Crate 0.54.9 发布数据存储系统

Crate 0.54.9 发布了。更新如下：

Improved retry logic of bulk inserts/updates in order to prevent stack overflows caused by recursive retry calls that could lead to hanging update and imports using “copy from“.
Fixed a race condition that could lead to relative write operations (`x = x+ 1`) to be executed twice.
In certain cases settings provided via “-Des.“ system properties could be overridden by the “crate.yml“ provided in “$CRATE_HOME/config“
Fixed an issue that prevented generated columns to work correctly with arrays.
Restrict creation of tables having the primary key constraint within an array column type or its children.
with arrays.
Fixed an issue that could lead to the `internal_action_long_timeout` and `recovery_activity_timeout` settings to be reset to a wrong default value on a setting update
Fixed an issue that could lead to stuck queries if a node dropped out of the cluster.
Fix: Changing the column policy on a partitioned table throws an exception if more than 1 partition exists.
Fix: Altering a setting on an empty partitioned table is changing the setting on all exising tables.
Fix: Throw proper error if a column is used twice in primary key constraint.
Fix: Improved error handling to make sure JOIN queries don’t get stuck on shard failures.
Fix bug which could cause the MATCH predicate in WHERE clause for partitioned tables to fail. This happens in cases when a column ident of the MATCH predicate is not a primary key column.
Fix: While restoring a snapshot with multiple indices inside, selecting concrete indices can result in a wrong index being restored.
Fix: Insert by query into a table with nested primary keys or nested partitioned by columns defined did not work.