What's New in v20.1.10

December 21, 2020

• For a comprehensive summary of features in v20.1, see the v20.1 GA release notes.
• To upgrade to v20.1, see Upgrade to CockroachDB v20.1.

Get future release notes emailed to you:

Thanks!

Downloads

Docker image

$ docker pull cockroachdb/cockroach:v20.1.10

Backward-incompatible changes

• The file names for heap profile dumps are now memprof.<date-and-time>.<heapsize>, where previously they were named memprof.<heapsize>.<date-and-time>. #55260

SQL language changes

• Fixed a bug where temporary tables may be included in BACKUP commands. #56475

Command-line changes

• CockroachDB now better attempts to retain heap profile dumps after a crash due to an out-of-memory error. #55260
• CockroachDB now better attempts to retain memory statistics corresponding to increases in total memory usage, not just heap allocations. #55260

Bug fixes

• Fixed a panic that could occur when running SHOW STATISTICS USING JSON for a table in which at least one of the columns contained all null values. #56515
• The file names for generated goroutine, CPU and memory profiles were sometimes incorrect, resulting in repeated warnings like strconv.ParseUint: parsing "txt": invalid syntax in log files. This has been corrected. #55260
• Fixed a bug when the Pebble storage engine is used with encryption-at-rest that could result in data corruption in some fairly rare cases after a table drop, table truncate, or replica deletion. #56680
• Previously, dumps of tables with a BIT type column would result in an error. This column type is now supported. #56452
• In 20.1.8, we attempted to fix age's lack of normalization of H:M:S into the years, months, and days field. However, this was also broken for values greater than 1 month, as well as breaking a::timestamp(tz) - b::timestamp(tz) operators. This has now been resolved. #56769
• CockroachDB previously would crash when executing a query with an AS OF SYSTEM TIME clause that used a placeholder (note that it wasn't a prepared statement, it was an attempt to use an unspecified placeholder value on a non-prepared statement). This is now fixed. #56781
• CockroachDB previously could encounter an internal error when DATE, TIMESTAMP, or TIMESTAMPTZ values that used year 1 BC were sent between nodes for execution. Additionally, previously it was not possible to specify DATE, TIMESTAMP, or TIMESTAMPTZ values with year 1 BC without using AD/BC notation. This is now fixed. #56743
• Fixed internal error when collecting a statement diagnostic bundle in some cases where the query hits an error. #56785
• Some boolean session variables would only accept string ("true" or "false") values. Now they also accept unquoted true or false values. #56814
• Fixed a bug which would prevent the dropping of hash sharded indexes if they were added prior to other columns. #55823
• Fixed a race condition in the tpcc workload with the --scatter flag where tables could be scattered multiple times or not at all. #56979
• Previously if a cluster backup that was taken during a schema change, a cluster restore of that backup would create duplicates of the ongoing schema changes. #56450
• Fixed a case where attempting to start a second BACKUP to the same location while the first was running using passphrase-based encryption could overwrite the metadata required to decrypt it and thus render it unreadable. #57025
• Fixed an internal error when using aggregates and window functions in an ORDER BY for a UNION or VALUES clause. #57522
• The CREATE TEMP TABLE AS statement previously created a non-temporary table. Now it makes a temporary one. #57550
• Fixed a bug where schema change jobs to add foreign keys to existing tables, via ALTER TABLE, could sometimes not be successfully reverted (either due to being canceled or having failed). #57810
• Fixed a bug where concurrent addition of a foreign key constraint and drop of a unique index could cause the foreign key constraint to be added with no unique constraint on the referenced columns. #57810
• Fixed a bug which can occur when canceling schema changes when there are multiple queued schema changes that could result in future schema changes being stuck. #55058
• Fixed a bug which can lead to canceled schema change jobs ending in the failed rather than canceled state. #55058

Performance improvements

• Interactions between Raft heartbeats and the Raft goroutine pool scheduler are now more efficient and avoid excessive mutex contention. This was observed to prevent instability on large machines (32+ vCPU) in clusters with many ranges (50k+ per node). #57009

Backward-compatible change

• The reserved, non-documented cluster settings server.heap_profile.xxx have been renamed to server.mem_profile.xxx. They now control collection of multiple sorts of memory profiles besides just Go heap allocations. #55260

Contributors

This release includes 34 merged PRs by 16 authors.

Was this helpful?