代码改变世界

【MySQL】MySQL压缩表的一个bug

  abce  阅读(168)  评论(0编辑  收藏  举报

版本:MySQL5.7

 

创建一张测试表:

1
2
3
4
5
6
7
8
9
10
create table abce_comp(
    id bigint(20) unsigned not null,
    identification_id int(10) unsigned default null,
    timestamp datetime not null,
    action varchar(50) not null,
    result varchar(50) not null,
    primary key(id),
    key index_abce_comp_result(result),
    key index_abce_comp_timestamp(timestamp)
);

插入10万行数据,用于测试:

1
for NUM in {1..100000}; do mysql -uroot -p"xxxxxxxx" abce -e "insert into abce_comp (id, identification_id, timestamp, action, result) values ($NUM,$NUM*100,now(),concat('string',$NUM),concat('VeryVeryLargeString',$NUM))"; done

我们来看看表的大小。再执行analyze table之前,将innodb_stats_persistent_sample_pages=100000。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
> set global innodb_stats_persistent_sample_pages=100000;
Query OK, 0 rows affected (0.00 sec)
 
> analyze table abce_comp;
+----------------+---------+----------+----------+
| Table          | Op      | Msg_type | Msg_text |
+----------------+---------+----------+----------+
| abce.abce_comp | analyze | status   | OK       |
+----------------+---------+----------+----------+
1 row in set (0.08 sec)
> select table_schema, table_name, table_rows,(data_length+index_length+data_free)/1024/1024 TOTAL_MB, create_options from information_schema.tables where table_name='abce_comp';
+--------------+------------+------------+-------------+----------------+
| table_schema | table_name | table_rows | TOTAL_MB    | create_options |
+--------------+------------+------------+-------------+----------------+
| abce         | abce_comp  |     100000 | 22.57812500 |                |
+--------------+------------+------------+-------------+----------------+
1 row in set (0.00 sec)

现在,开始压缩表,将key_block_size设置成4(块大小是我随机选的)。

1
2
3
> alter table abce_comp row_format=compressed,key_block_size=4,algorithm=inplace,lock=none;
Query OK, 0 rows affected (1.98 sec)
Records: 0  Duplicates: 0  Warnings: 0

按照上面的步骤,再次收集表的统计信息:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
> set global innodb_stats_persistent_sample_pages=100000;
Query OK, 0 rows affected (0.00 sec)
 
> analyze table abce_comp;
+----------------+---------+----------+----------+
| Table          | Op      | Msg_type | Msg_text |
+----------------+---------+----------+----------+
| abce.abce_comp | analyze | status   | OK       |
+----------------+---------+----------+----------+
1 row in set (0.06 sec)
 
> select table_schema, table_name, table_rows,(data_length+index_length+data_free)/1024/1024 TOTAL_MB, create_options from information_schema.tables where table_name='abce_comp';
+--------------+------------+------------+------------+----------------------------------------+
| table_schema | table_name | table_rows | TOTAL_MB   | create_options                         |
+--------------+------------+------------+------------+----------------------------------------+
| abce         | abce_comp  |     100000 | 7.14843750 | row_format=COMPRESSED KEY_BLOCK_SIZE=4 |
+--------------+------------+------------+------------+----------------------------------------+
1 row in set (0.00 sec)

可以看到,表已经被压缩了。再来看看表的结构:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
> show create table abce_comp\G
*************************** 1. row ***************************
       Table: abce_comp
Create Table: CREATE TABLE `abce_comp` (
  `id` bigint(20) unsigned NOT NULL,
  `identification_id` int(10) unsigned DEFAULT NULL,
  `timestamp` datetime NOT NULL,
  `action` varchar(50) NOT NULL,
  `result` varchar(50) NOT NULL,
  PRIMARY KEY (`id`),
  KEY `index_abce_comp_result` (`result`),
  KEY `index_abce_comp_timestamp` (`timestamp`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 ROW_FORMAT=COMPRESSED KEY_BLOCK_SIZE=4
1 row in set (0.00 sec)

如何给表取消压缩呢?是不是应该很简单。

1
2
3
> alter table abce_comp row_format=default,algorithm=inplace,lock=none;
Query OK, 0 rows affected (1.85 sec)
Records: 0  Duplicates: 0  Warnings: 0

语句执行成功,看起来是起效果了:

1
2
3
4
5
6
> select table_schema, table_name, table_rows,(data_length+index_length+data_free)/1024/1024 TOTAL_MB, create_options from information_schema.tables where table_name='abce_comp';
+--------------+------------+------------+------------+------------------+
| table_schema | table_name | table_rows | TOTAL_MB   | create_options   |
+--------------+------------+------------+------------+------------------+
| abce         | abce_comp  |     100000 | 7.14843750 | KEY_BLOCK_SIZE=4 |
+--------------+------------+------------+------------+------------------+

