On Mon, 2004-08-09 at 15:44, Matt Kettler wrote: > At 03:00 PM 8/9/2004, AltGrendel wrote: > >I started having problems with the bayse database always hitting on > >BAYES_99 on all inbound email. So thinking that it had been poisoned, I > >decided to rebuild it with my corpus. > > > >Instead of rebuilding, I get the following: > > Taking one step back.. what did you do BEFORE you called sa-learn to try to > rebuild? > > (ie: exactly what did you do with your old bayes database)
That's the confusing part, I've been out for a month (foot surgery) and everything has been on automatic. This all started this past Thursday as near as I can figure. I was running SA 2.6.3 and SpamCopURI 0.19. So as the spamc user, I ran check_bayes_db and got the following: 0.000 0 0 0 non-token data: db format = on-the-fly probs, expiry, scan-counting 0.000 0 0 0 non-token data: nspam 0.000 0 0 0 non-token data: nham 0.000 0 0 0 non-token data: ntokens 0.000 0 0 0 non-token data: oldest age 0.000 0 0 0 non-token data: current scan-count 0.000 0 0 0 non-token data: last expiry scan-count 0.500 5 0 4139 doctors 0.500 1 0 29682 splintery 0.500 13 0 2816 themselves 0.500 1 0 5466 cherish 0.500 1 0 3274 managing 0.500 7 0 5462 created 0.500 7 0 4339 N:H*M:OEANbNN 0.500 20 0 4297 Prozent 0.500 6 0 3848 health-care 0.500 8 0 4297 al-Harbi's 0.500 46 0 20293 registered 0.500 2 0 51037 68.00 0.500 2 0 43018 retrorocket 0.500 2 0 50095 H*M:sk:0vNrYrr 0.500 1 0 31796 sk:pS3wbN3 0.500 4 0 1257 mature 0.500 3 0 43795 Gamache 0.500 1 0 31995 H*c:sk:1uhZLYr 0.500 1 0 2906 H*r:SMTPD32-7.12 0.500 81 0 20370 Opt-out 0.500 2 0 58226 U*7xxUw0IXhQUi <snip out excessive tokens> After that, I tried sa-learn --rebuild, then sa-learn --force-expire. I have also moved the files off to a subdirectory and tried to build a new db with the same commands from the previous email. -------------------------------------------------------- Some parts of the past must be preserved, and some of the future prevented at all costs. --------------------------------------------------------
