Andrei Bintintan wrote:

> First... this is my first attempt to replication. Have no ideea how to 
> make it. Some of my problems probably will be stupid :(
>  

The stupid question is normally the pretty good one you didn't ask 
because you thought it was a dumb one :-)

> The scenario is the following:
> I have a main database on a server probably this will be the master.
>
> With this replication we want first to make a copy of this database 
> which will be at another location(another company) with 1 time/day 
> sincronization, so this will be a slave which will contain a part of 
> the database(so not all the tables) ~ asyncronus replication
> and
> we also would like to make a backup server in our local LAN which will 
> contain a complete backup of the main database in case of the server fail.
>
> We would like a free solution if possible, I read about Slony.
> It seems ok but facing problems: first I use Suse Linux 9.1, I 
> installed Postgresql 8.01 from rpm's and when I try to install Slony 
> it ask for sources. I didn't find the src's in the rpm, but I 
> downloaded the src's from postgresql site. Still Slony asks me to 
> configure and install those sources first and then try to install 
> Slony. Do I have to install manually postgre to use Slony? I would be 
> happy to leave the rpm installation running...
>  
> Although the documentation is a little "heavy" for me to understand, 
> does anyone know tipps and tricks?

Well, with version 1.0.5, which is what I'd have to recommend using if 
you need it today, Slony-I needs the full source tree for PostgreSQL, 
and it needs for the source tree to be configured for use with wherever 
your PG install is.  Furthermore, it installs a number of its components 
alongside the binaries (that we normally assume were installed from the 
PG source tree).

It is probably possible to use the RPM-based PG install, but once you 
have gone thru and configured and built everything from source, it's 
probably about as easy (and, since it means you control the exact 
provenance of all the components, it may be safer) to use the copy 
compiled from source.

With version 1.1, which has just gone into beta testing, it is no longer 
necessary to have the source tree; you just need to have the appropriate 
#include files and libraries available.

You can find an HTML copy of the documentation in the source tree at the 
URL below; that may be helpful...
-- 
<http://linuxdatabases.info/info/slony.html>