Spark sql is working, but it seems that without any cluster manager, it is possible? -
i have hadoop cluster 4 nodes(1master, 3slaves). , create hive tables files stored in hdfs. configure mysql hive metastore , copy hive-site.xml file inside conf folder of spark.
to install spark, download , extract spark in master node. , after copy hive-site.xml inside spark conf folder, start spark spark-shell command
. needed install in slave nodes also?
im asking because, im executing success spark sql queries below, if try acess cluster manager default page in localhost:8080, shows "unable connect". seems spark sql working fine, without cluster manager working, possible??
var hivecontext = new org.apache.spark.sql.hive.hivecontext(sc) query = hivecontext.sql("select * customers"); query.show()
master:8080
first , have let spark knows hadoop configurations are, setting env variable hadoop_conf_dir
in spark-env.sh file
then, when starting spark-shell
have tell spark use yarn master: spark-shell --master yarn-client
for more information see spark yarn docs
Comments
Post a Comment