Skip to content

Latest commit

 

History

History
94 lines (80 loc) · 8.85 KB

release-2.1.19.md

File metadata and controls

94 lines (80 loc) · 8.85 KB
Raw
title aliases summary
TiDB 2.1.19 Release Notes
/docs/dev/releases/release-2.1.19/
/docs/dev/releases/2.1.19/
TiDB 2.1.19 was released on December 27, 2019. It includes various fixes and optimizations for SQL optimizer, SQL execution engine, server, DDL, TiKV, PD, and TiDB Ansible. Some notable fixes include resolving incorrect query results, memory overhead reduction, and fixing issues related to timezone, data duplication, and panic occurrences. The release also includes upgrades and optimizations for TiDB Binlog and TiDB Ansible.

TiDB 2.1.19 Release Notes

Release date: December 27, 2019

TiDB version: 2.1.19

TiDB Ansible version: 2.1.19

TiDB

  • SQL Optimizer
    • Optimize the scenario of select max(_tidb_rowid) from t to avoid full table scan #13294
    • Fix the incorrect results caused by the incorrect value assigned to the user variable in the query and the push-down of predicates #13230
    • Fix the issue that the statistics are not accurate because a data race occurs when statistics are updated #13690
    • Fix the issue that the result is incorrect when the UPDATE statement contains both a sub-query and a stored generated column; fix the UPDATE statement execution error when this statement contains two same-named tables from different databases #13357
    • Fix the issue that the query plan might be incorrectly selected because the PhysicalUnionScan operator incorrectly sets the statistics #14134
    • Remove the minAutoAnalyzeRatio constraint to make the automatic ANALYZE more timely #14013
    • Fix the issue that the estimated number of rows is greater than 1 when the WHERE clause contains an equal condition on the unique key #13385
  • SQL Execution Engine
    • Fix the precision overflow when using int64 as the intermediate result of unit64 in ConvertJSONToInt #13036
    • Fix the issue that when the SLEEP function is in a query (for example, select 1 from (select sleep(1)) t;)), column pruning causes sleep(1) in the query to be invalid #13039
    • Reduce memory overhead by reusing Chunk in the INSERT ON DUPLICATE UPDATE statement #12999
    • Add more transaction-related fields for the slow_query table #13129:
      • Prewrite_time
      • Commit_time
      • Get_commit_ts_time
      • Commit_backoff_time
      • Backoff_types
      • Resolve_lock_time
      • Local_latch_wait_time
      • Write_key
      • Write_size
      • Prewrite_region
      • Txn_retry
    • Fix the issue that a subquery contained in an UPDATE statement is incorrectly converted; fix the UPDATE execution failure when the WHERE clause contains a subquery #13120
    • Support executing ADMIN CHECK TABLE on partitioned tables #13143
    • Fix the issue that the precision of statements such as SHOW CREATE TABLE is incomplete when ON UPDATE CURRENT_TIMESTAMP is used as a column attribute and floating point precision is specified #12462
    • Fix the panic occurred when executing the SELECT * FROM information_schema.KEY_COLUMN_USAGE statement because the foreign key is not checked when dropping, modifying or changing the column #14162
    • Fix the issue that the returned data might be duplicated when Streaming is enabled in TiDB #13255
    • Fix the Invalid time format error caused by daylight saving time #13624
    • Fix the issue that data is incorrect because the precision is lost when an integer is converted to an unsigned floating point or decimal type #13756
    • Fix the issue that an incorrect type of value is returned when the Quote function handles the NULL value #13681
    • Fix the issue that the timezone is incorrect after parsing the date from strings using gotime.Local #13792
    • Fix the issue that the result might be incorrect because the binSearch function does not return an error in the implementation of builtinIntervalRealSig #13768
    • Fix the issue that an error might occur when converting the string type to the floating point type in the INSERT statement execution #14009
    • Fix the incorrect result returned from the sum(distinct) function #13041
    • Fix the issue that data too long is returned when CAST converting the data in union of the same location to the merged type because the returned type length of the jsonUnquoteFunction function is given an incorrect value #13645
    • Fix the issue that the password cannot be set because the privilege check is too strict #13805
  • Server
    • Fix the issue that KILL CONNECTION might cause the goroutine leak #13252
    • Support getting the binlog status of all TiDB nodes via the info/all interface of the HTTP API #13188
    • Fix the failure to build the TiDB project on Windows #13650
    • Add the server-version configuration item to control and modify the version of TiDB server #13904
    • Fix the issue that the binary plugin compiled with Go1.13 does not run normally #13527
  • DDL
    • Use the table's COLLATE instead of the system's default charset in the column when a table is created and the table contains COLLATE #13190
    • Limit the length of the index name when creating a table #13311
    • Fix the issue that the length of the table name is not checked when renaming a table #13345
    • Check the width range of the BIT column #13511
    • Make the error information output from change/modify column more understandable #13798
    • Fix the issue that when executing the drop column operation that has not yet been handled by the downstream Drainer, the downstream might receive DML operations without the affected column #13974

TiKV

  • Raftstore
    • Fix the panic occurred when restarting TiKV and is_merging is given an incorrect value in the process of merging Regions and applying the Compact log #5884
  • Importer
    • Remove the limit on the gRPC message length #5809

PD

  • Improve the performance of the HTTP API for getting all Regions #1988
  • Upgrade etcd to fix the issue that etcd PreVote cannot elect a leader (downgrade not supported) #2052

Tools

  • TiDB Binlog
    • Optimize the node status information output through binlogctl #777
    • Fix the panic occurred because of the nil value in the Drainer filter configuration #802
    • Optimize the Graceful exit of Pump #825
    • Add more detailed monitoring metrics when Pump writes binlog data #830
    • Optimize Drainer's logic to refresh table information after Drainer has executed a DDL operation #836
    • Fix the issue that the commit binlog of a DDL operation is ignored when Pump does not receive this binlog #855

TiDB Ansible

  • Rename the Uncommon Error OPM monitoring item of TiDB service to Write Binlog Error and add the corresponding alert message #1038
  • Upgrade TiSpark to 2.1.8 #1063