Bayes ate my disk space...

Julian Field mailscanner at ecs.soton.ac.uk
Tue Dec 28 19:29:00 GMT 2004


    [ The following text is in the "ISO-8859-1" character set. ]
    [ Your display is set for the "US-ASCII" character set.  ]
    [ Some characters may be displayed incorrectly. ]

Christian Campbell wrote:

> I recently turned on bayes auto-learning and have just recently run
> out of space on my / volume.
>
> I've tracked it down to the /root/.spamassassin directory.  It seems
> there are a bunch of files called bayes_toks.expireNNNN (where NNNN is
> a number).
>
> First question...do these files need to be there for bayes to continue
> operating?  I'm not familiar with these files, and have limited SA
> experience.

You can safely delete all the expireNNNN files.

> Second question....  Can these be moved somewhere else?  If so, how to
> I reconfigure SA to make this change?

You should be able to make the .spamassassin directory a soft-link to
somewhere with more space.

> Third question...  Is there any place to establish a limit as to how
> much space these files take up or a way to purge old info?

They are created by MailScanner timing out while waiting for
SpamAssassin to do its bayes auto expiry. When MailScanner times out,
these temporary files get left behind.
You have a few options, among them being
a) Increase the SpamAssassin timeout time in MailScanner.conf, so that
SA gets time to do its own bayes auto expiry.
b) Switch off the auto-expiry (set "bayes_auto_expire 0" in
spam.assassin.prefs.conf). Then switch on MailScanner's regular bayes
expiry runs (set "Rebuild Bayes Every = 86400" in MailScanner.conf).
That way MailScanner gets to control the expiry process, and runs it
once a day (up to once every 28 hours in fact, but don't worry about the
details).

If you want MailScanner to continue processing mail while the Bayes db
is being rebuilt, but just ignoring the bayes result until the rebuild
has finished, set "Wait During Bayes Rebuild = no" in MailScanner.conf.

Quite a lot of people actually prefer option (a). I prefer option (b)
myself, which is why I wrote it :-)

Can someone add this to the FAQ please? This has to be the 50th time
this has been explained (no insult intended, the mail archive is not
easy to search well for random things like this).

>
> Thanks,
>
> Christian
>
>
> ------------------------ MailScanner list ------------------------
> To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
> 'leave mailscanner' in the body of the email.
> Before posting, read the MAQ (http://www.mailscanner.biz/maq/)
> and the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).
>
> *Support MailScanner development - buy the book off the website!*


--
Julian Field
www.MailScanner.info
Buy the MailScanner book at www.MailScanner.info/store
Professional Support Services at www.MailScanner.biz
MailScanner thanks transtec Computers for their support

PGP footprint: EE81 D763 3DB0 0BFD E1DC 7222 11F6 5947 1415 B654

------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
'leave mailscanner' in the body of the email.
Before posting, read the MAQ (http://www.mailscanner.biz/maq/) and
the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).

Support MailScanner development - buy the book off the website!




More information about the MailScanner mailing list