that right for autolearn? I read your config to mean that when above 4 hits it would auto learn it. Or has it auto learned it with bayes_auto_learn? If so why isn't bayes_auto_learn mentioned in the header?

    Australian Hosting
  13. ProfiTiger

    ProfiTiger Guest

    Hi Jllynch,
    I'm not really sure, which of the two settings (bayes_auto_learn_threshold_spam 4.0, auto_learn_threshold_spam 5) is the right one for SA 2.63, because SA changed some settings on version 3.
    You can try to set "auto_learn_threshold_spam" to 4, perhaps this would help. But there are some special cases for SA to stop autolearn. From the SA-Homepage (Mail_SpamAssassin_Conf. ):
    bayes_auto_learn ( 0 | 1 ) (default: 1)
    Whether SpamAssassin should automatically feed high-scoring mails (or low-scoring mails, for non-spam) into its learning systems. The only learning system supported currently is a naive-Bayesian-style classifier.
    Note that certain tests are ignored when determining whether a message should be trained upon:
    - rules with tflags set to 'learn' (the Bayesian rules) - rules with tflags set to 'userconf' (user white/black-listing rules, etc) - rules with tflags set to 'noautolearn'Also note that auto-training occurs using scores from either scoreset 0 or 1, depending on what scoreset is used during message check. It is likely that the message check and auto-train scores will be different.
    bayes_auto_learn_threshold_spam n.nn (default: 12.0)
    The score threshold above which a mail has to score, to be fed into SpamAssassin's learning systems automatically as a spam message.
    Since I turned on automatic deletion of email with score >=4, I recieve no spammails for looking into the header for autolearn. I turned of deletion for testing purposes.
    *ding* there is one. Here is the information from header:
    X-Spam-Status: Yes, score=4.4 required=4.0 tests=DEAR_SOMETHING,
    RCVD_IN_DSBL,RCVD_IN_SORBS_DUL autolearn=spam version=3.0.4
    So I think, SA < 3.0 uses other settings than SA >=3. In my actual user-prefs, there is only "bayes_auto_learn_threshold_spam 4.0".
    I hope this helps (but I think not...)