Announcing availability of Spark 3.5 on Cloudera Unified Runtime

[#1]

We are excited to announce long-term support for Apache Spark 3.5 in the Cloudera Unified Runtime (7.3.1 SP1). This update brings the newest Spark version and a more modern, high-performance experience for your data engineering teams. 

Key Benefits

• Up to 50% faster long-running queries, with 5-10% overall performance gains from advanced query planning, adaptive execution, and dynamic partition pruning.
• Modern developer experience with wider Python support (PySpark) and broader SQL coverage.
• Stronger stability and scalability with improved memory management, fault tolerance, and support for large-scale workloads.

Next Steps

As of Cloudera Runtime 7.3.1.200 SP1, Spark 3 is the default version, and Spark 2 has been fully removed.

• If you already use Spark 3, upgrading is simple – just recompile your applications for Spark 3.5 and continue running your jobs. Migrating Spark Applications has useful resources to consider when moving between versions.
• Migrating from Spark 2? Start with these resources: Upgrading Spark clusters and Migrating Spark Applications.

Learn More 

• Visit the Cloudera Spark documentation page for on-premises and cloud
• Read the blog on Cloudera’s Unified Runtime
• Try the free 5-day trial for data engineering