Fixed Issues in Cloudera Manager 6.3.0

The following sections describes issues fixed in Cloudera Manager 6.3.0:

NFS Gateway Startup failure

The NFS Gateway does not start when using Jsvc libraries with a version higher than 1.0.11. This is fixed for CDH versions 6.3 and higher.

Cloudera Bug: OPSAPS-50516

Kudu UNIX and Hive UNIX group

The host inspector in CM 6.3.0 now checks for the existence of the Kudu UNIX user and group. It also checks whether the Kudu user is a member of the Hive UNIX group. When upgrading to the CDH 6.3.0 parcel, the Kudu UNIX user will automatically be added to the Hive UNIX group. When upgrading via packages, this step must be performed manually; The Host Inspector will alert the user to this issue.

Cloudera Bug: OPSAPS-50360

Telemetry Publisher does not start with SSL enabled

Telemetry Publisher failed to start when SSL was enabled. This issue was fixed.

Cloudera Bug: OPSAPS-49256

Web server versions revealed in Cloudera Manager components

Cloudera Manager Admin Console and debug web UIs no longer send the "Server" header in responses.

Cloudera Bug: OPSAPS-24215

Failure to fetch Solr metrics when using non-default znode

Solr metric collection was unable to fetch data when a non-default ZooKeeper znode was set.

Cloudera Bug: OPSAPS-47239

Chronyd sync status not detected in some configurations

Fixed host monitoring issue for hosts using the "chrony" time synchronization daemon. Cloudera Manager now recognizes configurations with peer sync mode or local clock sync mode, in addition to server sync mode.

Cloudera Bug: OPSAPS-51143

Diagnostic bundle has zero-length agent log zip files

Fixed a bug causing zero-length agent log zip files in diagnostic bundles

Cloudera Bug: OPSAPS-49208

Null values in cluster templates fail silently

Fixed an issue where "null" values are passed in a Cloudera Manager cluster template. Cloudera Manager no longer silently fails during service installation.

Cloudera Bug: OPSAPS-49740

Renaming a cluster or service caused a server error

Renaming a cluster did not work correctly in Cloudera Manager 6.2.0 and is now fixed.

Cloudera Bug: OPSAPS-50167

Some services show up as stale when no configurations have changed

Fixed an issue where Compute clusters were indicating stale configurations in the Cloudera Manager Admin Console when no changes in the configuration have been made.

Cloudera Bug: OPSAPS-49698

Duplicate log file links in Network Performance Inspector

Fixed an issue that occurred when running "Inspect Network Performance" or "Inspect Cluster Network Performance" in the Cloudera Manager Admin Console. Links to log files were incorrectly repeated in the command output display.

Cloudera Bug: OPSAPS-49555

Telemetry Publisher does not decrypt the truststore password when testing the Altus connection

Fixed a bug where the Test Altus Connection command failed when SSL was enabled.

Cloudera Bug: OPSAPS-50248

Some default parameters are not set when using user-created Role Groups

Fixed an issue with the usage of user-created Role Groups used in Cluster Templates. Users do not have to provide the values for all configurations in the cluster template. Cloudera Manager, while creating cluster, will select (and auto-configure) values for missing configurations.

Cloudera Bug: OPSAPS-49893

Too many warnings about ephemeral ports

In previous Cloudera Manager versions, setting a role configuration (for example, HDFS DataNode HTTP Web UI Port) to an ephemeral port (e.g., 50010) would generate as many warnings as the number of DataNodes, and there was no option to suppress the warning. Now, only one warning is generated and the Suppress button is available in the Cloudera Manager Admin Console.

Cloudera Bug: OPSAPS-50314

Error handling in case of multithreaded Hive replication

Multi-threaded Hive replication introduced in CM 5.15 generated errors that could not be suppressed using the replication.hive.ignoreTableNotFound and replication.hive.igenoreDataBaseNotFound properties in Advanced Configuration Snippets. This is now fixed so that the new errors can be suppressed.

Cloudera Bug: OPSAPS-49987

Custom replication log paths now use automatic log retention settings

Previously, custom replication log paths set by 'Log Path' and 'Directory for metadata file' replication arguments did not utilize automatic log retention settings. This is fixed.

Cloudera Bug: OPSAPS-50878

Cloudera Manager 6.1 does not substitute {{CONF_DIR}}/library.leveldbjni.path for YARN processes

After upgrading to Cloudera Manager 6.1 or higher, if the /tmp directory is mounted using the noexec flag, then starting or restarting YARN NodeManagers will fail for all CDH versions. Note that this change may cause configuration staleness on Cloudera Manager when upgrading to 6.3.0 for clusters running YARN.

Cloudera Bug: OPSAPS-50253

Key Trustee Clusters

The button to enable Key Trustee Clusters no longer appears on the Administration -> Security -> Status page. This configuration is not required.

Cloudera Bug: OPSAPS-45389