Fixed Issues in Cloudera Manager 6.2.0
The following sections describes issue fixed in Cloudera Manager 6.2.0:
- Default thresholds for swap warnings are too low
- Service name auto-generation code is broken, causing a constraint violation
- Hive replication does not copy an empty database
- Cluster template import causes state corruption
- Summary and full views in the Replication API display the same result
- Setting displayName in cluster template does not impact display name
- Server fails to contact agent when TLS is on
- Host inspector incorrectly reports bad version of PSYCOPG2
- Diagnostic bundle has zero-length agent log .zip files
- Import/Export cluster deployment through API for inbuilt/custom roles
- Cannot negotiate TLS to Cloudera Manager Admin Console on OpenJDK 1.8 on Centos 7.2
- YARN jobs failure when using encrypted shuffle
- Uploading a valid license via the Cloudera Manager API Swagger client returns 500 Server Error
- Syntax highlighting for Java API client docs is broken
- Disable Hive > Actions > Update Hive Metastore NameNodes if high availability is not enabled for HDFS
- Cloudera Navigator Limitation when using Virtual Private Clusters
- Upgrade CDH screen does not list all the databases that need to be backed up
- dfs.client.read.shortcircuit needs to be set correctly for each Impala instance
- Maven dependency issue for jaxrs
- Java runtime error when setting static pool through Cloudera Manager
- Snapshot Policies screen hangs
- Make sure Spark app log collection works with authentication on
- New compile lock related configuration values to Cloudera Manager
- Add pre-upgrade confirmation box to HBase upgrade
- Diagnostic bundle creation slows down Cloudera Managers
- Broken Link on Replication History page
Default thresholds for swap warnings are too low
The default process swap size warning threshold has been increased from 0 to 200Mb. A new host level configuration property, Default Process Swap Memory Thresholds, has been added for process swap size alert thresholds. This allows for bulk updates of process level alerts that use the default settings.
Cloudera Bug: OPSAPS-44904
Service name auto-generation code is broken, causing a constraint violation
Fixed an issue where Cloudera Manager incorrectly auto-generated the Service Display Name when adding services.
Cloudera Bug: OPSAPS-48672
Hive replication does not copy an empty database
Empty databases without metadata or Hive/Impala UDFs are now replicated during Hive replication.
Cloudera Bug: OPSAPS-47224
Cluster template import causes state corruption
Fixed an issue with importing Cluster Templates. When hosts that are used in another cluster are, by mistake, specified in the template, then these hosts are dissociated with their cluster and the cluster is left in an invalid state and beyond repair.
Cloudera Bug: OPSAPS-48680
Summary and full views in the Replication API display the same result
Fixed an issue where using the Cloudera Manager API to return a summary view for a replication job returns a summary with limited information instead of the full view.
Cloudera Bug: OPSAPS-47182
Setting displayName in cluster template does not impact display name
Fixed an issue where setting the display name for a service using a cluster template did not change the service name.
Cloudera Bug: OPSAPS-46056
Server fails to contact agent when TLS is on
Fixed an issue with TLS encryption. When TLS encryption is enabled for agent-server communication, but no key and certificate is configured for an agent host, then the agent might erroneously report that it is serving the agent status server (port 9000) over HTTPS when it is not. This caused failures during diagnostic bundle collection.
To configure the agent status server to use HTTPS, configure the agent with a client_cert_file and client_key_file in the /etc/cloudera-scm-agent/config.ini file.
Cloudera Bug: OPSAPS-48958
Host inspector incorrectly reports bad version of PSYCOPG2
Fixed an issue where the Host Inspector may incorrectly report that an incompatible version of PSYCOPG2 was in use.
Cloudera Bug: OPSAPS-48649
Diagnostic bundle has zero-length agent log .zip files
Fixed an issue where zero-length agent log .zip files were included in diagnostic bundles.
Cloudera Bug: OPSAPS-49208
Import/Export cluster deployment through API for inbuilt/custom roles
Fixed an issue where using the Cloudera Manager API to import and export deployments failed when using custom roles.
