MySQL Database Backup Methods Season 2 – MySQL Data Dumper

Posted 2020-10-04

说到mysql数据库的备份, MySQL Data Dumper(项目)也是常用的工具, 其有两个可执行程序: mydumper, 负责导出数据; myloader,  负责导入数据. mydumper相对于mysqldump, 多了些特性, 在下面分析选项的过程中能体会到.

 

 

由于是第三方工具, 先来看下安装, 及可能遇到的问题.

a. mydumper需要依赖一些开发库, 使用yum安装即可.

[email protected]: ~# yum install glib* zlib* pcre* -y

 

b. 添加连接MySQL需要的动态链接库.

[email protected]: ~# cat /etc/ld.so.conf.d/mysql.conf

/opt/mysql/lib

 

[email protected]: ~#ldconfig

 

[email protected]: ~#ldconfig --print-cache | grep ‘mysql‘

         libmysqlclient.so.18 (libc6,x86-64)=> /opt/mysql/lib/libmysqlclient.so.18

 

[email protected]: ~# ls -l /opt/mysql/lib/libmysqlclient.so.18

lrwxrwxrwx 1 rootroot 26 Aug 25 14:21 /opt/mysql/lib/libmysqlclient.so.18 ->libmysqlclient_r.so.18.1.0

 

c. 编译安装.

[email protected]: ~# cmake -DCMAKE_INSTALL_PREFIX=/usr/local/mydumper

 

[email protected]: ~# make install

 

添加可执行命令的路径到环境变量PATH中.

[email protected]: ~$grep ‘PATH‘ .bash_profile

PATH=/usr/local/mydumper/bin:/opt/mysql/bin/:$PATH:$HOME/bin

export PATH

 

d. 在命令行敲入mydumper回车, 看下面的返回信息, 安装是正常的.

[email protected]: ~$mydumper

**(mydumper:723): CRITICAL **: Error connecting to database: Access denied foruser ‘root‘@‘localhost‘ (using password:NO)

 

[email protected]: ~$myloader

**(myloader:5288): CRITICAL **: a directory needs to be specified, see --help

 

若出现如下报错, 可能是步骤b有问题.

[email protected]: ~$mydumper

mydumper: errorwhile loading shared libraries: libmysqlclient.so.18: cannot open shared objectfile: No such file or directory

 

下面是演示用到的数据库数据表的信息:

([email protected])[(none)]> SELECT table_schema, table_name, engine FROM information_schema.tables WHERE (engine = ‘InnoDB‘ OR engine = ‘MyISAM‘) AND table_schema NOT IN(‘mysql‘, ‘performance_schema‘ ,‘information_schema‘);

+--------------+------------+--------+

| table_schema |table_name | engine |

+--------------+------------+--------+

| product      | pr1        | MyISAM |

| product      | pr2        | MyISAM |

| product      | pr3        | InnoDB |

| stage        | st1        | InnoDB |

| stage        | st2        | InnoDB |

| test         | tb1        | InnoDB |

| test         | tb2        |InnoDB |

+--------------+------------+--------+

7 rows in set(0.01 sec)

 

 

mydumper的选项也不少, 按照分析mysqldump一样, 将其分成若干组, 看看重点选项的含义.

 

Connection Options组

 

该组选项指明了如何连接数据库.

-h, --host      The host to connect to

-u, --user      Username with privileges to run the dump

-p,--password  User password

-P, --port       TCP/IPport to connect to

-S, --socket     domainsocket file to use for connection

 

Debug Options 组

 

改组指明了日志放在哪里, 以及日志的级别.

-L,--logfile   Log file name to use, by defaultstdout is used

-v,--verbose  Verbosity of output, 0 =silent, 1 = errors, 2 = warnings, 3 = info, default 2

 

Filtering Options组

 

改组指明了备份哪些数据库对象, 以及对备份文件做什么附加处理(压缩, 分割等).

-B,--database        Database to dump

-T,--tables-list        Comma delimitedtable list to dump (does not exclude regex option)

-o,--outputdir        Directory to outputfiles to

-s,--statement-size    Attempted size ofINSERT statement in bytes, default 1000000

-r, --rows            Try to split tables into chunks ofthis many rows. This option turns off --chunk-filesize

-F,--chunk-filesize      Split tables into chunks of this output filesize. This value is in MB

-c,--compress         Compress output files

-e,--build-empty-files   Build dump files even if no data availablefrom table

-x, --regex             Regular expression for ‘db.table‘matching

-m,--no-schemas       Do not dump tableschemas with the data

-d,--no-data           Do not dump tabledata

-G,--triggers           Dump triggers

-E, --events            Dump events

-R, --routines           Dump stored procedures and functions

 

Transactional Options 组

该组主要涉及到备份时如何加锁, 下面使用该命令行进行测试mydumper --regex ‘^(?!(mysql))‘--threads=1 [Option], 同时结合general log, 看mydumper是如何工作的.

