On Thu, 23 Jun 2011 10:26:48 +0200 Matus UHLAR - fantomas <uh...@fantomas.sk> wrote:
> On 21.06.11 14:19, David F. Skoll wrote: > >InnoDB may be better, but it's not as fast as CDB, and certainly > >slower than a local CDB file. > did you make any measurementsor are you just guessing? I did not measure InnoDB specifically, but I'm very confident. > > Just the TCP round-trip time will make MySQL slower than local CDB > > files. > spamd caches the DB connection, does it cache open CDB's? We don't use spamd. And it's not necessary to cache open CDB's because opening a CDB is extremely cheap: It's an open() followed by a mmap(). > I don't mean having the files in the memory cache, but caching the > DB indexes, since opening files is usually the slowest part, > similarly to using spamd vs, calling spamassassin binary each time... Again, in our measurements, CDB beat everything else we tried by a huge margin (BDB, PostgreSQL.) > > Furthermore, the design of CDB is > >inherently lock-free so there are no contention problems for multiple > >readers. > yes, the cdb has to be rebuilt every time from scratch. Indeed. I mentioned that. [...] > If you want to scale local files, you can scale local MySQL servers > the same way. Yes, it's a bit different, but proper scaling may not > be a good argument here. I've mentioned why I would never use MySQL. In my opninion, MySQL is defective by design. Furthermore, I also mentioned that SQL is overkill for simple key/value lookups that make up the Bayes workload. Regards, David.