再看看表结构:

1
2
3
4
5
6
7
8
9
10
11
12
13
> show create table abce_comp\G
*************************** 1. row ***************************
       Table: abce_comp
Create Table: CREATE TABLE `abce_comp` (
  `id` bigint(20) unsigned NOT NULL,
  `identification_id` int(10) unsigned DEFAULT NULL,
  `timestamp` datetime NOT NULL,
  `action` varchar(50) NOT NULL,
  `result` varchar(50) NOT NULL,
  PRIMARY KEY (`id`),
  KEY `index_abce_comp_result` (`result`),
  KEY `index_abce_comp_timestamp` (`timestamp`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 KEY_BLOCK_SIZE=4

哪里出了点差错,表的KEY_BLOCK_SIZE仍然是4。

 

再来一次尝试:

1
2
3
> alter table abce_comp row_format=default,key_block_size=0,algorithm=inplace,lock=none;
Query OK, 0 rows affected (0.91 sec)
Records: 0  Duplicates: 0  Warnings: 0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
> select table_schema, table_name, table_rows,(data_length+index_length+data_free)/1024/1024 TOTAL_MB, create_options from information_schema.tables where table_name='abce_comp';
+--------------+------------+------------+-------------+----------------+
| table_schema | table_name | table_rows | TOTAL_MB    | create_options |
+--------------+------------+------------+-------------+----------------+
| abce         | abce_comp  |     100000 | 18.54687500 |                |
+--------------+------------+------------+-------------+----------------+
> show create table abce_comp\G
*************************** 1. row ***************************
       Table: abce_comp
Create Table: CREATE TABLE `abce_comp` (
  `id` bigint(20) unsigned NOT NULL,
  `identification_id` int(10) unsigned DEFAULT NULL,
  `timestamp` datetime NOT NULL,
  `action` varchar(50) NOT NULL,
  `result` varchar(50) NOT NULL,
  PRIMARY KEY (`id`) KEY_BLOCK_SIZE=4,
  KEY `index_abce_comp_result` (`result`) KEY_BLOCK_SIZE=4,
  KEY `index_abce_comp_timestamp` (`timestamp`) KEY_BLOCK_SIZE=4
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4

又出错了!主键和辅助索引仍然是KEY_BLOCK_SIZE=4

 

尽管当表从压缩转换为非压缩表时,索引的KEY_BLOCK_SIZE在内部会遵循表的索引,但CREATE TABLE语句不会。起初这是一个美学/外观问题,但是当你进行dump时这是一个真正的问题,因为CREATE TABLE留下了KEY_BLOCK_SIZE值,这一点很不好。这是mysqldump的输出:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
# mysqldump -uroot -pxxxxxxxx abce abce_comp --no-data > abce_comp.sql
# more abce_comp.sql
-- MySQL dump 10.13  Distrib 5.7.37, for linux-glibc2.12 (x86_64)
--
-- Host: localhost    Database: abce
-- ------------------------------------------------------
-- Server version       5.7.37-log
 
/*!40101 SET @OLD_CHARACTER_SET_CLIENT=@@CHARACTER_SET_CLIENT */;
/*!40101 SET @OLD_CHARACTER_SET_RESULTS=@@CHARACTER_SET_RESULTS */;
/*!40101 SET @OLD_COLLATION_CONNECTION=@@COLLATION_CONNECTION */;
/*!40101 SET NAMES utf8 */;
/*!40103 SET @OLD_TIME_ZONE=@@TIME_ZONE */;
/*!40103 SET TIME_ZONE='+00:00' */;
/*!40014 SET @OLD_UNIQUE_CHECKS=@@UNIQUE_CHECKS, UNIQUE_CHECKS=0 */;
/*!40014 SET @OLD_FOREIGN_KEY_CHECKS=@@FOREIGN_KEY_CHECKS, FOREIGN_KEY_CHECKS=0 */;
/*!40101 SET @OLD_SQL_MODE=@@SQL_MODE, SQL_MODE='NO_AUTO_VALUE_ON_ZERO' */;
/*!40111 SET @OLD_SQL_NOTES=@@SQL_NOTES, SQL_NOTES=0 */;
SET @MYSQLDUMP_TEMP_LOG_BIN = @@SESSION.SQL_LOG_BIN;
SET @@SESSION.SQL_LOG_BIN= 0;
 