 

1. 先看不加选项时, 是什么情况.

Master线程, 获取GLOBAL READ LOCK, 开启一致性读事物, 得到二进制日志的坐标.

1587512Query   FLUSH TABLES WITH READ LOCK

1587512Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

1587512Query   SHOW MASTER STATUS

 

Dump线程, 设置事物隔离级别为REPEATABLE READ, 开启一致性读事物进行非事物数据表的备份.

1587513Query   SET SESSION TRANSACTION ISOLATIONLEVEL REPEATABLE READ

1587513Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

 

1587513Query   SELECT /*!40001 SQL_NO_CACHE */ *FROM `product`.`pr1`

1587513Query   SELECT /*!40001 SQL_NO_CACHE */ *FROM `product`.`pr2`

 

Master线程, 待Dump线程备份完非事物数据表后, 释放锁.

1587512Query   UNLOCK TABLES /* FTWRL */

 

Dump线程, 继续其它事物数据表的备份.

 

2. -k,--no-locks  Do not execute the temporaryshared read lock.  WARNING: This willcause inconsistent backups

使用该选项时, mydumper会有如下类似提示:

**(mydumper:4095): WARNING **: Executing in no-locks mode, snapshot will notbeconsistent

 

其主要作用过程如下:

Master线程, 开启一致性读事物, 得到二进制日志的坐标.

1586766Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

1586766Query   SHOW MASTER STATUS

 

Dump线程, 设置事物隔离级别为REPEATABLE READ, 开启一致性读事物进行数据表的备份.

1586767Query   SET SESSION TRANSACTION ISOLATIONLEVEL REPEATABLE READ

1586767Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

 

该过程由于未执行FLUSH TABLES WITH READ LOCK, 得到的二进制日志坐标可能不准确; (多个)线程开启一致性读事物时, 数据表可能会有变动, 这两点会造成备份数据不一致.                                           

 

3.--less-locking  Minimize locking time onInnoDB tables.

Master线程, 获取GLOBAL READ LOCK, 开启一致性读事物, 得到二进制日志的坐标.

1588054Query   FLUSH TABLES WITH READ LOCK

1588054Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

1588054Query   SHOW MASTER STATUS

 

Dump2线程, 设置事物隔离级别为REPEATABLE READ, 开启一致性读事物.

1588056Query   SET SESSION TRANSACTION ISOLATIONLEVEL REPEATABLE READ

1588056Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

 

Dump1线程, 锁定非事物数据表.

1588055 Query   LOCK TABLES `product`.`pr1` READ LOCAL,`product`.`pr2` READ LOCAL

 

Master线程, 释放锁.

1588054Query   UNLOCK TABLES /* FTWRL */

 

Dump1线程, 备份非事物数据表.

1588055Query   SELECT /*!40001 SQL_NO_CACHE */ *FROM `product`.`pr1`

1588055Query   SELECT /*!40001 SQL_NO_CACHE */ *FROM `product`.`pr2`

 

Dump1线程, 备份完成后, 释放锁.

1588055Query   UNLOCK TABLES /* Non Innodb */

 

Dump2线程, 继续其它事物数据表的备份.

 

4.--use-savepoints  Use savepoints toreduce metadata locking issues, needs SUPER privilege

该选项含义是, 尽快释放元数据锁, 其它过程和1相同.

1601611 Query         SAVEPOINT mydumper

1601611 Query         ROLLBACK TO SAVEPOINT mydumper

 

5.--lock-all-tables  Use LOCK TABLE forall, instead of FTWRL

Master线程, 获取有那些数据库和数据库表, 然后把需要备份的数据表加锁, 开启一致性读事物, 再后得到二进制日志的坐标.

