This may be a bit of a stupid question but... I have all the config files in /usr/share/spamassassin/ eg 10_misc.cf and 20_body_tests.cf etc....
Where does /etc/mail/spamassassin/locla.cf fit into this? The config below should go where? Into the locla.cf? Some explanation of how this works would really be appreciated . Mark Quoting Matt Kettler <[EMAIL PROTECTED]>: > For high-volume I'd recommend a couple tweaks if your CPU load average is > going too high. > > 1) disable razor, as you already suggested, along with razor2, dcc, and > pyzor. > > 2) zero the scores for several, or all, of the DNSBLs. I personally took > the approach of zeroing the scores of DNSBL's for ones which score under > 0.5, and I'm not high volume. If you do leave any DNSBLs on, make sure > your spamd server has FAST access to a nameserver. I'd even go as far as > to recommend a caching nameserver running on the same server as the machine > running spamd if it doesn't already have a normal DNS server on it. > > 3) I'd also consider zeroing some of the "very low scoring" rules. > Particularly those with evals. My cutoff was anything less than 0.01, but > you might want to take a wider swath if CPU usage is a problem for you. > > 4) Reduce some timeouts and retry counts. This helps a lot in preventing > mail from backloging whenever a blacklist becomes unavailable or a email > with "no mx for from" arrives. > > The following are the score tweaks I use with 2.43, they outline the basic > concepts above, but you might want to do this to a greater extent than I > have. Your mileage may vary, but this might be a good start for what you > want to do. > > ----------------------------------- > #Timeout reductions > #the following reduces the time for a run of SA > #but risks timing out valid blacklist data. Still far better to > #skip a blacklist sometimes than to choke a mailserver. > rbl_timeout 10 > razor_timeout 10 > #the following reduces the time for a run of SA > #but risks claiming a from: has no mx when it > #does indeed have one. This is a bit more risky > #as it increases chances of FPs. > check_mx_attempts 2 > check_mx_delay 3 > > #disabling razor 1, but allow razor2, that works much better > score RAZOR_CHECK 0 > #score RAZOR2_CHECK 0 > > #first, zero some low-scoring DNS lists. <0.5 is NOT worth cost of lookup. > score RCVD_IN_OSIRUSOFT_COM 0 > score X_OSIRU_DUL_FH 0 > score X_OSIRU_SPAMWARE_SITE 0 > #I don't like the idea of checking these, and the scores aren't very high > score RCVD_IN_MULTIHOP_DSBL 0 > score RCVD_IN_UNCONFIRMED_DSBL 0 > > #default scores too low to be worth running > # kill anything > 0.01 > #score EXPECT_TO_EARN 0.008 > #score SUPERLONG_LINE 0.009 > #score MIME_BOUND_DIGITS_3 0.009 > #score BIG_BUCKS 0.003 > #score GAPPY_TEXT 0.005 > #score RESERVES_RIGHT 0.008 > #score USER_AGENT_OUTLOOK -0.006 > #score GIFT_CERTIFICATE 0.004 > > score EXPECT_TO_EARN 0 > score SUPERLONG_LINE 0 > score RISK_FREE 0 > score MIME_BOUND_DIGITS_3 0 > score BIG_BUCKS 0 > score GAPPY_TEXT 0 > score RESERVES_RIGHT 0 > score USER_AGENT_OUTLOOK 0 > score GIFT_CERTIFICATE 0 > > At 01:00 PM 10/23/2002 -0600, Gustave Eiffel wrote: > > >Hello all, > > > >I am using spamc for all users through /etc/procmailrc on 4 servers and > >spamassassin works great. The problem is that it loads up the CPU to 4.0 > and > >often 10.0 or above. How can I reduce this? Eliminate razor check is one > I > >have seen on some other posts. What would be the best to try? > ------------------------------------------------------- This sf.net email is sponsored by: Influence the future of Java(TM) technology. Join the Java Community Process(SM) (JCP(SM)) program now. http://ads.sourceforge.net/cgi-bin/redirect.pl?sunm0002en _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
