You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Lucio Chiappetti <lu...@lambrate.inaf.it> on 2011/05/26 18:57:15 UTC
autolearn Re: major upgrade of spamassassin
On Wed, 18 May 2011, Lucio Chiappetti wrote:
> We are in the process of doing a long awaited and overdue upgrade of our
> servers (from Suse 9.2 to 11.4), which involves upgrading the (bundled)
> spamassassin (used with sendmail and amavis milter) from 3.0 to 3.3.
We progressed in our migration. We have now a machine running with the new
configuration, duly trained on ham and spam collected in the last weeks,
and we diverted to it a fraction of our incoming messages (we repointed
the secondary MX for one of our two domains).
It is catching and quarantining some spam (so far all ham is still
transiting through the primary MX) and we see that in the header
X-Spam-Status keyword, which has changed format, there is at the end
one of these
autolearn=unavailable
autolearn=no
autolearn=spam
The first one seems to occur for very high score spam, the second for
relatively low score one, and the third for mid-score spam similar to what
previously seen as "no"
Is any of those indicating a malfunction in autolearn ?
--
------------------------------------------------------------------------
Lucio Chiappetti - INAF/IASF - via Bassini 15 - I-20133 Milano (Italy)
------------------------------------------------------------------------
The reasonable man tries to adapt himself to the world
The unreasonable man tries to adapt the world to himself
Therefore all progress depends on the unreasonable man (G.B.Shaw)
------------------------------------------------------------------------
For more info : http://www.iasf-milano.inaf.it/~lucio/personal.html
------------------------------------------------------------------------
Re: autolearn Re: major upgrade of spamassassin
Posted by Bowie Bailey <Bo...@BUC.com>.
On 5/26/2011 12:57 PM, Lucio Chiappetti wrote:
> We progressed in our migration. We have now a machine running with the
> new configuration, duly trained on ham and spam collected in the last
> weeks, and we diverted to it a fraction of our incoming messages (we
> repointed the secondary MX for one of our two domains).
>
> It is catching and quarantining some spam (so far all ham is still
> transiting through the primary MX) and we see that in the header
> X-Spam-Status keyword, which has changed format, there is at the end
> one of these
>
> autolearn=unavailable
> autolearn=no
> autolearn=spam
>
> The first one seems to occur for very high score spam, the second for
> relatively low score one, and the third for mid-score spam similar to
> what previously seen as "no"
>
> Is any of those indicating a malfunction in autolearn ?
http://wiki.apache.org/spamassassin/AutolearningNotWorking
--
Bowie