What's New in v20.2.11

June 14, 2021

• For a comprehensive summary of features in v20.2, see the v20.2 GA release notes.
• To upgrade to v20.2, see Upgrade to CockroachDB v20.2.

Get future release notes emailed to you:

Thanks!

Downloads

Docker image

$ docker pull cockroachdb/cockroach:v20.2.11

SQL language changes

• Floating point infinity values are now formatted as Infinity (or -Infinity if negative). This is for compatibility with PostgreSQL #65105
• INSERT INTO ... ON CONFLICT ... DO UPDATE SET statements without predicates now acquire locks using the FOR UPDATE locking mode during their initial row scan, which improves performance for contended workloads. This behavior is configurable using the enable_implicit_select_for_update session variable and the [sql.defaults.implicit_select_for_update.enabled cluster setting].(../v20.2/cluster-settings.html) #65364
• The ST_GeomFromGeoJSON(string) spatial function is now marked as the preferred overload, meaning it will resolve correctly in more contexts. #65441
• The cancellation behavior for DistSQL flows has been improved. #65046

DB Console changes

• Fixed a bug where empty series would show up in metrics graphs and legends, and when data was incorrectly attributed to the wrong nodes on graphs for clusters with decommissioned nodes. #66034
• Removed shading on line graphs, which improves legibility when viewing more than a few series on the same plot. #66034
• Drag to zoom on metrics graphs now supports time ranges under 10 minutes. #66034

Bug fixes

• BACKUP no longer resolves intents one by one. This eliminates the need to run a high-priority query to clean up intents to unblock BACKUP in case of intent buildup. #64931
• Fixed a bug that could cause backups to be slower. LockTableWaiter now checks the finalized transaction cache before pushing transactions to avoid duplicating work. #64999
• Fixed a bug where interval math on a TIMESTAMPTZ on a DST boundary would incorrectly add or subtract an extra hour. #65097
• Fixed a bug where date_trunc on a TIME value on a DST boundary could switch timezones and produce the incorrect result. #65097
• Fixed a bug causing the ZONECONFIG privilege on tables and databases to be incorrectly interpreted as USAGE, which could corrupt a table and/or database because USAGE is an invalid privilege for tables and databases. Also fixed the case when the ZONECONFIG privilege would be incorrectly restored as USAGE for tables and databases created and backed up in v20.1 and then restored through a full cluster RESTORE in v20.2 or later. #65159
• Fixed a bug which could cause a panic when running an EXECUTE of a previously prepared statement with a REGCLASS or REGTYPE parameter or a user-defined type argument after running BEGIN AS OF SYSTEM TIME with an invalid timestamp. #65150
• Fixed a bug which could cause a panic when issuing a query referencing a user-defined type as a placeholder. #65151
• Fixed a bug introduced in v20.2 that caused rows to be incorrectly de-duplicated from a scan with a non-unique index. #65288
• Fixed a bug causing revision_history cluster backups to not include dropped databases. This means that, previously, dropped databases could not be restored from backups that were taken after the database was dropped. #65317
• Fixed a bug where empty zone configurations get created for certain indexes during ALTER PRIMARY KEY. #65175
• Fixed a bug causing SHOW CREATE TABLE output to not display the zone configurations of a table or index if there were no partitions, even if there were zone configurations on the index or table. #65175
• Fixed a bug where ALTER DATABASE ... SET OWNER ... did not work if the database name was a keyword. #65367
• Fixed a bug where SHOW CREATE TABLE would show the zone configurations of a table with the same name from a different schema. #65369
• Previously, CockroachDB would crash when attempting to create a table using CREATE TABLE ... AS syntax where AS selects from crdb_internal.node_statement_statistics, crdb_internal.node_transaction_statistics, or crdb_internal.node_txn_stats virtual tables. #65544
• Fixed a bug where binary TIMETZ values were not being decoded correctly when sent as a parameter in the wire protocol. #65579
• A certain percentage of cases in which a node could have served a follower read were not handled correctly, resulting in the node routing the request to another nearby node for no reason. This has been fixed. #65470
• Fixed a race condition during transaction cleanup that could leave old transaction records behind until MVCC garbage collection. #65384
• Improved transaction cleanup for disconnected clients to reduce intent buildup. #65384
• Fixed a bug where the storage layer would under some rare write-heavy workloads start off by writing SSTable files that were too small for CockroachDB to manage effectively. The storage layer now creates fewer, larger files under those workloads, in line with expected behavior. #65481
• Scheduled backups with interleaved tables can now be created with the include_deprecated_interleaves option. #65730
• Calling get_bit or set_bit on a byte array argument now goes to the correct index of the underlying bitstring, which matches the Postgres behavior. #65787
• Previously, ALTER DATABASE ... CONVERT TO SCHEMA could potentially leave the schema with invalid privileges thus causing the privilege descriptor to be invalid. #65813
• Fixed a scenario in which a rapid sequence of splits could trigger a storm of Raft snapshots. This would be accompanied by log messages of the form would have dropped incoming MsgApp, but allowing due to ..., which tended to occur as part of RESTORE/IMPORT operations. #65500
• Previously, a schema's privilege descriptor could become corrupted upon executing ALTER DATABASE ... CONVERT TO SCHEMA. This is due to privileges that are invalid on a schema being copied over to the schema, rendering the schema unusable due to invalid privileges. #65999
• Previously, TRUNCATE transactionally scanned the meta range to find ranges to unsplit, creating contention with TRUNCATE's new behavior of preserving splits. #65942
• Fixed the error classification for duplicate index names where the later index was a UNIQUE index. #64002
• Fixed the error classification for ALTER TABLE ... ADD CONSTRAINT ... UNIQUE with the same name as an existing index. #64002
• Fixed a bug in SHOW RANGES that misattributed localities to nodes when using multiple stores. #66038
• Fixed a bug where incremental backups with revision_history would sometimes fail if a view was altered between incremental backups. #66072

Performance improvements

• Fixed an issue in the optimizer that prevented spatial predicates of the form (column && value) = true from being index-accelerated. These queries can now use a spatial index, if one is available. #65988

Contributors

This release includes 55 merged PRs by 32 authors. We would like to thank the following contributors from the CockroachDB community:

• Max Neverov
• Mohammad Aziz (first-time contributor)

Was this page helpful?

Yes No