MLSQL v1.1.7 Release roadmap

2018-12-23 本文已影响277人祝威廉

MLSQL v1.1.7 plans to release in Mid Jan 2019, this version will take almost three weeks.

MLSQL v1.1.7 Release Window:

Date	Event
Late Dec 2018	New features and Improvement
Early Jan 2019	Code freeze. Release branch cut. QA period. Focus on bug fixes, tests, stability, and docs. Generally, no new features merged.
Mid Jan 2019	Release candidates (RC), voting, etc. until final release passes

Hive JDBC supports. In spark, if you use JDBC data source to connect hive thrift jdbc server, it will fail since there is no HiveJDBCDialert implemented. Please check PR-828
MongoDB supports. More detail please check PR-822
Solr supports
Docker release of MLSQL-Cluster
Docker release of MLSQL

Refactor DataSource adaptor which is used in load/save statement. This improvement will make people involved in MLSQL community more easy to contribute DataSource implementation. People no need to worry about affecting the core code in MLSQL when adding new data source adaptor. More detail please check PR-815
Docker introduced to unit test in MLSQL. MLSQL provides an abstract server class which you can easy to use Docker to create a data source. This makes you can test if MLSQL is working properly with DataSource e.g. Kafka(multi-version), MongoDB, MySQL, and more.
DataSource direct query mode. As we know, there are standard DataSource API in spark, but sometimes it will trigger full scan on the original table which is slow and resource waste.
For example, we use SQL like select count(*) from table1 in MySQL, it will be quick if it executed in MySQL instead of spark. So we provide a direct query mode. JDBC and ElasticSearch should support this mode in v1.1.7.
The group-id will be changed from streaming.king to tech.mlsql.