技术分享 | 用好 MySQL 的 MRR 优化器

于由社区 蒋乐兴发布

作者：蒋乐兴

MySQL DBA，擅长 python 和 SQL，目前维护着 github 的两个开源项目：mysqltools 、dbmc 以及独立博客：https://www.sqlpy.com。

本文来源：原创投稿

*爱可生开源社区出品，原创内容未经授权不得随意使用，转载请联系小编并注明来源。

MRR 要解决的问题

MRR 是 MySQL 针对特定查询的一种优化手段。假设一个查询有二级索引可用，读完二级索引后要回表才能查到那些不在当前二级索引上的列值，由于二级索引上引用的主键值不一定是有序的，因此就有可能造成大量的随机 IO，如果回表前把主键值给它排一下序，那么在回表的时候就可以用顺序 IO 取代原本的随机 IO。

环境准备

为了实验我们要准备一下表结构和数据。
  1. -- 创建表
  2. mysql> show create table t;
  3. +----------------------------------------------------------------------+
  4. | Table | Create Table                                                                                                                                                                                                                                                                                                                                                                               |
  5. +----------------------------------------------------------------------+
  6. | t     | CREATE TABLE `t` (
  7.   `id` int NOT NULL AUTO_INCREMENT,
  8.   `i0` int NOT NULL,
  9.   `i1` int NOT NULL,
  10.   `i2` int NOT NULL,
  11.   `i3` int NOT NULL,
  12.   `c0` varchar(128) NOT NULL,
  13.   `c1` varchar(128) NOT NULL,
  14.   `f0` float NOT NULL,
  15.   `f1` float NOT NULL,
  16.   PRIMARY KEY (`id`),
  17.   KEY `idx_i0` (`i0`)
  18. ) ENGINE=InnoDB
  19. +----------------------------------------------------------------------+
  20. 1 row in set (0.00 sec)

  22. -- 造数据
  23. mysql> select count(*) from t;
  24. +----------+
  25. | count(*) |
  26. +----------+
  27. |  1120000 |
  28. +----------+
  29. 1 row in set (0.77 sec)

  31. --
  32. update t set i0 = id % 100;

MRR 的优化效果

1. 有 MRR 优化（Using MRR）时 SQL 的耗时情况。

  1. mysql> explain select i0,i3 from t where i0 between 1 and 2;
  2. +----+-------------+-------+------------+-------+---------------+--------+---------+------+-------+----------+----------------------------------+
  3. | id | select_type | table | partitions | type  | possible_keys | key    | key_len | ref  | rows  | filtered | Extra                            |
  4. +----+-------------+-------+------------+-------+---------------+--------+---------+------+-------+----------+----------------------------------+
  5. |  1 | SIMPLE      | t     | NULL       | range | idx_i0        | idx_i0 | 4       | NULL | 43968 |   100.00 | Using index condition; Using MRR |
  6. +----+-------------+-------+------------+-------+---------------+--------+---------+------+-------+----------+----------------------------------+
  7. 1 row in set, 1 warning (0.00 sec)

  9. mysql> select i0,i3 from t where i0 between 1 and 2;
  10. 22400 rows in set (0.80 sec)

2. 关闭 MRR 优化。

  1. set optimizer_switch = 'index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,engine_condition_pushdown=on,index_condition_pushdown=on,mrr=off,mrr_cost_based=on,block_nested_loop=on,batched_key_access=off,materialization=on,semijoin=on,loosescan=on,firstmatch=on,duplicateweedout=on,subquery_materialization_cost_based=on,use_index_extensions=on,condition_fanout_filter=on,derived_merge=on,use_invisible_indexes=off,skip_scan=on,hash_join=on';

  3. mysql> explain select i0,i3 from t where i0 between 1 and 2;
  4. +----+-------------+-------+------------+-------+---------------+--------+---------+------+-------+----------+-----------------------+
  5. | id | select_type | table | partitions | type  | possible_keys | key    | key_len | ref  | rows  | filtered | Extra                 |
  6. +----+-------------+-------+------------+-------+---------------+--------+---------+------+-------+----------+-----------------------+
  7. |  1 | SIMPLE      | t     | NULL       | range | idx_i0        | idx_i0 | 4       | NULL | 43968 |   100.00 | Using index condition |
  8. +----+-------------+-------+------------+-------+---------------+--------+---------+------+-------+----------+-----------------------+
  9. 1 row in set, 1 warning (0.00 sec)

  11. mysql> select i0,i3 from t where i0 between 1 and 2;
  12. 22400 rows in set (2.56 sec)
结论
就刚才的测试场景开启 MRR 优化可以得到 3 倍的性能提升。

MRR 的优化器参数调整

如果想关闭 MRR 优化的话，就要把优化器开关 mrr 设置为 off。

默认只有在优化器认为 MRR 可以带来优化的情况下才会走 MRR，如果你想不管什么时候能走 MRR 的都走 MRR 的话，你要把 mrr_cost_based 设置为 off，不过最好不要这么干，因为这确实是一个坑，MRR 不一定什么时候都好，全表扫描有时候会更加快，如果在这种场景下走 MRR 就完成了。

开启 MRR 关闭基于开销的优化。

  1. -- mrr=on,mrr_cost_based=off
  2. set optimizer_switch = 'index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,engine_condition_pushdown=on,index_condition_pushdown=on,mrr=on,mrr_cost_based=off,block_nested_loop=on,batched_key_access=off,materialization=on,semijoin=on,loosescan=on,firstmatch=on,duplicateweedout=on,subquery_materialization_cost_based=on,use_index_extensions=on,condition_fanout_filter=on,derived_merge=on,use_invisible_indexes=off,skip_scan=on,hash_join=on';

