Re: Replication

Rodrigo,

Rodrigo Botana wrote:

> There is any way to have a replication (master -> slave) whithout
> install anything on the slave computer. I ask this because i would
> like to have my website (that don't have any replication service
> avaliable) with the same information of my internal system.


In theory, the slon daemon of Slony-I (http://slony.info/) can be run on
a separate machine from the database to which it connects (it just uses
a libpq connection). With this technique, you could run the website's
slon daemon on your local box, along with the slon daemon for the
"master" database. YMMV on how well it works with the slon-->postmaster
connection over a WAN (it is typically over a local socket).

Depending on your requirements, however, Slony may be overkill. What is
the size of your data set? How many seconds/minutes/hours/days can your
website afford to be out of sync with the data on your internal system?
If you could be as much as say, a day out of date, and your data set is
less than 100MB, and you have a pretty fast connection between your
internal system and the website, you might get away with a nightly
pg_dump, piped to a psql process that connects to the pgsql server on
your website.

There are many ways to skin the replication cat, and the right one
depends on your requirements. There is no one universal solution that's
appropriate for everyone, but with more information, we could be of
better help.

Best Regards,

Bill Montgomery