bayes expire tokens

[Kai Schaetzl]

    [ 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. ]

Brady Tucker wrote on         Tue, 8 Mar 2005 18:35:00 -0600:

> I'm assuming that the bayes 
> DB gets so big that doing the expire job only once/day it takes to long to 
> finish and it gives up and tries again repeatedly ?

Exactly, that's the problem. Additionally, if your db builds up over a long 
time and expiry eventually kicks in it's quite possible that it cannot find a 
good expiry delta because of the algorithm they use for this. That can extend 
the run to several minutes and can even fail completely to expire. Running the 
expiry regularly prohibits building up of token data which could spoil the 
expiry. (The algorithm for expiry takes time of last run, last delta, expected 
reduction and one or two other things into account and if it doesn't produce a 
sensible figure which could match the expected expiry maximum SA will try to go 
thru ten trial loops which can quite often fail.)

  (even when running in 
> tmpfs)????  I don't know why..

There's a timeout for SA in mailscanner.conf. As I understand if that occurs 
the operation gets called off. Which means expiry also gets called off in the 
middle of operation.



Kai

-- 
Kai Schätzl, Berlin, Germany
Get your web at Conactive Internet Services: http://www.conactive.com
IE-Center: http://ie5.de & http://msie.winware.org

------------------------ 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!