On Tue, 25 Nov 2003, Yevgeniy Miretskiy wrote: > Hello, > > sa-learn stopped learning messages. Debugging shows that it can > successfully tie Bayes db, extracts tokens, etc, but never actually > writes data to the database. > > I had a db corruption issue some time ago, so, this could very > well be remnants of that. > > Anyway, I'm trying to run "sa-learn -D --force-expire" because my db > grew to be very large (~100MB). > > No matter what bayes_expiry_max_db_size is set to (I tried anything from 100K to > 3Mil), > sa-learn reports, after running for quite some time: > bayes: couldn't find a good delta atime, need more token difference, skipping > expire. > > here is the output of "sa-learn --dump magic" > 0.000 0 2 0 non-token data: bayes db version > 0.000 0 50976 0 non-token data: nspam > 0.000 0 1050 0 non-token data: nham > 0.000 0 2895991 0 non-token data: ntokens > 0.000 0 1013932420 0 non-token data: oldest atime > 0.000 0 1100005170 0 non-token data: newest atime > 0.000 0 1069789057 0 non-token data: last journal sync atime > 0.000 0 1069787463 0 non-token data: last expiry atime > 0.000 0 0 0 non-token data: last expire atime delta > 0.000 0 0 0 non-token data: last expire reduction count > > > Any suggestions on how to force expiration run in this case?
You've got a corrupted or poisoned database. Looking at that 'non-token data: newest atime' line, you've got some tokens in there with a date from the future. That 'non-token data: oldest atime' date is a fair ways in the past too, so my guess is that those bogus dates are throwing off the expire code. Something similar to that happened to me, (dates extreme and whacko) couldn't expire it and finally had to dump the whole kit. Maybe worth writing some kind of DB editing tool to manually remove bogus records, easiest just to dump it. Dave -- Dave Funk University of Iowa <dbfunk (at) engineering.uiowa.edu> College of Engineering 319/335-5751 FAX: 319/384-0549 1256 Seamans Center Sys_admin/Postmaster/cell_admin Iowa City, IA 52242-1527 #include <std_disclaimer.h> Better is not better, 'standard' is better. B{ ------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk