[RELEASE] ScyllaDB Enterprise 2024.2.10

The ScyllaDB team announces ScyllaDB Enterprise 2024.2.10, a bug-fix production-ready ScyllaDB Enterprise patch release for ScyllaDB Enterprise 2024.2 Feature (Short Term Support) Release.

Note there is a later LTS (Long term Support) Release 2025.1, and you are encouraged to upgrade to it.

Moving an existing cluster to TLS without downtime

This release introduces a new Transitional state for node-node encryption.

The transitional state encrypts all outgoing traffic but allows non-encrypted incoming traffic, allowing an upgrade from non-encrypted to encrypted node-to-node traffic without downtime.

More here #18903

Related Links

The following issues are fixed in this release (with an open-source reference, if available):

Stability

Tablets

  • Fix the message when creating a new Tablet Keyspace #24006

Monitoring

  • node_ops_finished_percentage metric for removenode with RBNO never reaching 1, even though removenode finishes #21174

Tooling

  • scylla-nodetool: status: negative load size crashes nodetool #24134

Config

  • Fix description of commitlog_use_hard_size_limit #23780

Performance

  • Share IO queues between mountpoints seastar#2733 scylladb#23820
    Currently, Seastar assigns IO queues based on mountpoints listed in io-properties.yaml, mapping each mountpoint’s device number to an IO queue. However, this approach fails when a single physical disk is split into multiple virtual block devices with separate mountpoints, each gets its own IO queue, which is incorrect since they share the same underlying disk.

    This PR enhances the configuration format to allow a single disk entry to specify multiple mountpoints. Seastar will then create one IO queue for all specified mountpoints, correctly reflecting the shared physical disk.