  4. mysql> explain select i0,i3 from t where i0 between 1 and  10;
  5. +----+-------------+-------+------------+-------+---------------+--------+---------+------+--------+----------+----------------------------------+
  6. | id | select_type | table | partitions | type  | possible_keys | key    | key_len | ref  | rows   | filtered | Extra                            |
  7. +----+-------------+-------+------------+-------+---------------+--------+---------+------+--------+----------+----------------------------------+
  8. |  1 | SIMPLE      | t     | NULL       | range | idx_i0        | idx_i0 | 4       | NULL | 218492 |   100.00 | Using index condition; Using MRR |
  9. +----+-------------+-------+------------+-------+---------------+--------+---------+------+--------+----------+----------------------------------+
  10. 1 row in set, 1 warning (0.00 sec)

  12. select i0,i3 from t where i0 between 1 and  10;
  13. 112000 rows in set (4.86 sec)

开启 MRR 开启基于开销的优化。

  1. -- mrr=on,mrr_cost_based=on
  2. set optimizer_switch = 'index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,engine_condition_pushdown=on,index_condition_pushdown=on,mrr=on,mrr_cost_based=on,block_nested_loop=on,batched_key_access=off,materialization=on,semijoin=on,loosescan=on,firstmatch=on,duplicateweedout=on,subquery_materialization_cost_based=on,use_index_extensions=on,condition_fanout_filter=on,derived_merge=on,use_invisible_indexes=off,skip_scan=on,hash_join=on';

  4. mysql> explain select i0,i3 from t where i0 between 1 and  10;
  5. +----+-------------+-------+------------+------+---------------+------+---------+------+---------+----------+-------------+
  6. | id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows    | filtered | Extra       |
  7. +----+-------------+-------+------------+------+---------------+------+---------+------+---------+----------+-------------+
  8. |  1 | SIMPLE      | t     | NULL       | ALL  | idx_i0        | NULL | NULL    | NULL | 1121902 |    19.48 | Using where |
  9. +----+-------------+-------+------------+------+---------------+------+---------+------+---------+----------+-------------+
  10. 1 row in set, 1 warning (0.00 sec)

  12. mysql> select i0,i3 from t where i0 between 1 and  10;
  13. 112000 rows in set (1.52 sec)

可以看到当 mrr_cost_based = OFF 的情况下用时 4.86s，mrr_cost_based = ON 的情况下用时 1.52s，总的来说 mrr_cost_based 是非常关键的建议始终打开。

MRR 的参数优化

MRR 要把主键排个序，这样之后对磁盘的操作就是由顺序读代替之前的随机读。从资源的使用情况上来看就是让 CPU 和内存多做点事，来换磁盘的顺序读。然而排序是需要内存的，这块内存的大小就由参数 read_rnd_buffer_size 来控制。

read_rnd_buffer_size 太小无法启用 MRR 功能。

  1. mysql> select @@read_rnd_buffer_size;
  2. +------------------------+
  3. | @@read_rnd_buffer_size |
  4. +------------------------+
  5. |                 262144 |
  6. +------------------------+
  7. 1 row in set (0.00 sec)

  9. mysql> explain select i0,i3 from t where i0 between 1 and  12;
  10. +----+-------------+-------+------------+------+---------------+------+---------+------+---------+----------+-------------+
  11. | id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows    | filtered | Extra       |
  12. +----+-------------+-------+------------+------+---------------+------+---------+------+---------+----------+-------------+
  13. |  1 | SIMPLE      | t     | NULL       | ALL  | idx_i0        | NULL | NULL    | NULL | 1121902 |    23.57 | Using where |
  14. +----+-------------+-------+------------+------+---------------+------+---------+------+---------+----------+-------------+
  15. 1 row in set, 1 warning (0.00 sec)

放大 read_rnd_buffer_size 让 MySQL 有足够的资源用于 MRR 。

  1. mysql> set read_rnd_buffer_size = 32 * 1024 * 1024;
  2. Query OK, 0 rows affected (0.00 sec)

  4. mysql> explain select i0,i3 from t where i0 between 1 and  12;
  5. +----+-------------+-------+------------+-------+---------------+--------+---------+------+--------+----------+----------------------------------+
  6. | id | select_type | table | partitions | type  | possible_keys | key    | key_len | ref  | rows   | filtered | Extra                            |
  7. +----+-------------+-------+------------+-------+---------------+--------+---------+------+--------+----------+----------------------------------+
  8. |  1 | SIMPLE      | t     | NULL       | range | idx_i0        | idx_i0 | 4       | NULL | 264436 |   100.00 | Using index condition; Using MRR |
  9. +----+-------------+-------+------------+-------+---------------+--------+---------+------+--------+----------+----------------------------------+
  10. 1 row in set, 1 warning (0.00 sec)
分类： 技术分享
标签：

相关文章

技术分享

技术译文 | 回顾 Oracle 在 MySQL 8.0 中的管理工作

对于新兴技术而言，社区管理并不难。经过 29 年的发展，MySQL 已 阅读更多…

技术分享

技术分享 | MySQL Undo 工作机制历史演变

前几天遇到一个关于 Undo 变大的 CASE，为了方便后续排查问题， 阅读更多…

OceanBase

OceanBase 的 Oracle 模式竟然可以使用 Repeatable Read？

OceanBase 的 Oracle 模式不是只支持 2 种隔离级别： 阅读更多…