How do you guys deal with database performance and the huge number of tokens and signatures, aside from the usual db engine tweaks?
I'm migrating from SpamAssassin to DSPAM, and it works amazingly well in small-scale tests, but MySQL is just eating disk i/o for lunch after about a day of learning. I've tried Postgres, and it fell on its face even harder.
What would be considered reasonable specs (CPU, RAM, RAID level, FS type, etc) for an SQL server that would have to deal with millions of rows of data?
Any help would be greatly appreciated...
[link] [7 comments]