21.35. Release 0.186
This release has a stability issue that may cause query failures in large deployments due to HTTP requests timing out.
- Fix excessive GC overhead caused by map to map cast.
- Fix implicit coercions for
ROWtypes, allowing operations between compatible types such as
- Fix issue that may cause queries containing expensive functions, such as regular expressions, to continue using CPU resources even after they are killed.
- Fix performance issue caused by redundant casts.
json_parse()to not ignore trailing characters. Previously, input such as
[1,2]abcwould successfully parse as
- Fix leak in running query counter for failed queries. The counter would increment but never decrement for queries that failed before starting.
- Reduce coordinator HTTP thread usage for queries that are queued or waiting for data.
- Reduce memory usage when building data of
- Estimate memory usage for
GROUP BYmore precisely to avoid out of memory errors.
- Add queued time and elapsed time to the client protocol.
query_max_execution_timesession property and
query.max-execution-timeconfig property. Queries will be aborted after they execute for more than the specified duration.
- Add Geospatial Functions including functions for processing Bing tiles.
- Add Spill to Disk for joins.
- Add Redshift Connector.
Resource Groups Changes
- Query Queues are deprecated in favor of Resource Groups and will be removed in a future release.
- Rename the
hardConcurrencyLimit. The old property name is deprecated and will be removed in a future release.
- Fail on unknown property names when loading the JSON config file.
JDBC Driver Changes
- Allow specifying an empty password.
FileSystem closederrors when using Kerberos authentication.
- Add support for path style access to the S3 file system. This can be enabled
by setting the
- Add an
- Remove the
These are backwards incompatible changes with the previous SPI. If you have written a connector, you will need to update your code before deploying this release.