Mauro Chiarugi - ASP Srl wrote:

> Hi,
> i'm new to this mailing list..
> I'd like to know if Sloney is what that i need.. The problem is:
> We need to replicate a database in three location because of the slow 
> band avaible
>
> -----                     -----
>   1 | ------Internet------| 2 |
> -----           |         -----
>               -----
>               | 3 |
>               -----
>
> In each site where postgresql is installed, a lot of client make 
> connections to PG to read data.. but only few transactions happend to 
> store data on the db.. So, we need a solutions to synchronize all 
> database when data changes..
> I've read that Sloney has a master/slave design, with only a master.. 
> what does it means? The master is the only server in which data 
> changes or not?

No, that is exactly the design that Slony-I DOESN'T have.  It is NOT a 
"master/slave" system.

It involves a "subscription" design, where data has an origin (which is 
a single node), and where multiple nodes can subscribe to updates to 
that data source.

Updates can only be put in at the origin node.

> Are there any other solutions? Does it run with postgresql 8 and windows?

PostgreSQL 8 has not been released, so it is dangerous to make promises 
on that, though testing has taken place with various CVS checkouts of PG 
"head."

No one has done any porting work of Slony-I tools to Windows.  Someone 
(I do not recall who) has compiled the triggers for Windows(tm) and was 
able to get replication to run so long as they ran slon (the daemon that 
performs replication) and slonik (the configuration control tool) on a 
Unix-like system.