1586979Query   SELECT TABLE_SCHEMA, TABLE_NAMEFROM information_schema.TABLES WHERE TABLE_TYPE =‘BASE TABLE‘ AND TABLE_SCHEMANOT IN (‘information_schema‘, ‘performance_schema‘, ‘data_dictionary‘) AND NOT(TABLE_SCHEMA = ‘mysql‘ AND (TABLE_NAME = ‘slow_log‘ OR TABLE_NAME =‘general_log‘))

1586979Query   LOCK TABLE `product`.`pr1` READ,`product`.`pr2` READ, `product`.`pr3` READ, `stage`.`st1` READ, `stage`.`st2`READ, `test`.`tb1` READ, `test`.`tb2` READ

1586979Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

1586979Query   SHOW MASTER STATUS

 

Dump线程, 设置事物隔离级别为REPEATABLE READ, 开启一致性读事物进行非事物数据表的备份.

1586980Query   SET SESSION TRANSACTION ISOLATIONLEVEL REPEATABLE READ

1586980Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

 

1586980Query   SELECT /*!40001 SQL_NO_CACHE */ *FROM `product`.`pr1`

1586980Query   SELECT /*!40001 SQL_NO_CACHE */ *FROM `product`.`pr2`

 

Master线程, 待Dump线程备份完非事物数据表后, 释放锁.

1586979Query   UNLOCK TABLES /* FTWRL */

 

Dump线程, 继续其它事物数据表的备份.

 

此种加锁方式, 若数据库数据表比较多时, 加锁效率不高.

 

6.--trx-consistency-only  Transactionalconsistency only 

使用该选项时, mydumper会有如下类似提示:

**(mydumper:2573): WARNING **: Using trx_consistency_only, binlog coordinateswill not be accurate if you are writing to non transactional tables

 

Master线程, 获取GLOBAL READ LOCK, 开启一致性读事物, 得到二进制日志的坐标.

1588315Query   FLUSH TABLES WITH READ LOCK

1588315Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

1588315Query   SHOW MASTER STATUS

 

Dump线程, 设置事物隔离级别为REPEATABLE READ, 开启一致性读事物.

1588316Query   SET SESSION TRANSACTION ISOLATIONLEVEL REPEATABLE READ

1588316Query   START TRANSACTION /*!40108 WITHCONSISTENT SNAPSHOT */

 

Master线程, 释放锁.

1588315Query   UNLOCK TABLES /* trx-only */

 

Dump线程, 备份数据表.

 

此方式, 从加锁到释放锁, 时间最短, 效率最高.

 

经上面的分析, 可得到加锁过程影响大小顺序如下:

--lock-all-tables> 不加该组选项 = --use-savepoints >--less-locking > --trx-consistency-only > --no-locks

 

Performance Options 组

该组指定了线程数量, 和如何处理长查询.

-t,--threads          Number of threads touse, default 4

 

-l,--long-query-guard  Set long query timerin seconds, default 60

-K,--kill-long-queries   Kill long runningqueries (instead of aborting)

 

 

参数了解完了, 看两个实际工作中例子.

1. 备份除数据库mysql之外的其它数据库.

[email protected]:~/dbbackup$ mydumper --outputdir=20170826 --compress --build-empty-files--regex ‘^(?!(mysql))‘ --triggers --events --routines --logfile=error.txt--use-savepoints --trx-consistency-only --threads=4 --verbose=3

 

2. 备份全部数据库.

[email protected]:~/dbbackup$ mydumper --outputdir=20170826 --compress --build-empty-files--triggers --events --routines --long-query-guard=60 --kill-long-queries--logfile=error.txt --use-savepoints --trx-consistency-only --threads=4--verbose=3

 

 

经过选项分析和实践过程, 总结下mydumper的特点:

1. 多线程备份, 可指定线程数量, 其也是速度优于mysqldump的关键.

 

2. 对于备份数据一致性方面考虑较多, 主要体现在非事物数据表的备份上.

 

3. 分析选项时, 没有指定字符集的, 查看general log后, 发现是这样处理的/*!40101 SET NAMES binary*/, 即省去了转换字符集的开销.

 

4. 提供了如何应对长查询的选项.

 

 

myloader并没有太多需要说明的, 看下选项解释, 实践下即可.

 

 

mydumper在备份时, 效率有了很大提升, 但其终究还是将数据转化为SQL语句, 即常说的逻辑备份.

 

若能直接备份数据文件, 效率是否更高, 又是否有这样的方法呢... 那就是XtraBackup了, 稍后会在"MySQL Database BackupMethods Season 3 – XtraBackup"中介绍.

本文出自 “感动自己” 博客，请务必保留此出处http://coveringindex.blog.51cto.com/12533373/1959683