On Thu, 30 Sep 2004 18:42:59 +0200
Sune Kloppenborg Jeppesen <[EMAIL PROTECTED]> wrote:

> On Thursday 30 September 2004 18:05, Robin Lynn Frank wrote:
> > While I can see the advantage of keeping awl and prefs in a sql
> > database, I can't see an advantage to keeping bayes data in a sql
> > db.
> >
> > Can someone point out an advantage?  Would there be any disadvantage
> > in keeping everything except bayese in sql?
> You could have two front end servers using the same Bayes backend with
> SQL. This makes autolearning work easily across multiple servers.
> 
An interesting thought.  I'm not sure we'd need it, but I'll remember
it.  The reason I brought this up is because we'd once tested dspam with
mysql and found that mysql seemed more likely to slow down considerably
as disk i/o demands rose.  It appeared to me that berkeleydb was less
touchy that way, though I suspect issues of lock contention might be 
greater.

-- 
Robin Lynn Frank
Director of Operations
Paradigm-Omega, LLC
http://www.paradigm-omega.com
==============================
Sed quis custodiet ipsos custodes?

Attachment: pgpE7Ad9dV9fy.pgp
Description: PGP signature

Reply via email to