[Sqlgrey-users] Major updates to sqlgrey in CVS.

[Dan F. <da...@ha...>]

Allright..

Ive been putting this update off for a while, but its finally committed 
to CVS.
It adds 3 new features. 1 small and 2 big ones.

Ive tried to make the patch as un-intrusive as possible. Everything 
should work as it always has if you do not enabled theese features.
My production environment is  is running with all theese features 
enabled and has for several months without problems, but id still like 
it if someone would test this version to confirm that. Id like 
confirmation that it runs like it always has for ppl who dont use the 
new features. If someone tries the new features id love to get feedback 
on that as well.
(Again, it HAS been thoroughly tested in my setup).

New features:
1. reject_code option.
By setting this in sqlgrey.conf you can change the default "450" 
response to eg. 451.

2. DBClustering
Allows for use of several sql-servers for reading, while still writing 
to just 1 master server.
See README.DBCLUSTER
Tested and running in production on MySQL's. If anyone tries this on 
other SQL brands let me know how it works out.

3. Discrimination (aka. Selective Greylisting)
Alternative approach to greylisting. Allows you to specify a set of 
rules (as regular expressions) that decides whom can bypass greylisting 
entirely, based on characteristics in eg. reverse-dns name, HELO string, 
sender address ect.
See READ.DISCRIMINATION

(Remember; you have to checkout CVS version, not download the packaged 
1.7.3)

Any feedback is greatly appreciated. :)

- Dan Faerch