spark1.6 configuring Sparksql metadata to be stored in PostgreSQL

1: Configure Hive-site.xml First

<Configuration>   < Property>       <name>Javax.jdo.option.ConnectionURL</name>       <value>Jdbc:postgresql://192.168.56.103:5432/sparksql</value>   </ Property>   < Property>       <name>Javax.jdo.option.ConnectionDriverName</name>       <value>Org.postgresql.Driver</value>   </ Property>   < Property>       <name>Javax.jdo.option.ConnectionUserName</name>       <value>Postgres</value>   </ Property>      < Property>       <name>Javax.jdo.option.ConnectionPassword</name>       <value>Gaoxing</value>   </ Property></Configuration>

2: Configure the JDBC jar path for PostgreSQL

Configuring in Spark-default.properties

Spark.driver.extraclasspath=/opt/spark/lib/postgresql-9.4.jar

Question 1

Start Thriftserver, 10000 listening port is not open. Rename the Hive-site.xml and use the default Derby database to start

Find answers on the Web:

The original hive automatically created PostgreSQL's representation, PostgreSQL will automatically lock the dead, your sister, so stupid.

You need to extract the PostgreSQL SQL statement from the source of hive to automatically create

Https://github.com/apache/hive/blob/master/metastore/scripts/upgrade/postgres/hive-schema-1.2.0.postgres.sql

Question 2

The name of the created table is uppercase, the dead can not query, asked the next PG DBA said is the problem of the pattern.

CREATE TABLE "CDS" (    "cd_id" bigint not NULL);

PG with quotation marks is case sensitive, elder brother, not according to the routine out of the card AH

spark1.6 configuring Sparksql metadata to be stored in PostgreSQL