On lør 03 okt 2009 23:41:41 CEST, "Steven W. Orr" wrote
Thank you. I am still confused in one area:
no problem
These scripts do not touch the bayes_token table, and it is this table that has by far the most number of rows.
i do not touch it for one reason is that it will autoflush oldest tokens, if the db gets to big you simply have to much tokens know in the db, its not a error
I currently have over 230000 rows in that table. Do I manage these myself,
no this is part of how bayes works
or is there something that is supposed to make this happen automatically?
nope, my setup above is all needed to make it optimized, i could make a bug on this for 3.3.x but it will be nice others can confirm if i miss something :)
I admit that I am confused by the man page for sa-learn because it seems to suggest that "expiry" (whatever that is) is performed there, but I just don't see anything that says exactly what to do. Also, the man page refers to a "journal" that I know nothing about.
this is for non mysql setup imho -- xpoint
