You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by David Jones <dj...@ena.com> on 2018/01/19 14:27:10 UTC
Re: Autolearn says it learned but dump magic stays at zero
On 01/19/2018 08:12 AM, Heiler Bemerguy wrote:
> Hi guys,
>
> I'm new to the list so pardon any stupidity I may say.. lol
>
> I'm using SpamAssassin 3.4.1 with Postfix 3.1.6 on Debian 9.
>
> ii spamassassin *3.4.1-6+deb9u1 * all Perl-based spam
> filter using text analysis
>
> local.cf:
>
> use_bayes 1
> bayes_auto_learn 1
> bayes_auto_learn_threshold_nonspam -1.0
> bayes_auto_learn_threshold_spam 3.0
> bayes_auto_expire 1
>
> I had cleared bayes database with --clear some days ago and had
> restarted spamassassin service. Today I saw some autolearning on
> mail.log, but all the "dump magic" values are still 0
>
> see: *spamd: result: Y 8 -
> DEAR_FRIEND,FREEMAIL_FROM,FREEMAIL_REPLYTO_END_DIGIT,LOTS_OF_MONEY,MONEY_FRAUD_5,RCVD_IN_DNSWL_NONE,RCVD_IN_MSPIKE_H3,RCVD_IN_MSPIKE_WL,SUBJ_ALL_CAPS,TRACKER_ID,T_DKIM_INVALID,T_HK_NAME_FM_MR_MRS
> scantime=7.1,size=4797,user=debian-spamd,uid=110,required_score=2.3,rhost=::1,raddr=::1,rport=36112,mid=<CA...@mail.gmail.com>,autolearn=spam
> autolearn_force=no,shortcircuit=no*
>
> and: *spamd: result: . -5 -
> MSGID_FROM_MTA_HEADER,RCVD_IN_DNSWL_HI,RCVD_IN_MSPIKE_H3,RCVD_IN_MSPIKE_WL
> scantime=6.8,size=8775,user=debian-spamd,uid=110,required_score=2.3,rhost=::1,raddr=::1,rport=45058,mid=<15...@spamassassin.apache.org>,autolearn=ham
> autolearn_force=no,shortcircuit=no*
>
> BUT:
>
> root@mailer:~# sa-learn --dump magic
> 0.000 0 3 0 non-token data: bayes db version
> 0.000 0 0 0 non-token data: nspam
> 0.000 0 0 0 non-token data: nham
> 0.000 0 0 0 non-token data: ntokens
> 0.000 0 0 0 non-token data: oldest atime
> 0.000 0 0 0 non-token data: newest atime
> 0.000 0 0 0 non-token data: last journal
> sync atime
> 0.000 0 0 0 non-token data: last expiry atime
> 0.000 0 0 0 non-token data: last expire
> atime delta
> 0.000 0 0 0 non-token data: last expire
> reduction count
>
> I think most numbers should be non-zero, right? Why it seems to be not
> storing what it learned?
>
> Any comments will be appreciated.
>
>
> --
> Atenciosamente / Best Regards,
>
> Heiler Bemerguy
> Network Manager - CINBESA
> 55 91 98151-4894/3184-1751
>
What Bill Cole said plus make sure you are running your commands as the
same user as the spamd "glue" that is calling/launching SpamAssassin.
Normally this is the best way to make sure you are using the same
configuration as the "glue" -- spamd in this case so maybe this is the
postfix user. I have never used spamd so I don't know for sure.
# ps -elf | grep spamd
--
David Jones
Re: Autolearn says it learned but dump magic stays at zero
Posted by David Jones <dj...@ena.com>.
On 01/19/2018 08:56 AM, Heiler Bemerguy wrote:
>
> Em 19/01/2018 11:27, David Jones escreveu:
>> On 01/19/2018 08:12 AM, Heiler Bemerguy wrote:
>>> Hi guys,
>>>
>>> I'm new to the list so pardon any stupidity I may say.. lol
>>>
>>> I'm using SpamAssassin 3.4.1 with Postfix 3.1.6 on Debian 9.
>>>
>>> ii spamassassin *3.4.1-6+deb9u1 * all Perl-based spam
>>> filter using text analysis
>>>
>>> local.cf:
>>>
>>> use_bayes 1
>>> bayes_auto_learn 1
>>> bayes_auto_learn_threshold_nonspam -1.0
>>> bayes_auto_learn_threshold_spam 3.0
>>> bayes_auto_expire 1
>>>
>>> I had cleared bayes database with --clear some days ago and had
>>> restarted spamassassin service. Today I saw some autolearning on
>>> mail.log, but all the "dump magic" values are still 0
>>>
>>> see: *spamd: result: Y 8 -
>>> DEAR_FRIEND,FREEMAIL_FROM,FREEMAIL_REPLYTO_END_DIGIT,LOTS_OF_MONEY,MONEY_FRAUD_5,RCVD_IN_DNSWL_NONE,RCVD_IN_MSPIKE_H3,RCVD_IN_MSPIKE_WL,SUBJ_ALL_CAPS,TRACKER_ID,T_DKIM_INVALID,T_HK_NAME_FM_MR_MRS
>>> scantime=7.1,size=4797,user=debian-spamd,uid=110,required_score=2.3,rhost=::1,raddr=::1,rport=36112,mid=<CA...@mail.gmail.com>,autolearn=spam
>>> autolearn_force=no,shortcircuit=no*
>>>
>>> and: *spamd: result: . -5 -
>>> MSGID_FROM_MTA_HEADER,RCVD_IN_DNSWL_HI,RCVD_IN_MSPIKE_H3,RCVD_IN_MSPIKE_WL
>>> scantime=6.8,size=8775,user=debian-spamd,uid=110,required_score=2.3,rhost=::1,raddr=::1,rport=45058,mid=<15...@spamassassin.apache.org>,autolearn=ham
>>> autolearn_force=no,shortcircuit=no*
>>>
>>> BUT:
>>>
>>> root@mailer:~# sa-learn --dump magic
>>> 0.000 0 3 0 non-token data: bayes db version
>>> 0.000 0 0 0 non-token data: nspam
>>> 0.000 0 0 0 non-token data: nham
>>> 0.000 0 0 0 non-token data: ntokens
>>> 0.000 0 0 0 non-token data: oldest atime
>>> 0.000 0 0 0 non-token data: newest atime
>>> 0.000 0 0 0 non-token data: last journal
>>> sync atime
>>> 0.000 0 0 0 non-token data: last expiry
>>> atime
>>> 0.000 0 0 0 non-token data: last expire
>>> atime delta
>>> 0.000 0 0 0 non-token data: last expire
>>> reduction count
>>>
>> What Bill Cole said plus make sure you are running your commands as
>> the same user as the spamd "glue" that is calling/launching
>> SpamAssassin. Normally this is the best way to make sure you are using
>> the same configuration as the "glue" -- spamd in this case so maybe
>> this is the postfix user. I have never used spamd so I don't know for
>> sure.
>>
>> # ps -elf | grep spamd
>>
>
>
> root@mailer:~# ps -elf | grep spamd
> 5 S root 11629 111283 4 80 0 - 44831 core_s 11:47 ? 00:00:20
> spamd child
> 4 S postfix 12150 930 0 80 0 - 20808 SyS_ep 11:51 ? 00:00:00
> pipe -n spamassassin -t unix user=debian-spamd argv=/usr/bin/spamc -f -e
> /usr/sbin/sendmail -oi -f ${sender} ${recipient}
>
>
Don't know for sure but try this:
root@mailer# su - debian-spamd -s /bin/sh
debian-spamd@mailer$ sa-learn --dump magic
See if that has better numbers in the output.
