Fixed issues in Cloudera Runtime 7.1.9 SP1 CHF 17

Know more about the list of fixes that are shipped for CDP Private Cloud Base version 7.1.9 SP1 CHF 17.

CDPD-99837: Mixed case key names can break HDFS encryption after KTS to Ranger KMS migration

When you migrate encryption keys from Key Trustee Server (KTS) to the Ranger KMS DB, encryption zone key names that were originally mixed case (uppercase, lowercase, or camelCase) in KTS can cause HDFS decryption to fail after migration.

The issue has been fixed.

CDPD-102110: Query failure when cost-based optimization is disabled

Previously, a query failure occurred with a runtime error when you ran a join query containing string functions with cost-based optimization disabled.

This issue is now resolved.

Apache Jira: HIVE-29488

CDPD-91439: Support for python-oracledb connectivity in Hue

Previously, Hue experienced database connectivity incompatibilities due to limitations with the deprecated cx_Oracle driver. This issue is now fixed by transitioning Hue to use the new python-oracledb library. This change enables official support for modern Python-OracleDB connectivity, ensures compatibility with newer versions like Python 3.11, and secures the Oracle database integration within Hue.

CDPD-101682: DirectoryDeletingService is using single thread

Previously, the DirectoryDeletingService in Ozone used an unbounded queue, which resulted in directory deletion tasks being processed sequentially by a single thread, even when multiple threads were available. This could lead to slow deletion performance for large datasets.

This issue is now resolved by ensuring that directory deletion properly supports multi-threaded operations. The number of threads can now be configured to optimize performance.

Apache Jira: HDDS-15080

CDPD-101434: Possibility of Missing blocks in majority of Datanodes for Ratis Pipeline

Previously, when data was written through a Ratis pipeline, blocks were occasionally written only to the leader DataNode, despite the system returning a successful write acknowledgment to the client. Consequently, if the leader DataNode failed or went offline before the data was replicated to other nodes, the associated blocks or chunk data went missing, resulting in temporary data unavailability. This vulnerable state persisted until the Storage Container Manager (SCM) background processes detected the under-replicated container and completed the proper replication.

This issue is fixed now.

Apache Jira: HDDS-15052

CDPD-72842: Ozone DataNode reporting negative container values for UsedBytes and BlockCount parameters

Previously, the duplicate delete requests in the block-deletion backlog caused DataNodes to report incorrect or negative container metrics.

This issue is now resolved and the background delete task now skips repeat deletes within the same run to ensure consistent space and block accounting.

Apache Jira: HDDS-11267

CDPD-97209: Inconsistency in handling REPLICATION_NOT_HEALTHY_BEFORE_MOVE in Container Balancer

Previously, ContainerBalancerSelectionCriteria used a less stringent check that allowed unhealthy containers to proceed, only for them to be rejected later by MoveManager’s stricter health requirements. By adopting the same rigorous health validation upfront, only healthy containers are selected for movement.

This issue is fixed by returning the source Datanodes to the priority queue upon failure so they remain eligible for subsequent balancing attempts.

Apache Jira: HDDS-14614