ScyllaDB Enterprise 2021.1.8

The ScyllaDB team announces ScyllaDB Enterprise 2021.1.8, a bug-fix production-ready ScyllaDB Enterprise patch release for ScyllaDB Enterprise 2021.1. ScyllaDB Enterprise customers are encouraged to upgrade to ScyllaDB Enterprise 2021.1.8 in coordination with the ScyllaDB support team.

ScyllaDB Enterprise 2021.1.8 brings focus on bug fixes, stabilization and tooling improvement. More below.

Related Links

• Get ScyllaDB Enterprise 2021.1.8 (customers only, or 30-day evaluation)
• Upgrade from 2020.1.x to 2021.1.y
• Upgrade from ScyllaDB Open Source 4.3 to ScyllaDB Enterprise 2021.1
• Submit a ticket

A list of stability improvements and bug fixes below, each with open source reference, is available:

• CQL: ScyllaDB now correctly rejects CQL insert/update statements with NULLs in key columns. #7852
• CQL: paxos_grace_seconds table option: prevent applying a negative value #7906
• CQL: Secondary Index-based queries yield less data than they should if base queries are short-reads #9198
• Install: scylla_setup failed to setup with error /etc/bashrc: line 99: TMOUT: readonly variable #8049
• Install: GCP: scylla_io_setup generates invalid io_properties.yaml on n2-highmem-48 instance with 24 SSDs #9429
• Install: scylla_io_setup on GKE: fails with workload identity enabled, since 4.3.x #9471
  Init: scylla can’t start when running in ARM based ami image #9493
• Install:  scylla_raid_setup: failed due to the “mdmonitor.service: Failed with result 'protocol'” on CentOS8  #9540
• Performance: stalls in repair caused by unnecessary copy #8902
• Performance (ARM): computing probabilistic read repair decisions when probability is 0 inefficient for ARM #9107
• Docker: stopwaitsecs missing from scylla-server supervisord config #9485
• Admin REST API: /storage_service/repair_status invalid status code when repair id not found #9576.
• UX: Uninformative error messages when client connects with wrong TLS version #8827,
• TLS handshake retry should on repeated failure retain original exception. Seastar #911
• Stability: The multi-shard combining reader is responsible for merging data from multiple cores when a range scan runs. A bug that is triggered by very small token ranges (e.g. 1 token) causes shards that have no data to contribute to be queried, increasing read amplification. #8161
• Stability: Set gossiper scheduling class explicitly #7154
• Stability: SSTable kl writer creates the index incorrectly, causing reads to miss range tombstones #9690
• UX: workload prioritization: Reduce the logging sensitivity
• Stability: Reshape stop API not working for TWCS tables #8755
• Stability: transport: Fix abort on certain configurations of native_transport_port(_ssl) #7866 #7783
• Stability: Failed fromJson() should result in FunctionFailure error, not an internal error #7911
• Stability: segfault due to corrupt _last_status_change map in cql_transport::cql_server::event_notifier::on_down during shutdown #8143
• Stability: possible Fix use-after-free in gossip real_mark_alive and mark_dead #8859
• Stability: a bug in CDC log forced a non optimized execution path (but did not hurt correctness) #9103
• ScyllaDB Manager uses this API. The issue is mitigated for earlier versions of ScyllaDB with Manager 2.5.4
• Stability: Bloom filter memory consumption can be increased with segregation on memtable flush #9581
• Stability: background reads out-living the parent reader can access freed range and slice #9719

14 Jan 2022