Uploaded image for project: 'Spark'
  1. Spark
  2. SPARK-4758

Make metastore_db in-memory for HiveContext

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Resolved
    • Minor
    • Resolution: Won't Fix
    • 1.2.0, 1.3.0
    • None
    • SQL
    • None

    Description

      HiveContext by default will create a local folder metastore_db.

      This is not very user friendly as the metastore_db will be locked by HiveContext and thus will block multiple Spark process to start from the same directory.

      I would propose adding a default hive-site.xml in conf/ with the following content.

      <configuration>

      <property>
      <name>javax.jdo.option.ConnectionURL</name>
      <value>jdbc:derby:memory:databaseName=metastore_db;create=true</value>
      </property>

      <property>
      <name>javax.jdo.option.ConnectionDriverName</name>
      <value>org.apache.derby.jdbc.EmbeddedDriver</value>
      </property>

      <property>
      <name>hive.metastore.warehouse.dir</name>
      <value>file://${user.dir}/hive/warehouse</value>
      </property>

      </configuration>

      jdbc:derby:memory:databaseName=metastore_db;create=true Will make sure the embedded derby database is created in-memory.

      Jianshi

      Attachments

        Activity

          People

            Unassigned Unassigned
            huangjs Jianshi Huang
            Votes:
            2 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: