Upgrading to Cloudera Runtime 7.3.2

Cloudera Runtime 7.3.2 introduces significant new features and improvements. As a result, if you are planning to upgrade the Cloudera Runtime version in your existing Data Lake or Cloudera Data Hub clusters to 7.3.2, you might be required to perform this in multiple steps.

Cloudera Runtime 7.3.2.10000 Service Pack 1 (SP1)

Some of the important Cloudera Runtime 7.3.2.10000 SP1 features and changes include:

Cloudera Streams Messaging cluster:
- Migrating existing ZooKeeper-based Kafka clusters to KRaft is now supported. For more information, see the Migrating Kafka from ZooKeeper to KRaft documentation.
- Scaling KRaft nodes is generally available on public cloud Streams Messaging clusters. For more information, see Scaling KRaft.
- Streams Messaging components can use centralized Cloudera Manager TLS protocol and cipher settings, including TLS 1.3, when the TLS_ADVANCED_CONTROL feature flag is enabled.
Default Python version is 3.12

For the service specific list of new features, known issues, fixed issues and improvements in Cloudera Runtime 7.3.2.10000 SP1, see the Cloudera Runtime Release Notes.

Cloudera Runtime 7.3.2

Some of the important Cloudera Runtime 7.3.2 features include:

Default Operating System (OS) version is RHEL 9.6
Default Java Development Kit (JDK) version is 17
Cloudera Streams Messaging cluster:
- Kafka Raft (KRaft) is generally available.
- Kafka 3.9 rebase
Cloudera Flow Management clusters:
- Removing NiFi 1 from the cluster templates

For the service specific list of new features, known issues, fixed issues and improvements in Cloudera Runtime 7.3.2 release, see the Cloudera Runtime Release Notes.

Cumulative Hotfix: Cloudera Runtime 7.3.2.100 Cumulative Hotfix 1 (CHF1) includes a set of fixed issues and improvements. For more information, see the Cloudera Runtime Release Notes.

important

Cloudera Runtime 7.3.2 requires JDK 17. After you upgrade to 7.3.2, rolling back to an earlier Java version is not possible. If your cluster includes custom Java solutions, Cloudera recommends performing a phased JDK upgrade on Cloudera Runtime 7.3.1 before upgrading to 7.3.2. This allows you to validate your custom solutions against JDK 17 and roll back to Java 8 on 7.3.1 if issues arise. For more information, see Phased JDK upgrade on 7.3.1 before upgrading to 7.3.2.
On JDK 17, interactive spark-shell (Scala REPL) requires the driver JVM option -Dscala.usejavacp=true in spark.driver.extraJavaOptions. Without it, spark-shell can fail with compiler-mirror errors while spark-submit and pyspark continue to work. Configure this option in Cloudera Manager before you test Spark 3 after upgrade. For steps, see spark-shell on JDK 17 in Running Apache Spark 3.5.4 applications.
Upgrades are not supported for Cloudera Streams Messaging clusters that use the technical preview version of KRaft. Upgrading a Streams Messaging cluster running in KRaft mode from Cloudera Runtime 7.3.1 or earlier is not supported. You must decommission the existing Cloudera Streams Messaging cluster and deploy a new one on the target version.
Upgrading a Cloudera Flow Management Data Hub cluster that uses NiFi1 to Cloudera Runtime 7.3.2 is not supported. Ensure that you first migrate to NiFi 2 to be able to upgrade to Cloudera Runtime 7.3.2. For more information, see the Using the Flow Management Migration Tool documentation.

What to do next?

The following pages provide step-by-step guidance on identifying the upgrade path for your source cluster version and upgrading to Cloudera Runtime 7.3.2.