TiDB 2.1 GA Release Notes

On November 30, 2018, TiDB 2.1 GA is released. See the following updates in this release. Compared with TiDB 2.0, this release has great improvements in stability, performance, compatibility, and usability.

TiDB

  • SQL Optimizer

    • Optimize the selection range of Index Join to improve the execution performance

    • Optimize the selection of outer table for Index Join and use the table with smaller estimated value of Row Count the as the outer table

    • Optimize Join Hint TIDB_SMJ so that Merge Join can be used even without proper index available

    • Optimize Join Hint TIDB_INLJ to specify the Inner table to Join

    • Optimize correlated subquery, push down Filter, and extend the index selection range, to improve the efficiency of some queries by orders of magnitude

    • Support using Index Hint and Join Hint in the UPDATE and DELETE statement

    • Support pushing down more functions: ABS/CEIL/FLOOR/IS TRUE/IS FALSE

    • Optimize the constant folding algorithm for the IF and IFNULL built-in functions

    • Optimize the output of the EXPLAIN statement and use hierarchy structure to show the relationship between operators

  • SQL executor

    • Refactor all the aggregation functions and improve execution efficiency of the Stream and Hash aggregation operators

    • Implement the parallel Hash Aggregate operators and improve the computing performance by 350% in some scenarios

    • Implement the parallel Project operators and improve the performance by 74% in some scenarios

    • Read the data of the Inner table and Outer table of Hash Join concurrently to improve the execution performance

    • Optimize the execution speed of the REPLACE INTO statement and increase the performance nearly by 10 times

    • Optimize the memory usage of the time data type and decrease the memory usage of the time data type by fifty percent

    • Optimize the point select performance and improve the point select efficiency result of Sysbench by 60%

    • Improve the performance of TiDB on inserting or updating wide tables by 20 times

    • Support configuring the memory upper limit of a single statement in the configuration file

    • Optimize the execution of Hash Join, if the Join type is Inner Join or Semi Join and the inner table is empty, return the result without reading data from the outer table

    • Support using the EXPLAIN ANALYZE statement to check the runtime statistics including the execution time and the number of returned rows of each operator

  • Statistics

    • Support enabling auto ANALYZE statistics only during certain period of the day

    • Support updating the table statistics automatically according to the feedback of the queries

    • Support configuring the number of buckets in the histogram using the ANALYZE TABLE WITH BUCKETS statement

    • Optimize the Row Count estimation algorithm using histogram for mixed queries of equality query and range queries

  • Expressions

    • Support following built-in function:

      • json_contains

      • json_contains_path

      • encode/decode

  • Server

    • Support queuing the locally conflicted transactions within tidb-server instance to optimize the performance of conflicted transactions

    • Support Server Side Cursor

    • Add the HTTP API

      • Scatter the distribution of table Regions in the TiKV cluster

      • Control whether to open the general log

      • Support modifying the log level online

      • Check the TiDB cluster information

  • DDL

    • Support the parallel execution of the Add index statement and other statements to avoid the time consuming Add index operation blocking other operations

    • Optimize the execution speed of ADD INDEX and improve it greatly in some scenarios

    • Support the select tidb_is_ddl_owner() statement to facilitate deciding whether TiDB is DDL Owner

    • Support the ALTER TABLE FORCE syntax

    • Support the ALTER TABLE RENAME KEY TO syntax

    • Add the table name and database name in the output information of admin show ddl jobs

    • Support using the ddl/owner/resign HTTP interface to release the DDL owner and start electing a new DDL owner

  • Compatibility

    • Support more MySQL syntaxes

    • Make the BIT aggregate function support the ALL parameter

    • Support the SHOW PRIVILEGES statement

    • Support the CHARACTER SET syntax in the LOAD DATA statement

    • Support the IDENTIFIED WITH syntax in the CREATE USER statement

    • Support the LOAD DATA IGNORE LINES statement

    • The Show ProcessList statement returns more accurate information

Placement Driver (PD)

TiKV

Tools

Upgrade caveat

  • TiDB 2.1 does not support downgrading to v2.0.x or earlier due to the adoption of the new storage engine
  • Parallel DDL is enabled in TiDB 2.1, so the clusters with TiDB version earlier than 2.0.1 cannot upgrade to 2.1 using rolling update. You can choose either of the following two options:

    • Stop the cluster and upgrade to 2.1 directly
    • Roll update to 2.0.1 or later 2.0.x versions, and then roll update to the 2.1 version
  • If you upgrade from TiDB 2.0.6 or earlier to TiDB 2.1, check if there is any ongoing DDL operation, especially the time consuming Add Index operation, because the DDL operations slow down the upgrading process. If there is ongoing DDL operation, wait for the DDL operation finishes and then roll update.

Was this page helpful?