--
David Jones
Re: Autolearn says it learned but dump magic stays at zero
Posted by Heiler Bemerguy <he...@cinbesa.com.br>.
Em 19/01/2018 11:27, David Jones escreveu:
> On 01/19/2018 08:12 AM, Heiler Bemerguy wrote:
>> Hi guys,
>>
>> I'm new to the list so pardon any stupidity I may say.. lol
>>
>> I'm using SpamAssassin 3.4.1 with Postfix 3.1.6 on Debian 9.
>>
>> ii spamassassin *3.4.1-6+deb9u1 * all Perl-based spam
>> filter using text analysis
>>
>> local.cf:
>>
>> use_bayes 1
>> bayes_auto_learn 1
>> bayes_auto_learn_threshold_nonspam -1.0
>> bayes_auto_learn_threshold_spam 3.0
>> bayes_auto_expire 1
>>
>> I had cleared bayes database with --clear some days ago and had
>> restarted spamassassin service. Today I saw some autolearning on
>> mail.log, but all the "dump magic" values are still 0
>>
>> see: *spamd: result: Y 8 -
>> DEAR_FRIEND,FREEMAIL_FROM,FREEMAIL_REPLYTO_END_DIGIT,LOTS_OF_MONEY,MONEY_FRAUD_5,RCVD_IN_DNSWL_NONE,RCVD_IN_MSPIKE_H3,RCVD_IN_MSPIKE_WL,SUBJ_ALL_CAPS,TRACKER_ID,T_DKIM_INVALID,T_HK_NAME_FM_MR_MRS
>> scantime=7.1,size=4797,user=debian-spamd,uid=110,required_score=2.3,rhost=::1,raddr=::1,rport=36112,mid=<CA...@mail.gmail.com>,autolearn=spam
>> autolearn_force=no,shortcircuit=no*
>>
>> and: *spamd: result: . -5 -
>> MSGID_FROM_MTA_HEADER,RCVD_IN_DNSWL_HI,RCVD_IN_MSPIKE_H3,RCVD_IN_MSPIKE_WL
>> scantime=6.8,size=8775,user=debian-spamd,uid=110,required_score=2.3,rhost=::1,raddr=::1,rport=45058,mid=<15...@spamassassin.apache.org>,autolearn=ham
>> autolearn_force=no,shortcircuit=no*
>>
>> BUT:
>>
>> root@mailer:~# sa-learn --dump magic
>> 0.000 0 3 0 non-token data: bayes db version
>> 0.000 0 0 0 non-token data: nspam
>> 0.000 0 0 0 non-token data: nham
>> 0.000 0 0 0 non-token data: ntokens
>> 0.000 0 0 0 non-token data: oldest atime
>> 0.000 0 0 0 non-token data: newest atime
>> 0.000 0 0 0 non-token data: last journal
>> sync atime
>> 0.000 0 0 0 non-token data: last expiry
>> atime
>> 0.000 0 0 0 non-token data: last expire
>> atime delta
>> 0.000 0 0 0 non-token data: last expire
>> reduction count
>>
> What Bill Cole said plus make sure you are running your commands as
> the same user as the spamd "glue" that is calling/launching
> SpamAssassin. Normally this is the best way to make sure you are using
> the same configuration as the "glue" -- spamd in this case so maybe
> this is the postfix user. I have never used spamd so I don't know for
> sure.
>
> # ps -elf | grep spamd
>
root@mailer:~# ps -elf | grep spamd
5 S root 11629 111283 4 80 0 - 44831 core_s 11:47 ?
00:00:20 spamd child
4 S postfix 12150 930 0 80 0 - 20808 SyS_ep 11:51 ?
00:00:00 pipe -n spamassassin -t unix user=debian-spamd
argv=/usr/bin/spamc -f -e /usr/sbin/sendmail -oi -f ${sender} ${recipient}
--
Atenciosamente / Best Regards,
Heiler Bemerguy
Network Manager - CINBESA
55 91 98151-4894/3184-1751