Accumulo also has a pluggable load balancer that could be used to balance
with rack and data-center awareness. This can be done using the currently
released versions of Accumulo.
What problem are you trying to solve, exactly? Are you concerned about
durability and availability issues like disaster recovery in the event of a
datacenter failure, or more about optimizing performance under the
constraint of non-uniform network topology?
On Tue, Jan 21, 2014 at 11:41 AM, Keith Turner <[EMAIL PROTECTED]> wrote:
> I started experimenting w/ a multi-level rack aware batch writer, see
> ACCUMULO-551. I even prototyped the code (there is a github link in the
> ticket). For this I used the hadoop code and config to make decisions
> about rack locality. I never got around to seriously testing this
> On Mon, Jan 20, 2014 at 4:56 PM, Mastergeek <[EMAIL PROTECTED]>
> > I realize this is a heavy topic, but I was curious to know what
> > implementations would have to be added to a build of Accumulo to make it
> > Rack and Datacenter aware? I found this
> > https://issues.apache.org/jira/browse/ACCUMULO-378 but it has not been
> > updated in some time. My main interests, for now at least, revolve around
> > understanding the components necessary to realize a rack and datacenter
> > aware Accumulo.
> > - Jeff
> > -----
> > --
> > View this message in context:
> > Sent from the Developers mailing list archive at Nabble.com.