|
Hi!
I have a doubt...
In this same scenario, I need that when cluster stand alone, it continue working in read-write mode; when to start it (with rsync mode) the data in the all clusters have consistent (with data updated in the cluster that was down).
Is it possible?
If yes, what the configuration?
Thanks!
2006/8/23, Diego Gutierrez <dgutierr@xxxxxxxxxxxxxxx>:
Hi Eduardo.
If I understand you correctly, you have 9 database nodes to replicate using
VPN links, each of the nodes accessed only by local clients.
If so, you don't need any load balancers. Your minimum configuration would
then be: 9 pgcluster servers (db engines) and just 1 replicator, I suggest
you place it in the network central point.
Optionally, you can configure more than 1 replicator, in a master/slave
fashion (the docs are more or less clear about that).
Have in mind, however, that using such a configuration would have the
following drawbacks:
- Performance will be limited by the bandwidth connecting the servers
- If any of the nodes dies unexpectedly, that node's clients will be left
without service
Cheers,
DiegoG
On Wednesday 23 August 2006 17:51, Eduardo Gaio wrote:
> Hi!!!
>
> What configuration (cluster, replication, load balancer?) it's necessary to
> this scenario, using PGCluster: 9 database (PostgreSQL) in remote location,
> connected by VPN? Note: just replication is need; each database work in
> local mode and replicate to the others.
_______________________________________________
Pgcluster-general mailing list
Pgcluster-general@xxxxxxxxxxxxx
http://pgfoundry.org/mailman/listinfo/pgcluster-general
|
