Migrating Hive Workloads to Cloudera Private CloudPDF version

Identifying and fixing invalid Hive schema versions

As Administrator, after upgrading from Ambari-managed HDP to Cloudera Private Cloud Base, you need to identify Hive metastore operations that might fail due to Hive schema version incompatibility.

Incompatibility might exist if the upgrade process failed to make schema updates. You need to turn on the Hive Metastore Schema validation process for the metastore during the migration of your workloads to CDP. The Hive metastore captures any schema updates that occur during the upgrade, and displays issues in the Hive metastore logs. With this information, you can use the Apache Hive Schema tool to fix any problems.
  1. In Cloudera Manager, click Clusters > HIVE > Configuration.
  2. Check the hive.metastore.server.max.message.size.
  3. Set hive.metastore.server.max.message.size to the recommended value: 10% of the value of your Java heap size for Hive Metastore Server in bytes, but no more than 21478364. Recommended value: 214748364
  4. Click Clusters > HIVE > Configuration, and search for schema.
  5. Check Strict Hive Metastore Schema Validation to set hive.metastore.schema.verification to true.
  6. Check the Hive metastore logs and set a compatible metastore schema for the current Hive version using the Apache Hive Schema Tool.