Known Issues in Cloudera Manager 7.4.4

Known issues in Cloudera Manager 7.4.4

OPSAPS-63881: When CDP Private Cloud Base is running on RHEL/CentOS/Oracle Linux 8.4, services fail to start because service directories under the /var/lib directory are created with 700 permission instead of 755.

Run the following command on all managed hosts to change the permissions to 755. Run the command for each directory under /var/lib:

chmod -R 755 [***path_to_service_dir***]

OPSAPS-61022: Role logging process is getting stuck during Cloudera Manager Agent upgrade

After you upgrade Cloudera Manager Agent, role logging process is stopped. Those roles that happen to log regularly will have their logging threads stuck, possibly causing the entire role to hang.

Restart the cluster after the Cloudera Manager Agent upgrade to restart the Role logging process.

OPSAPS-60943 Clicking on Historical Disk Usage causes error

On the Reports page in the Cloudera Manager Admin Console, clicking the link for Historical Disk Usage by (User or Group) causes the following error, caused by a Null Pointer Exception, to appear:

Server Error

A server error has occurred. The full stack trace is not shown here due to security reasons. See Cloudera Manager Server log for details.

Set the start and end times to a shorter time period by adjusting the following values in the URL:

&start=2021-11-09&end=2021-12-09

OPSAPS-61905 Spark jobs will not be able to run due to topology.py file that is not compatible with python3.

Manually edit topology.py file in the /etc/hadoop/conf* directories. Fix python3 incompatible lines.

Cloudera bug: OPSAPS-63881: When CDP Private Cloud Base is running on RHEL/CentOS/Oracle Linux 8.4, services fail to start because service directories under the /var/lib directory are created with 700 permission instead of 755.

Run the following command on all managed hosts to change the permissions to 755. Run the command for each directory under /var/lib:

chmod -R 755 [***path_to_service_dir***]
x

CDPD-26099 Extra steps required to run Hue on RHEL 8

See the following Hue administration topics:

• MySQL - Installing and configuring MySQL on RHEL 8
• MariaDB - Installing and configuring MariaDB on RHEL 8

CDPD-28390 Rolling restart of HDFS JournalNodes may timeout on Ubuntu 20

If the restart operation times out, you can manually stop and restart the NameNode and JournalNode services one by one.

CDPD-27663 RPC TLS configuration for Ozone is supported only on new CDP 7.1.7 clusters

Known Issue Description: gRPC TLS configuration for Ozone is supported only on new CDP 7.1.7 clusters and not on clusters upgraded from CDP 7.1.6 to CDP 7.1.7.

If you want to enable gRPC TLS on the upgraded CDP 7.1.7 clusters, you must contact Cloudera Support for more information.

OPSAPS-59163 Client Configuration downloaded from Cloudera Manager Admin Console is not the same as client configuration downloaded to cluster hosts

If you deploy the client configuration to a host that is not managed by Cloudera Manager, you will need to make the following change:

1. Unzip the client configuration bundle.
2. Edit the hadoop-env.sh file and change the line that begins with export HADOOP_MAPRED_HOME= to

   export HADOOP_MAPRED_HOME=<path to Jar files>

OPSAPS-59802 – Zeppelin and Livy roles should be co-located on the same host.

When installing or upgrading to CDP Private Cloud Base, you must co-locate all Zeppelin and Livy roles on the same cluster host due to an issue with certificate generation.

OPSAPS-60412 Extra step required when using Cloudera Manager Trial installer on Ubuntu 20

When using cloudera-manager-installer.bin to install a trial version of Cloudera Manager, the installation will fail.

Before running cloudera-manager-installer.bin run the following command:

apt-get install libncursesw5

OPSAPS-60721 Ozone fails to start when SCM High Availability is enabled

The Ozone SCM Primordial Node ID is a required field that needs to be specified with one of the SCM hostnames during Ozone HA installation.

Ozone SCM can either be deployed in a 3 node or 1 node configuration. For the 1 node configuration, the Ozone SCM Primordial Node ID configuration property is not a required configuration. However for 3 nodes, this field is a required configuration to be added during initialization of SCM.

When adding the Ozone service, enter the hostname of one of the SCM host in the Ozone SCM Primordial Node ID field on the Review Changes page. If this value is not set, the startup of Ozone services will fail.

OPSAPS-61045 Upgrading cluster with Hive requires Hue

If you have upgraded a cluster to CDP 7.1.7 and the source cluster includes the Hive service and the Hue service is not included, the Hive Metastore may fail after the upgrade. Perform the following steps to restore the Hive Metastore:

1. Ensure that you add a Hue user before proceeding to add the required Ranger policies for the Hue user.

2. Log in to the Cloudera Manager Admin Console.

3. See if Health check issue is observed.

4. Navigate to Clusters

5. Select the Ranger service

6. Click Web UI. This redirects to the Ranger service page.

7. On the Ranger Admin UI, click Service Manager

8. Add the required Ranger policies for Hue user

   The Policies include all - hiveservice, all - global, all - url, all - database, table, column and all - database,udf policies created under cl1_hive. For more information, see Ranger policies for components.

   After the Hue user is added to the policies metastore health check issue is no longer observed in the Cloudera Manager portal.

OPSAPS-61278: The SRM Client’s secure storage fails to generate correctly in FIPS-enabled clusters

In a FIPS enabled cluster, the SRM Client’s secure storage fails to generate correctly. As a result, the automatically generated configuration used by the srm-control tool will contain unresolvable references, making it unusable.

There are two workarounds for this issue. Choose one of the following:

• Configure SRM using the Streams Replication Manager’s Replication Configs Cloudera Manager property.

  Ensure that all cluster connection related properties (cluster aliases, bootstrap servers, security-related properties) are added, with the appropriate prefixes, for all clusters taking part in the replication process.

• Manually create a custom configuration file and use it with the srm-control tool.
  1. Create a copy of the default configuration located at /etc/streams_replication_manager/conf/srm.properties.
  2. Update all security related properties in the copy with the appropriate values.
     The values that you must update are similar to the following example:

     ${secure:/var/lib/streams_replication_manager/client.store:ext_ssl_truststore_password}

     Ensure that you replace all ${...} references with actual values such as the key and truststore locations, passwords, and so on.
  3. Run the srm-control tool using the --config option.
     For example:

     srm-control topics --config [***PATH TO CUSTOM CONFIGURATION FILE***] --source [***SOURCE_CLUSTER***] --target [***TARGET_CLUSTER***] --add [***TOPIC1***],[***TOPIC2***]

OPSAPS-61523 Failure when installing Cloudera Manager Agents

When installing the Cloudera Manager Agent package on new hosts through either the Add Hosts wizard or Add Cluster wizard, if you select Cloudera Repository during the Select Repository step, the Agent Installation fails with message "Failed to Copy Installation Files".

Select Custom Repository instead and enter the URL and your license credentials in the following format:

https://[username]:[password]@archive.cloudera.com/p/cm7/7.4.4