Title: spamassassin time outs

We are running SA 3.01 and the latest MailScanner.  Besides the delivered SA rule sets, we have the following:

r--r--r--    1    31854 May 31 19:24 70_sare_adult.cf
-r--r--r--    1     3927 Apr 24  2004 70_sare_bayes_poison_nxm.cf
-r--r--r--    1   211390 Oct  3 20:18 70_sare_header.cf
-r--r--r--    1   103436 Sep 12 20:22 70_sare_html.cf
-r--r--r--    1    11559 Sep 16 12:45 70_sare_oem.cf
-r--r--r--    1    17548 Aug 16 09:38 70_sare_random.cf
-r--r--r--    1    19899 Oct  3 20:20 70_sare_specific.cf
-r--r--r--    1     7023 Oct  8 00:45 70_sare_spoof.cf
-r--r--r--    1     7937 Nov 11 00:45 70_sc_top200.cf
-r--r--r--    1    13211 May 11  2004 72_sare_bml_post25x.cf
-r--r--r--    1    57580 May  6  2004 99_FVGT_Tripwire.cf
-r--r--r--    1    10147 May  1  2004 99_sare_fraud_post25x.cf
-r--r--r--    1     8043 Nov 12 23:12 bakerbotts.cf
-r--r--r--    1   100721 Nov  7 14:13 bogus-virus-warnings.cf
-r--r--r--    1    23470 May 25 11:56 chickenpox.cf
-r--r--r--    1    18052 Nov  2 00:45 evilnumbers.cf
-r--r--r--    1     4883 May 25 11:03 random.current.cf
-r--r--r--    1     3880 Apr 27  2004 weeds.cf

The bakerbotts.cf is our own custom rule set.

We process between 100k and 165k emails a day on two primary servers, with 3 overflow servers.  Since we have implemented 3.01, we have experienced a higher number of timeout messages from SA.  Yesterday 487 and the day before 544.

Nov 15 15:57:00 dalgate-inside MailScanner[8379]: SpamAssassin timed out and was killed, failure 1 of 20
Nov 15 15:58:25 houmx01-inside MailScanner[1377]: SpamAssassin timed out and was killed, failure 1 of 20

Does the software need more CPU resources; or do we have too many add-on rule sets; or is there another explanation?

We are using pyzor, but not dcc or razor2.  Using all three was slowing down the spam check process.

Any ideas would be greatly appreciated.

Thanks,
Donald

Reply via email to