On 10/21/09 8:01 PM, Dale Puch wrote:
> Not a network or SQL engineer. But a google search turned up this.
> http://pgcluster.projects.postgresql.org/feature.html From the mailing list
> I see the code is up to 1.9 even if the page isn't updated.
>
> There also seemed like several basic configurations that can be used with
> different good and bad points.
>
um, no.
synchronous multi-master requires that the database servers be in close
physical proximity,
as latency is deadly. async allows for geographical distribution, but
requires application
design to take it into account because of conflict resolution problems.
there is no silver
bullet. anyone who claims otherwise is either poorly informed or a salesman.
database replication is a non-trivial engineering exercise, which is why
people get paid
a lot of money to design and implement solutions, and database vendors
can charge so
much money for solutions that aren't necessarily even all there.
richard
(not a database administrator but as a developer, i annoy a bunch of
them every day)
_______________________________________________
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us