--
-- Table structure for table `abce_comp`
--
 
DROP TABLE IF EXISTS `abce_comp`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */;
CREATE TABLE `abce_comp` (
  `id` bigint(20) unsigned NOT NULL,
  `identification_id` int(10) unsigned DEFAULT NULL,
  `timestamp` datetime NOT NULL,
  `action` varchar(50) NOT NULL,
  `result` varchar(50) NOT NULL,
  PRIMARY KEY (`id`) KEY_BLOCK_SIZE=4,
  KEY `index_abce_comp_result` (`result`) KEY_BLOCK_SIZE=4,
  KEY `index_abce_comp_timestamp` (`timestamp`) KEY_BLOCK_SIZE=4
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;
/*!40101 SET character_set_client = @saved_cs_client */;
 
--
-- GTID state at the end of the backup
--
 
SET @@GLOBAL.GTID_PURGED='10ccf388-efb0-11ec-b477-08002762387f:1-100014';
/*!40103 SET TIME_ZONE=@OLD_TIME_ZONE */;
 
/*!40101 SET SQL_MODE=@OLD_SQL_MODE */;
/*!40014 SET FOREIGN_KEY_CHECKS=@OLD_FOREIGN_KEY_CHECKS */;
/*!40014 SET UNIQUE_CHECKS=@OLD_UNIQUE_CHECKS */;
/*!40101 SET CHARACTER_SET_CLIENT=@OLD_CHARACTER_SET_CLIENT */;
/*!40101 SET CHARACTER_SET_RESULTS=@OLD_CHARACTER_SET_RESULTS */;
/*!40101 SET COLLATION_CONNECTION=@OLD_COLLATION_CONNECTION */;
/*!40111 SET SQL_NOTES=@OLD_SQL_NOTES */;
 
-- Dump completed on 2022-07-11 21:49:52

如你所见,似乎无法使用全局ALTER TABLE命令(如果我们可以这样称呼它)来反转表定义中的索引的key_block_size,因此我们将进行最后一次尝试:

1
2
3
4
5
> alter table abce_comp
    -> drop primary key, add primary key (id),
    -> drop key index_abce_comp_result, add key index_abce_comp_result (result),
    -> drop key index_abce_comp_timestamp, add key index_abce_comp_timestamp (timestamp),
    -> row_format=default,key_block_size=0,algorithm=inplace,lock=none;

现在再看看结果:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
> show create table abce_comp\G
*************************** 1. row ***************************
       Table: abce_comp
Create Table: CREATE TABLE `abce_comp` (
  `id` bigint(20) unsigned NOT NULL,
  `identification_id` int(10) unsigned DEFAULT NULL,
  `timestamp` datetime NOT NULL,
  `action` varchar(50) NOT NULL,
  `result` varchar(50) NOT NULL,
  PRIMARY KEY (`id`),
  KEY `index_abce_comp_result` (`result`),
  KEY `index_abce_comp_timestamp` (`timestamp`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4
1 row in set (0.00 sec)
 
> select table_schema, table_name, table_rows,(data_length+index_length+data_free)/1024/1024 TOTAL_MB, create_options from information_schema.tables where table_name='abce_comp';
+--------------+------------+------------+-------------+----------------+
| table_schema | table_name | table_rows | TOTAL_MB    | create_options |
+--------------+------------+------------+-------------+----------------+
| abce         | abce_comp  |     100000 | 18.54687500 |                |
+--------------+------------+------------+-------------+----------------+

显然,这是一个bug:https://bugs.mysql.com/bug.php?id=56628

在MySQL 5.7中,完全还原的唯一方法(至少在表及其索引的定义中)是重新生成主键及其所有索引。这听起来像是一个终端解决方案,但如果你使用mysqldump进行备份(我们总是建议使用 Percona XtraBackup 来实现这些目的,它更快更高效)这是一个需要考虑的问题,因为它在其定义中保留了那些错误的定义。

幸运的是,这在MySQL 8中已修复。MySQL 8中,执行以下语句,就可以将压缩表转换成非压缩表

1
> alter table abce_comp ROW_FORMAT=DEFAULT, KEY_BLOCK_SIZE=0,ALGORITHM=INPLACE,LOCK=NONE;
相关博文:
阅读排行:
· DeepSeek 开源周回顾「GitHub 热点速览」
· 物流快递公司核心技术能力-地址解析分单基础技术分享
· .NET 10首个预览版发布:重大改进与新特性概览!
· AI与.NET技术实操系列(二):开始使用ML.NET
· .NET10 - 预览版1新功能体验(一)
历史上的今天:
2015-07-17 Putty设置删除
2015-07-17 ssh/scp 远程连接ssh非默认端口方法
2015-07-17 查看LINUX版本
点击右上角即可分享
微信分享提示