-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, Today it noticed a high percentage of (what I considered) "false negatives" and very few postives. Checking mail headers I found out that most of the messages had no SA markup at all, i.e. was very probably never processed with SA. To make it short, spamc would not work and no spamd process running. Checking syslog I found out that spamd was started successfully and checked some messages:
Nov 8 19:26:59 jan root: spamd starting Nov 8 19:27:05 jan spamd[194]: server started on port 783 (running version 2.43 ) Nov 8 19:27:06 jan spamd[194]: connection from localhost [127.0.0.1] at port 32 784 Nov 8 19:27:06 jan spamd[359]: info: setuid to jk succeeded Nov 8 19:27:06 jan spamd[359]: processing message <[EMAIL PROTECTED]> f or jk:1000, expecting 3168 bytes. Nov 8 19:27:06 jan spamd[194]: connection from localhost [127.0.0.1] at port 32 787 Nov 8 19:27:06 jan spamd[367]: info: setuid to jk succeeded Nov 8 19:27:06 jan spamd[367]: processing message <Pine.LNX.4.44.0211062301160. [EMAIL PROTECTED]> for jk:1000, expecting 3906 bytes. Nov 8 19:27:07 jan spamd[194]: connection from localhost [127.0.0.1] at port 32 793 Nov 8 19:27:07 jan spamd[378]: info: setuid to jk succeeded Nov 8 19:27:07 jan spamd[378]: processing message <Pine.LNX.4.30.0211071145430. [EMAIL PROTECTED]> for jk:1000, expecting 4182 bytes. Nov 8 19:27:08 jan spamd[194]: connection from localhost [127.0.0.1] at port 32 798 Nov 8 19:27:08 jan spamd[392]: info: setuid to jk succeeded Nov 8 19:27:08 jan spamd[392]: processing message <Pine.LNX.4.44.0211061906340. [EMAIL PROTECTED]> for jk:1000, expecting 5316 bytes. Nov 8 19:27:08 jan spamd[367]: clean message (-0.4/5.0) for jk:1000 in 1.7 seco Then, not quite surprisingly: Nov 8 19:27:22 jan spamd[194]: hit max-children limit (10): waiting for some to exit A few lines later, spamd seems to have disappeared: Nov 8 19:27:28 jan spamc[703]: connect() to spamd at 127.0.0.1 failed, retrying (1/3): Connection refused I attached all SA related syslog messages in case I might have missed something important (gziped). There's nothing strange in fetchmail's or exim's log. Is this a know problem that spamd (or perl??) crashed? Should I file bugreport? Should a crashing spamd have left any traces I should have found (logs, core dumps)? Can this have something to do with my low 'max children' (10) limit? My machine could process more messages at a time, I just like it that way, leaving most possible CPU time to foreground processes. Jan -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: pgpenvelope 2.10.2 - http://pgpenvelope.sourceforge.net/ iD8DBQE9zB6RY6Nk2Nv6ZRcRAlpGAKCVKkAlZlObFvPaimTcmKGyotDTwACggUbC C1mRdcWOlWMcRabMo9etKzs= =kfv9 -----END PGP SIGNATURE-----
syslog.gz
Description: syslog