Fixed issues in Cloudera Data Warehouse Private Cloud

This section lists issues fixed in this release of the Cloudera Data Warehouse (CDW) service.

DWX-16826: Executor pods do not get scheduled on the same node after the scale down

Earlier, due to the PersistentVolumeClaim (PVC) behavior, when the executor pods attempted to get rescheduled back on the node on which they were running before scaling down, but could not because other pods got scheduled on that node and consumed the resources. This issue has been fixed by removing the PVCs whenever the Virtual Warehouse scales down. The executor pods can still be scheduled on the same node if the node is healthy and suitable. But the executors no longer wait to get rescheduled on that particular node. They are automatically scheduled on any suitable node.

DWX-17179: Hue backend and Impala catalog pods are scheduled on the same node in an HA deployment

Earlier, multiple replicas of Hue backend, frontend, and Impala catalog pods got scheduled on the same node in HA mode and you had to manually schedule them on other nodes. This issue has been fixed.

DWX-16433: Logging Improvements

Writing production logs and workload profiles using Fluentd have been improved to write logs more reliably to HDFS under heavy load.

DWX-16891: Hive-Ranger integration issue after a refresh

Earlier, CDW displayed an error while compiling a statement: FAILED: HiveAccessControlException Permission denied. This happened because Hive failed to evaluate Ranger group policies when the Virtual Warehouse was either by upgrading or refreshing it, and you had to rebuild the Hive Virtual Warehouse to fix the Ranger integration issues. This issue has been resolved.