Version 5.10.11

Released on 2025-07-14.

Note

If you are upgrading a cluster, you must be running CrateDB 4.0.2 or higher before you upgrade to 5.10.11.

We recommend that you upgrade to the latest 5.9 release before moving to 5.10.11.

A rolling upgrade from 5.9.x to 5.10.11 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.

Warning

(Added on 2025-08-07)

Version 5.10.11 has critical known issues. If you haven’t updated to 5.10.11 yet, it is recommended to skip 5.10.11 and move to 5.10.12 directly.

  • Replication errors due to the seqNo and primaryTerm information missing when replicating new records written as part of INSERT INTO or COPY FROM statements from the primary to replica shards when running in a mixed cluster with 5.10.11 and an earlier version.

  • A regression introduced with 5.10.11 that caused values of undeterministic DEFAULT and GENERATED columns to be NULL (instead of having generated values) when using UPDATE or INSERT... ON CONFLICT... statements.

Table of contents

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

Fixes

  • Fixed an issue that allowed to CREATE tables with more columns than the limit defined by the mapping.total_fields.limit setting.

  • Fixed an issue that would prevent the cluster.routing.allocation.awareness.attributes and cluster.routing.allocation.awareness.force.*.values to be shown under the settings column of the Cluster table.

  • Fixed an issue that wrongly allowed columns of tables created by logical replication subscriptions to be renamed, dropped or added.

  • Fixed an issue that wrongly allowed tables created by logical replication publications to be swapped.

  • Fixed an issue that could cause streaming errors, leading to shard failures and re-replication if using INSERT INTO statements with ON CONFLICT clause combined with bulk requests if some operations ran into a conflict while others didn’t.

  • Fixed an issue that caused analysis errors of SQL batch/bulk statements to be wrongly reported as runtime errors, resulting in an error per batch instead of a single error for the whole batch operation. Both client interfaces, HTTP and PG, are affected.

  • Fixed an issue which caused inserts of object arrays with columns having mixed types to partially fail. It used the type of the first value seen, and failed for the values with a different type. The new behavior is that it will use the type with the higher precendence, and cast the other values if possible.

  • Fixed incorrect JSON response formatting for HTTP bulk operations with a single argument that results in a runtime error. The response now follows the structure specified in bulk-errors, independent of the number of bulk arguments.

    Warning

    This change may break clients that rely on the previous error response format when using bulk_args with a single bulk argument.

  • Fixed an issue that caused a user to have duplicate roles granted by different grantors.

  • Fixed an issue that prevented the superuser crate from revoking granted roles.