Fixed Issues in Cloudera Manager 7.12.0

Fixed issues in Cloudera Manager 7.12.0.

OPSAPS-69709: Set Sqoop Atlas hook to send notifications synchronously

Sqoop has an Atlas hook which by default runs asynchronously to send notifications to the Atlas server. In certain cases, the Java Virtual Machine (JVM) in which Sqoop is running can shut down before the Kafka notification of the Atlas hook is sent. This can result in lost notifications.

This issue is fixed by ensuring that the notifications are synchronous.

OPSAPS-69759: Multiple TestDFSIO(Mapreduce job) failure during COD ZDU

This issue has been fixed and Mapreduce job failures will no longer occur.

OPSAPS-69340: Dlog4j.configurationFile annotation is not working with the log4j library of the Cloudera Manager Server

The incorrect notation used in defining the log4j configuration file name (which is Dlog4j.configurationFile annotation) is preventing the Cloudera Manager Server from receiving updates made to thelog4j.properties file. This issue is fixed now.

OPSAPS-69485: Invalid mapred-site.xml due to double dash in comments

The string '--' is not allowed in XML comments. Cloudera Manager incorporates values from the safety valve into XML comments. Therefore, XML configuration file generation fails if the safety valve contains '--'.

Cloudera Manager replaces the "--" characters in XML configuration file comments with &#8212 which is the Unicode character of '--'.

OPSAPS-66908: Refresh Cluster command degrades with high node count

The refresh cluster command performance is improved now with asynchronous config generation.

OPSAPS-64516: Unable to clear user's local cache files for YARN in Cloudera Manager

A new YARN command, CleanNmLocalDirCommand, was created to delete the cache files. This command clears, but does not delete the directories under the YARN local directories. This command can only be used if the YARN service is stopped. Users can also now clear the local cache through Cloudera Manager.

OPSAPS-67041: Telemetry Publisher throws FileNotFoundException for Spark application for ifile log

This fix addresses an issue where a modification in the YARN log path caused an inability to access executor logs for Spark in client and cluster modes, and driver logs for Spark in cluster mode through the observability user interface when submitting a Spark job. A similar scenario was observed for Hive as well.

OPSAPS-68500: The cloudera-manager-installer.bin fails to reach Ubuntu 20 repository on the Archive URL due to redirections

Agent Installation with Cloudera Manager on Ubuntu20 platform does not function when the self-installer method (using the installer.bin file) is employed to install Cloudera Manager. The failure mode is that Cloudera Manager Agent installation step will fail with an error message saying "The repository 'https://archive.cloudera.com/p/cm7/7.11.3/ubuntu2004/apt focal-cm7 InRelease' is not signed."

This issue is fixed now.

OPSAPS-66023: Error message about an unsupported ciphersuite while upgrading or installing cluster with the latest FIPS compliance

When upgrading or installing a FIPS enabled cluster, Cloudera Manager is unable to download the new CDP parcel from the Cloudera parcel archive.

Cloudera Manager displays the following error message:

HTTP ERROR 400 java.net.ConnectException: Unsupported ciphersuite TLS_EDH_RSA_WITH_3DES_EDE_CBC_SHA

This issue is fixed now by correcting the incorrect ciphersuite selection.

OPSAPS-67641: Hive ACID replication UX improvement

The Next Run column on the Cloudera Manager > Replication > Replication Policies page was showing None Scheduled for recurring Hive ACID replication policy jobs, which is incorrect. The column now displays the correct message.

OPSAPS-68524: Updating OzoneReplicationType in UI

The Listing type option now displays all the available options where you can choose a replication method to replicate Ozone data using Ozone replication policies.

OPSAPS-69063: Concurrent policy creation to multiple targets

Sometimes, standard error or standard output retrieval of Cloudera Manager commands would fail because of a Java-related issue which affected the HTTPS connections with TLSv1.3 protocol. This resulted in different failures when the HBase replication commands were run remotely from the destination cluster on the source cluster. This issue is now resolved.

OPSAPS-68995: Convert some DistCp feature checks from CM version checks to feature flags

To ensure interoperability between different cumulative hotfixes (CHF), the NUM_FETCH_THREADS, DELETE_LATEST_SOURCE_SNAPSHOT_ON_JOB_FAILURE, and RAISE_SNAPSHOT_DIFF_FAILURES DistCp features must be published as feature flags.

OPSAPS-69481: Some Kafka connect metrics missing from Cloudera Manager due to conflicting definitions

Cloudera Manager now registers kafka_connect_connector_task_metrics_batch_size_avg and kafka_connect_connector_task_metrics_batch_size_max metrics correctly.