Version 5.10.2

Released on 2025-02-28.

Note

If you are upgrading a cluster, you must be running CrateDB 4.0.2 or higher before you upgrade to 5.10.2.

We recommend that you upgrade to the latest 5.9 release before moving to 5.10.2.

A rolling upgrade from 5.9.x to 5.10.2 is supported. Before upgrading, you should back up your data.

Warning

Tables that were created before CrateDB 4.x will not function with 5.x and must be recreated before moving to 5.x.x.

You can recreate tables using COPY TO and COPY FROM or by inserting the data into a new table.

Table of contents

See the Version 5.10.0 release notes for a full list of changes in the 5.10 series.

Fixes

  • The status field within the JSON payload returned from the HTTP root endpoint (/) now shows 503 instead of 200 if the cluster state is considered unavailable to match how the ok field behaves.

  • Fixed an issue where a filter was not pushed down to the source when the filtered column was aliased as part of a view. This could lead to an unnecessary full table scan.

  • Fixed a regression introduced in 5.6.5 that caused empty partitions to be left behind when DELETE condition contained a non-deterministic function such as NOW(), e.g.:

    DELETE FROM t WHERE day < now() - INTERVAL '3 days';

    where ‘day’ is TIMESTAMP type that is also the PARTITIONED BY column.

  • Fixed an issue that caused a NullPointerException when binding to a non-existing prepared statement via PostgreSQL wire protocol.

  • Fixed an issue that caused a SELECT query to fail if a WHERE clause had a comparison of a non-boolean column with a boolean literal, e.g.:

    SELECT int_col FROM t where int_col = true;

  • Fixed an issue that caused a SELECT query with a WHERE clause to return incorrect result if a table had a PRIMARY KEY on a boolean column and the column was referenced as a standalone operand in a logical expression. E.g.:

    SELECT * FROM tbl WHERE boolean_pk_col OR boolean_pk_col = false

  • Fixed a performance regression introduced in Version 5.10.0 that caused CREATE SNAPSHOT to use more memory. That could cause the statement to fail with OutOfMemoryError under memory pressure.

  • Fixed an issue that prevented nodes from starting when upgrading to >= 5.8 with tables containing a generated column using a user defined function.