Drillbits can be installed on all nodes or any subset. Drill prefers to have colocated nodes for better performance but this isn't a requirement (e.g. Drillbit on same node as data node or regionserver).
On Sun, May 3, 2015 at 7:55 AM, George Lu <[email protected]> wrote: > Hi all, > > For the distributed mode installation, I have some questions, > > 1) Do I need to install Drill to each Hadoop Data Node or HBase > RegionServer? Or I can install to any nodes in the cluster? > > 2) From the "Deploying Apache Drill in a Clustered Environment" in the > confluence page, mentions below: > "If you are using external JAR files, edit drill-env.sh, located in > /opt/drill/conf/, and define HADOOP_HOME:" > > What is external JAR files? What is the purpose if I set the HADOOP_HOME? > > Thanks! > > George Lu >
