Alex,
If you have 2 different sets of data, you might want to replicate both
ways in separate data bases. I mean set up on both machines 2 data
bases, one master and one slave on both machines. Then you can use 2
slony replication sets to replicate in both directions. This way if one
of the servers fail, you have both DBs on the other one.
I think you can also replicate in both directions in the same data base,
except you can't have a single table replicated both ways, you should
have each table being part of just 1 replication set. So if you have the
same tables in both application, then you should use separate data
bases.
I didn't try this with slony, but from what I have seen from how it's
set up I guess it should work.
Cheers,
Csaba.
On Fri, 2006-07-14 at 13:55, Alexander Bluem wrote:
> Hello,
>
> I have a certain setup, so that two computers are running nearly
> identical databases: identical setup, tables, users and permissions,
> only the contents differ. Now I'd like to keep them in sync, WITHOUT an
> extra machine, hence master-slave setup. The problem is, that either one
> could fail eventually. Either one or the other machine (with the same
> database) get data, but not yet both at the same time. This is some sort
> of load balancing.
> Is there software out there that "rsyncs" database tables both ways? Or
> will I have to write scripts for this task? I've already taken a look at
> Slony but it is unfortunately a "master to multiple slaves". And I want
> both (or maybe in future three) machines to communicate with each other.
> That means one gets data, and sends it to other machines running the
> same db.
>
>
> Cheers,
> Alex
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
