Known issues in 7.1.9 SP1 CHF 4

You must be aware of the known issues and limitations, the areas of impact, and workaround in Cloudera Runtime 7.1.9 SP1 CHF 4.

The list of known issues for the Runtime release 7.1.9 SP1 CHF 4 includes the known issues from Runtime release 7.1.9 SP1. For more information, see Known issues in Cloudera Runtime 7.1.9 SP1.

CDPD-70450: Impala SQL queries that include the “WITH” clause should populate lineage in Atlas

Impala SQL queries that do not use the WITH clause can show lineage in Atlas, but queries that do use the WITH clause cannot show lineage in Apache Atlas. Impala SQL queries using the WITH clause are not supported.

CDPD-75665: Importing a table generates a DDL with an incorrect location

When creating a table using the IMPORT command, this table's partitions could point to an incorrect location, that is, an external imported table can have his partitions located under the managed warehouse directory, which violates the metastore.warehouse.external.dir and metastore.warehouse.dir that intend to host different type of tables.

Apache Jira: HIVE-28580

None

CDPD-77738: Atlas hook authorization issue causing HiveCreateSysDb timeout

Atlas hook authorization error causes HiveCreateSysDb command to time out due to repeated retries.

None

COMPX-18909:YARN NodeManagers potentially getting into unhealthy state caused by YARN-11709

A logic was introduced to mark the NodeManager unhealthy if the container-executor encounters an unrecoverable or configuration error. For example, if the container-executor script is missing from the configured place. There is an edge-case in this behaviour that, if the application is killed just before one of the containers is trying to access the localizer syslog file, an IOException occurs because the NodeManager has already deleted the application log directory. This IOException is then catched by the logic of CDPD-49702/YARN-11709, and the NodeManager is marked unhealthy incorrectly.

Restart the unhealthy NodeManager.

OPSAPS-72509, CDPD-32440: Hive metadata transfer to GCS fails with ClassNotFoundException

Hive replication policies from an on-premises cluster to cloud fails during the “Transfer Metadata Files” step if the following conditions are true:

• the target is a GCS Data Lake
• the source Cloudera Manager version is 7.11.3 CHF7, 7.11.3 CHF8, 7.11.3 CHF9, 7.11.3 CHF9.1, 7.11.3 CHF10, or 7.11.3 CHF11

This is because the fs.gs.delegation.token.binding property is already defined in the configuration and cannot be unset to disable the delegation tokens in the cloud connector service.

None