Skip to content

Latest commit

 

History

History

hadoop-hive-1.2.x

README.md

Hadoop 2.7.x and Hive 1.2.x Example

The docker-compose.yml contains the following services:

  • namenode - Apache Hadoop NameNode
  • datanode - Apache Hadoop DataNode
  • resourcemanager - Apache Hadoop YARN Resource Manager
  • nodemanager - Apache Hadoop YARN Node Manager
  • historyserver - Apache Hadoop YARN Timeline Manager
  • hs2 - Apache Hive HiveServer2
  • metastore - Apache Hive Metastore
  • metastore-db - Postgres DB that supports the Apache Hive Metastore

Configuration

Hadoop configuration parameters are provided by the following .env files. Ultimately these values are written to the appropriate Hadoop XML configuration file. For Example, properties beginning with the following keys map the following files:

  • CORE_CONF_* > core-site.xml
  • HDFS_CONF_* > hdfs-site.xml
  • HIVE_SITE_CONF_* > hive-site.xml
  • YARN_CONF_* > yarn-site.xml

Key names use the following character conversions:

  • a single underscore _ equals dot .
  • a double underscore __ equals a single underscore _
  • a triple underscore ___ equals a dash -

For example, the key HDFS_CONF_dfs_namenode_datanode_registration_ip___hostname___check would result in the property dfs.namenode.datanode.registration.ip-hostname-check being written to hdfs-site.xml.

Another example, the key YARN_CONF_yarn_resourcemanager_resource__tracker_address would result in the property yarn.resourcemanager.resource_tracker.address being written to yarn-site.xml.

Exiting configuration files and their default values are listed below. Please note the value for YARN_CONF_yarn_nodemanager_resource_memory___mb assumes that your docker host has at least 8gb of memory. Feel free to modify as necessary.

core.env

HADOOP_LOG_DIR=/var/log/hadoop
YARN_LOG_DIR=/var/log/hadoop

CORE_CONF_fs_defaultFS=hdfs://namenode:8020
CORE_CONF_hadoop_http_staticuser_user=root

HDFS_CONF_dfs_namenode_datanode_registration_ip___hostname___check=false
HDFS_CONF_dfs_permissions_enabled=false
HDFS_CONF_dfs_webhdfs_enabled=true

YARN_CONF_yarn_nodemanager_resource_memory___mb=6144
YARN_CONF_yarn_resourcemanager_recovery_enabled=true
YARN_CONF_yarn_resourcemanager_store_class=org.apache.hadoop.yarn.server.resourcemanager.recovery.FileSystemRMStateStore
YARN_CONF_yarn_resourcemanager_system___metrics___publisher_enabled=true
YARN_CONF_yarn_timeline___service_enabled=true
YARN_CONF_yarn_timeline___service_generic___application___history_enabled=true

hive.env

HIVE_SITE_CONF_javax_jdo_option_ConnectionURL=jdbc:postgresql://metastore-db/metastore
HIVE_SITE_CONF_javax_jdo_option_ConnectionDriverName=org.postgresql.Driver
HIVE_SITE_CONF_javax_jdo_option_ConnectionUserName=hive
HIVE_SITE_CONF_javax_jdo_option_ConnectionPassword=hive
HIVE_SITE_CONF_hive_server2_transport_mode=binary
HIVE_SITE_CONF_hive_execution_engine=tez
HIVE_SITE_CONF_hive_metastore_uris=thrift://metastore:9083
HIVE_SITE_CONF_datanucleus_autoCreateSchema=false

yarn-node-manager.env

YARN_CONF_yarn_resourcemanager_resource___tracker_address=resourcemanager:8031

yarn-remote.env



yarn-resource-manager.env




yarn-timeline.env




Docker Compose


Start the Containers


docker-compose up


Stop and Destroy the Containers


docker-compose down


Testing


Once all services are up you can create a simple hive table to test functionality.  For example:


$ docker-compose exec hs2 bash
# /opt/hive/bin/beeline -u jdbc:hive2://localhost:10000
> CREATE TABLE pokes (foo INT, bar STRING);
> LOAD DATA LOCAL INPATH '/opt/hive/examples/files/kv1.txt' OVERWRITE INTO TABLE pokes;
> SELECT * FROM pokes;
> !quit


Exposed UI Interfaces



Docker Images



Open Interactive Shells


docker exec -ti namenode /bin/bash
docker exec -ti datanode /bin/bash
docker exec -ti resourcemanager /bin/bash
docker exec -ti nodemanager /bin/bash
docker exec -ti historyserver /bin/bash
docker exec -ti hs2 /bin/bash
docker exec -ti metastore /bin/bash
docker exec -ti metastore-db /bin/bash