[RELEASE] ScyllaDB Manager 3.9.0

Release Notes
, ,
1

The ScyllaDB team announces the release of ScyllaDB Manager 3.9.0, a production-ready minor release of the stable 3.9 branch.

ScyllaDB Manager is a centralized cluster administration and recurrent tasks automation tool.

This release focuses on supporting new ScyllaDB 2026.1 features like native backup to Google Cloud Storage and native restore from both Google Cloud Storage and Amazon S3.

Below are the changes in this release.

Native backup support for Google Cloud Storage (#4588)

ScyllaDB Manager 3.6 introduced native backup support for Amazon S3, which allows uploading SSTables directly from ScyllaDB to the backup location without the ScyllaDB Manager Agent (rclone) proxy. ScyllaDB Manager 3.9 extends native backup support to Google Cloud Storage (GCS), available starting from ScyllaDB 2026.1. For more information, see native backup documentation.

Native restore support (issues)

ScyllaDB Manager 3.9 introduces native restore, which is a similar feature to the already released native backup. It allows ScyllaDB to stream SSTables directly from the backup location (S3 and GCS) into the cluster without downloading them to disk first via the rclone proxy. This feature is available starting from ScyllaDB 2026.1. For more information, see native restore documentation.

Native restore can be used by setting the new sctool restore --method flag. It specifies the API used for streaming SSTables:

  • ‘auto’: Use the supported native restore functionalities, otherwise use rclone restore.
  • ‘native’: Use the native restore functionalities only.
  • ‘rclone’: Use the rclone restore functionalities only.

Using native restore functionalities requires additional configuration and has its own limitations which are described in detail in the documentation.

Adjust default ScyllaDB Manager keyspace configuration (#4765)

ScyllaDB Manager 3.9 adjusts the default keyspace configuration used to store task definitions, progress, and other task related information. ScyllaDB Manager keyspace now uses NetworkTopologyStrategy with a single replica per rack.

ScyllaDB Manager requires access to its own ScyllaDB cluster for storing ScyllaDB Manager data. This usually is a local single node ScyllaDB instance, but it can also be any ScyllaDB cluster. When ScyllaDB Manager keyspace is created automatically on server startup, it will have a replication factor equal to the number of racks in each data center. In cases where this setup could lead to reduced availability (total replication factor smaller than 3), the replication factor will be increased up to 3 (depending on node count) regardless of the rack count. This ensures both rack-fault-tolerance and high availability out of the box.

For non-standard ScyllaDB Manager database deployments, especially multi-datacenter ones, it is still recommended to create the ScyllaDB Manager keyspace manually before starting ScyllaDB Manager server. For more information, see ScyllaDB Manager installation documentation.

Other improvements:

  • Support new format of S3 object storage endpoints (#4728)
  • Don’t batch SSTables during native restore (#4592)

Upgrade to the new release

ScyllaDB customers are encouraged to upgrade to ScyllaDB Manager 3.9.0 in coordination with the ScyllaDB support team.

The new release includes upgrades of both ScyllaDB Manager Server and Agent.

Useful Links:

ScyllaDB Manager 3.9.0 supports the following ScyllaDB releases

  • 2026.1
  • 2025.4
  • 2025.3
  • 2025.1
  • 2024.1

You can install and run ScyllaDB Manager on Kubernetes using ScyllaDB Operator. More here.