Version 5.8.1¶
Released on 2024-07-29.
Note
If you are upgrading a cluster, you must be running CrateDB 4.0.2 or higher before you upgrade to 5.8.1.
We recommend that you upgrade to the latest 5.7 release before moving to 5.8.1.
A rolling upgrade from 5.7.x to 5.8.1 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 5.8.0 release notes for a full list of changes in the 5.8 series.
Fixes¶
Changed behavior to partitioned tables by automatically calculating number_of_routing_shards, so that they behave the same as simple tables, for which the change was already introduced in Version 5.8.0. With this fix, it’s possible to alter the number_of_shards for a single partition of a partitioned table.
Fixed an issue that would cause the number_of_replicas setting of a table to be reset to
1
, when the number of shards for that table is increased or decreased.Fixed an issue that would add a whitespace character at the beginning of some lines in the files containing the rows which are exported by executing COPY TO using the local filesystem on a whole partitioned table.
Changed the behavior of LIKE and ILIKE operators to throw an error, when the pattern to match ends with the
ESCAPE
character. Previously, theESCAPE
character was ignored, and the result was computed based on the rest of the pattern.Fixed an issue that caused an
IndexOutOfBoundsException
when the max_by aggregation was called with a literal as searchfield parameter instead of a column e.g.:SELECT MAX_BY(x, 1) from tbl;
Fixed an issue that caused write operations to fail if the table contained generated columns with a cast to
geo_shape
.Fixed a rare race condition that could lead to queries appearing stuck and eventually time out after 60 seconds if they were executed while a shard was being created. This could happen right after a table or a partition is created, during shard relocation or node restarts.
Fixed an issue that could lead to
ANALYZE
over account the number of documents in a table if shards were relocating.Fixed a regression introduced in 5.8.0 that led to error when running a query
SELECT * FROM sys.nodes
in a mixed cluster.Fixed a regression introduced in 5.8.0 that prevented cluster to become GREEN after a rolling upgrade.
Fixed a regression introduced in 5.8.0 that lead to error when running an aggregations in a mixed cluster.
Fixed an issue that caused
WHERE
clause to fail to filter rows when the clause contained array scalar functions underNOT
operator. The affected scalars include array_min(array), array_max(array), array_sum(array), array_avg(array), array_upper(anyarray, dimension), array_lower(anyarray, dimension) and array_length(anyarray, dimension).