4.26.7, bayes rebuild, confused.

Julian Field mailscanner at ecs.soton.ac.uk
Sun Feb 1 21:33:46 GMT 2004


At 20:39 01/02/2004, you wrote:
>Julian,
>
>    I tried an upgrade from 4.25-14 to 4.26.7.  Setup: Sol 9,
>SA 2.63.  I got gobs of:
>
>    Skipping SpamAssassin while waiting for Bayes database to rebuild

This should only happen every day or so. Depends on your setting of
"Rebuild Bayes Every".
How often do you get a bunch of these?

>messages with 4.26.7.  The zero byte file /tmp/MS.bayes.rebuild.lock
>appeared after I restarted MS, and it never seems to go away.

That's fine. It's a lock file that each of the child processes will
maintain a shared lock on.

>   I
>tried things with both "Rebuild Bayes Every = 0" and with this set
>to 86400.  Same deal.  I'm afraid to try "Wait During Bayes Rebuild = yes"
>because if it never rebuilds then no mail gets delivered, right?

You'll soon see.


>A rebuild should only take a few seconds, right?

It can take a minute or two if your Bayes database is quite large.

>I tried stopping MS, removing /tmp/MS.bayes*, restarting -- still no
>luck.  I've fallen back to 4.25-14 for the moment.
>
>BTW, I have a cron job to do bayes spam/ham learning with
>
>   $SALEARN --prefs-file=$PREFS --rebuild  --force-expire
>
>at the top.  Should I still do this rebuild and force-expire in this
>script?

No. My scheduled rebuild is there to replace this. It is designed to solve
the bayes_toks.new problem by locking out SpamAssassin during the rebuild
without causing SA to just timeout. Instead of SA timing out, it skips it
or waits for it to complete.
--
Julian Field
www.MailScanner.info
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



More information about the MailScanner mailing list