Re: Berkeley or PBL?

Brian Burton <[email protected]> Thu, 02 Feb 2006 11:30:27 -0500
Newsgroups gmane.mail.spam.spamprobe.general
Message-ID <[email protected]>
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

David A. Lee wrote:
> I recommend totally AGAINST  BDB ... its so full of problems you will 
> get ulcers.
> PBL is very good, but I've recently made the switch to the internal 
> "hash" DB and am very happy.

Hash was introduced later.  It wasn't available in 0.9.  BDB 4.2 
"should" compile with 0.9 but I've always recommended PBL or hash 
instead of BDB for people.  Some people had corruption problems with BDB.


> The only major downside of "hash" DB is you cant actually get the terms 
> dumped out for debugging,
> you only get the hash-value of the terms.

One strength of the hash format is it's wicked fast.  Not that that's 
much of an issue for most people but can be nice if you have a lot of 
users.  Also the file size remains fixed.  Again that's best if you have 
lots of users.

David covered the drawbacks nicely.  Run cleanup consistently and 
aggressively to keep it from becoming more than around 1/2 full. 
Something like "spamprobe cleanup 2 7 8 14 100 90" might be good.

All the best,
++Brian

-----BEGIN PGP SIGNATURE-----
Version: PGP Desktop 9.0.4 (Build 4042)

iQEVAwUBQ+IzqzxRyEoJfXIFAQgpSwf/amb+vgv5sw50e9SC50NLo1+xiklvuxUA
GxNIRwa+YOY/FcphNS2sYHrPwGeR67+rDGH5BMWVB4789RKab6amuE5W0BnA5//M
kDCu//UqfE1hchCzaQYiFWWSw1yygIvlM7zfzvIGrwnOKlCbqJG/nQlSav3CdGcu
3UQntlZX7OjAPiINx6nStZtSqt7kXR6nZ4J2DvW4B0r5ClHPkWSazvH7ek/SxqN7
q+5otK1eH61eNqNf1Ds8kGREoMCsDvTiPSY33O6TFE+a2k+ttrhCQmhoRBzjeBmu
vUPO4MdguESsaMXpT1UeMhmnJ/MYVhvxpXbEomznENlakKJLM4atsA==
=6TUB
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642