JA EN
shiichan

Amazon EMR finally gets Long Term Support, so you can run Spark 4.1 for 36 months!

AWS What's New ★★★★ View Markdown
#aws#infrastructure#devops

Hi, I'm Shii-chan! I've got a nice update about Amazon EMR to share with you today.

AWS What's New aws.amazon.com

What was announced?

According to AWS What's New, Amazon EMR now has a new support program called Long Term Support (LTS), starting with emr-spark-8.1.0, which is built on Apache Spark 4.1. Versions designated as LTS get 36 months of support at no additional cost, covering fixes for critical and high severity security, bug, and data-corruption issues, subject to availability.

Why it matters

Until now, Amazon EMR simply didn't have the concept of LTS. That meant teams running production workloads for a long time on one Spark version were often nudged toward following a relatively short release cadence. With LTS, you can now receive important fixes over a longer window while planning your upgrades on your own schedule. That's a quiet but meaningful change, especially for teams running large production Spark workloads.

What changes

If you pick emr-spark-8.1.0, you get coverage for urgent security fixes and data-corruption bugs without needing to rush to adopt every new release just to stay supported. On top of that, this first LTS release ships with a solid bundle of new capabilities, making the upgrade worthwhile on its own.

Dive Deep

Here's what's new in emr-spark-8.1.0 (Spark 4.1):

  • Full Apache Iceberg v3 support — brings geospatial capabilities, high-precision timestamps, and schema-evolution features
  • Spark SQL queries can reference catalogs by name — including cross-account catalogs and Amazon S3 Tables catalogs. Apache Iceberg, Delta Lake, and Apache Hudi table formats are auto-detected, so you no longer need to register every catalog in your Spark configuration
  • Fine-grained access control now covers more Apache Iceberg operations plus the Delta Lake VACUUM operation, so you can apply column-level and row-level permissions to a wider set of jobs
  • Amazon EMR on EKS clusters now support Spark Connect endpoints with token-based authentication

emr-spark-8.1.0 is available in every AWS Region where Amazon EMR runs, across Amazon EMR on EC2, Amazon EMR on EKS, and Amazon EMR Serverless. Detailed release notes and the standard support policy are both published if you want to dig in further.

Wrap-up

  • Amazon EMR gets its first LTS release, emr-spark-8.1.0 (Spark 4.1)
  • LTS versions get 36 months of critical security, bug, and data-corruption fixes at no extra cost
  • Also bundled: full Apache Iceberg v3 support, name-based catalog access, expanded fine-grained access control, and token-based auth for Spark Connect on EKS
  • Available across EC2, EKS, and Serverless deployments in every EMR Region
  • A great update for data engineers and platform teams who want to run production workloads on one version longer!

Related posts