Cloudera Bug: OPSAPS-48104
Cannot negotiate TLS to Cloudera Manager Admin Console on OpenJDK 1.8 on Centos 7.2
Fixed an issue where older versions of OpenJDK may not implement certain TLS ciphers, causing an inability to log into the Cloudera Manager Admin Console when TLS is enabled. Either upgrade the version of OpenJDK, or allow Cloudera Manager to use less secure ciphers by editing /etc/default/cloudera-scm-server and uncommenting the line that contains "export CMF_OVERRIDE_TLS_CIPHERS=".
Cloudera Bug: OPSAPS-49578
YARN jobs failure when using encrypted shuffle
On some operating systems, the default file mask may cause the Auto-TLS truststore file to not be world-readable, causing YARN jobs to fail when encrypted shuffle is enabled. This issue has been fixed.
Cloudera Bug: OPSAPS-48731
Uploading a valid license via the Cloudera Manager API Swagger client returns 500 Server Error
Fixed the update_license() method in ClouderaManagerResourceApi in the Cloudera Manager Python Swagger-based API client that allows updating the license file for Cloudera Manager. Without the fix, the update_license() method fails with the following error: "No multipart with content id license found" HTTP error 500.
Cloudera Bug: OPSAPS-49116
Syntax highlighting for Java API client docs is broken
Fixed an issue with syntax highlighting for the Cloudera Manager Java API client SDK documentation.
Cloudera Bug: OPSAPS-49406
Disable Hive > Actions > Update Hive Metastore NameNodes if high availability is not enabled for HDFS
Fixed an issue where re-running the Update Hive Metastore NameNodes Command when HDFS High Availability was disabled corrupted the URI paths in Hive tables by appending an additional port number.
Cloudera Bug: OPSAPS-46970
Cloudera Navigator Limitation when using Virtual Private Clusters
Navigator does not capture auditing and lineage information from Compute clusters in version 6.2.0.
Cloudera Bug: OPSAPS-48694
Upgrade CDH screen does not list all the databases that need to be backed up
When you run the Upgrade Cluster wizard to upgrade from CDH 6.1 to 6.2, Hive now appears in the list of databases to be backed up.
Cloudera Bug: OPSAPS-49314
dfs.client.read.shortcircuit needs to be set correctly for each Impala instance
When Impala daemons are not co-located with a DataNode, Advanced Configuration Snippets are no longer necessary to disable short circuit reads when using CDH 6.2 or higher.
Cloudera Bug: OPSAPS-46971
Maven dependency issue for jaxrs
Fixed an issue where some of the Cloudera Manager API endpoints stopped working in version 6.1.0.
Cloudera Bug: OPSAPS-49159
Java runtime error when setting static pool through Cloudera Manager
Fixed a RuntimeException that occurs when configuring Static Pools in Cloudera Manager.
Cloudera Bug: OPSAPS-48476
Snapshot Policies screen hangs
Fixed an issue where the Snapshots Policies screen in the Cloudera Manager Admin Console hangs when there are scheduled replications with snapshots.
Cloudera Bug: OPSAPS-49511
Make sure Spark app log collection works with authentication on
Fixed several issues with the YARN diagnostic bundle, which could not collect Spark event logs when SSL or Kerberos authentication was enabled for the Spark History Server.
Cloudera Bug: OPSAPS-39280
New compile lock related configuration values to Cloudera Manager
- hive.driver.parallel.compilation
- hive.driver.parallel.compilation.global.limit
Cloudera Bug: OPSAPS-47503
Add pre-upgrade confirmation box to HBase upgrade
Added two checkboxes to the HBase upgrade screen for validating co-processors and tables.
Cloudera Bug: OPSAPS-48815
Diagnostic bundle creation slows down Cloudera Managers
Fixed an issue where running a diagnostic bundle command slows down other commands running in Cloudera Manager.
Cloudera Bug: OPSAPS-49233
Broken Link on Replication History page
Fixed a broken link on the Replication History page to the respective YARN jobs.
Cloudera Bug: OPSAPS-29486