Dear Matt Bullock,
> Would that bypass the rules completely? Part of the reason for the > upgrade was to put all the rules into effect to aid in catching a higher > percentage of spam. > > One thing is for sure though, I wont try the newer version until I > upgrade the memory to 512 MB
Well, the following are the statistics on a Athlon 2100XP with 512MB of RAM(no bayes used, as bayes db size was almost zero):
$time spamassassin < /tmp/spam.test > /dev/null
real 0m2.918s user 0m1.133s sys 0m0.074s
$time spamassassin --siteconfigpath=/etc/MailScanner/SA/RulesDuJour/ < /tmp/spam.test > /dev/null
real 0m4.919s user 0m3.028s sys 0m0.202s
$wc -l /etc/MailScanner/SA/RulesDuJour/* 1342 /etc/MailScanner/SA/RulesDuJour/99_FVGT_Tripwire.cf 253 /etc/MailScanner/SA/RulesDuJour/antidrug.cf 11278 /etc/MailScanner/SA/RulesDuJour/bigevil.cf 2487 /etc/MailScanner/SA/RulesDuJour/bogus-virus-warnings.cf 182 /etc/MailScanner/SA/RulesDuJour/evilnumbers.cf 212 /etc/MailScanner/SA/RulesDuJour/random.current.cf
15754 total
So, to compare each line of the mail to those 15k odd lines in the rulesets directory, the total time to spam check a mail increases from 2.9 seconds to 4.9 seconds, a difference of 2 seconds. This may get more as the load in the mail server increases.
So, either you can go for more RAM+Clock speed or you can try increasing the SA timeout in /etc/MailScanner/MailScanner.conf:
SpamAssassin Timeout = 40
But, a timeout period of 40 seconds is more than enough.If it takes longer than that, the server may probably be feeling extremely high load, in which case, bayes is faster than rulesets.
Eager to have your feedback on this, cheers, Karthikeyan, S. -- S.Karthikeyan | Ph: +91 (0) 44 52166646 Fax: +91 (0) 44 52079957 Opencomputing Technologies | http://opencompt.com Server Side E-Mail Protection.
spam.test.gz
Description: Unix tar archive
