Next platform deadline: Amazon EKS Nov 26, 2026  

All 14 deadlines

Data & Databases

Cassandra 3.x is end of life. The upgrade path to 4.1 and 5.0

Cassandra 3.0 and 3.11 have been unmaintained since September 2024, and 4.0 is next in line. How the project supports versions, why you cannot skip a major, and how to plan a rolling upgrade.

By the SRASH team · · 5 min read

Key facts

  • Cassandra 3.0 and 3.11 reached end of life when 5.0 became generally available on Sep 5, 2024.
  • The project maintains about three releases: today 5.0, 4.1 and 4.0. 4.0 is maintained until the 6.0 release.
  • Online, rolling upgrades are supported between adjacent major versions; skipping a major is not supported.
  • CVE fixes may still be applied to unmaintained versions case by case, but there is no commitment.

Where support stands

The Apache Cassandra project announced in 2023 that 3.0 and 3.11 would reach end of life when 5.0 went GA. That happened on September 5, 2024. The project's download page now lists 5.0 as the latest release, 4.1 as the previous stable release (maintained until 7.0) and 4.0 as the older stable release (maintained until 6.0).

The project's release policy aims to keep three releases supported at any time, and its wiki states that support for 4.0 drops when 6.0 ships. If you are on 4.0 today, you have time, but not unlimited time.

Why you cannot jump straight to 5.0

Cassandra keeps compatibility between adjacent major versions so that clusters can be upgraded node by node while serving traffic. Upgrades that skip a major version are not supported. From 3.11 the supported route is to 4.0 or 4.1 first, then to 5.0.

A practical rolling upgrade

  1. Get healthy first. Run a full repair, clear pending compactions and confirm schema agreement across the ring.
  2. Check clients. Upgrade drivers to versions that support both the old and new protocol before touching servers.
  3. Review configuration changes. 4.x renamed and removed settings; 5.0 adds new features such as storage-attached indexes and trie-based memtables and SSTables that you can adopt later.
  4. Snapshot each node, then upgrade one node at a time: drain, stop, upgrade, start, verify.
  5. Upgrade SSTables once all nodes run the new version, then repeat for the next major.
  6. Rehearse the whole sequence on a staging ring with production-like data and traffic.

Managed alternatives

Some teams use the upgrade as a moment to move to a managed service such as DataStax Astra DB, Amazon Keyspaces, Azure Managed Instance for Apache Cassandra or a managed Cassandra provider. That removes upgrade work but changes cost, features and control, so compare them on your workload before deciding.

Start with a free 30-minute review.

Databases, cloud or AI: talk to a senior engineer about what you run and what is coming. We reply within one business day and sign your NDA on request.