[RELEASE] ScyllaDB 2025.2.2

Release Notes
, ,
1

The ScyllaDB team announces ScyllaDB 2025.2.2, a bug-fix production-ready patch release for ScyllaDB 2025.2 Feature Release.

Related Links

The following issues are fixed in this release:

Stability

  • Remove throwing (some) protocol_exceptions when handling new connections #24567 #25272. avoid negative side effects of throwing these exceptions during a connection storm.

  • auth: default user creation may cause a crash if done during upgrade to raft topology #24975

  • Coredump during truncate: Data written after truncation time was incorrectly truncated #25013

  • Node crashes at shutdown with “Assertion this->_con->get()->sink_closed() failed” while restore in progress #25165

  • non-RBNO streaming may hang on failure if IO permits are deadlocked #24925

  • seastar_memory - oversized allocation (locator::describe_ring) #24158

  • Segmentation fault in scheduling group streaming during tablets streaming after node was decommissioned #23162

  • streaming,repair: db::view::check_needs_view_update_path() can deadlock with the repair/streaming permit #24807. This issue happens with removenode, when RBNO is disabled, so range

streamer is used.

  • test_mv_tablets_empty_ip: node got stuck during shutdown #23665

  • token_metadata_ptr may be destroyed without gentle cleaning #13381

  • Upgrade to raft topology may cause reads from system_distributed, leading to timeouts and defuncting group0 fiber #24963

Admin

Change Data Capture (CDC)

CDC is a feature that allows you to not only query the current state of a database’s table, but also query the history of all changes made to the table.

  • It’s possible to manually drop columns from CDC log #24643

  • node crashes when dropping a column while writing to it with CDC #24952

Tracing and logging

  • The system.clients virtual table now lists active Alternator requests in addition to CQL connections.

  • Change severity of Failed force_keyspace_cleanup - gate_closed_exception error to WARNING #16732

Monitoring

  • tasks: task::impl::get_progress may mix up different progress unit #24650

Performance

  • Destroying tablet_metadata holding many tables or even clearing it gently queues a long burst of seastar tasks #24814

  • service levels: cache is reloaded too often #25114 #23065

  • slow replace in the presence of a table with tablets #25163

  • utils::http::dns_connection_factory creates new system trust credential on each instance #24447

1 Like