You could also add SPARK_MASTER_IP to bind to a specific host/IP so that it won't get confused with those hosts in your /etc/hosts file.
Thanks Best Regards On Fri, Mar 13, 2015 at 12:00 PM, Du Li <[email protected]> wrote: > Hi Spark community, > > I searched for a way to configure a heterogeneous cluster because the need > recently emerged in my project. I didn't find any solution out there. Now I > have thought out a solution and thought it might be useful to many other > people with similar needs. Following is a blog post I just wrote. Feedback > is welcome. > > The capability to accommodate heterogeneous hardware has already been > there in Spark. We learn how to use it. Oscar goes to those contributors. > :-) > > Du > > How to set up a Spark cluster with heterogeneous hardware > <http://scala4fun.tumblr.com/post/113486166952> > > > > > > > How to set up a Spark cluster with heterogeneous hardware > <http://scala4fun.tumblr.com/post/113486166952> > Apache Spark (http://spark.apache.org) is a big data processing platform > that has become widely adopted across the industries in just a few years. > The strength of Spark is to make use of the main memory that is getting > larger and cheaper to support low-latency use cases. In a pra... > View on scala4fun.tumblr.com > <http://scala4fun.tumblr.com/post/113486166952> > Preview by Yahoo > > > >
