Crate 0.55.1 发布数据存储系统

Crate 0.55.1 发布了，主要更新内容如下：

– Fix: Failures on large bulk operations could lead to an stackoverflow error and unresponsive cluster.
– Updated crash to 0.16.2 which contains following changes:
– Fix: crash_standalone could crash because it contained non-ascii character but no encoding was declared
– Fixed “read file” command (“\r“) for legacy Python 2.7
– Fixed result info command output for whitespace other than space characters in queries.
– Fixed call of `sysinfo` command when launching the shell. stackoverflow error and unresponsive cluster.
– Fix: Ordered queries with limit 0 failed
– Fixed error handling while preparing execution plans.
– Fix: Number of bulk results were equal to the total number of items on bulk inserts/updates, ignoring possible multi-value bulk inserts/updates.
– Fixed path to sigar lib in crate.bat for Windows
– Fixed proper validation of the “blocks.*“ table settings at all
SQL operations.
– Fix: Allow select statements with more than one level nested arrays in table columns.
eg: SELECT categories[‘items’][‘id’] FROM assets
– Fixed an edge case that could cause sorted join queries to fail.
– Fix: Allow concurrent deletion of partition without raising an exception.
– Throw a more meaningful error if an attempt is made to order by an invalid field in a query with aggregations.
– Fixed cpu used bug. It was returning -1, even if cpu info loaded from sigar.
– Update crate-admin to 0.18.1 which contains following changes:
– Resolve path to “/_sql“ endpoint correctly when Crate is served at a different location than “/“
– Note that 0.55.0 didn’t include 0.18 but 0.17, although it was mentioned in the changelog. For a list of changes from 0.17 to 0.18 see the 0.55.0 release section.
– Fix in Cloud AWS plugin: Use the current region of the instance if no EC2 region is specified in the configuration.
– Fixed a validation issue that allowed setting a primary key column value to null