使用Anemometer分析MySQL慢查询记录

数据库管理员一般是用percona的toolkit工具来分析MySQL慢查询记录，但是不够直观。

下面介绍一款比较直观的工具来统计分析MySQL慢查询记录anemometer。

在使用之前需要安装percona的toolkit工具，anemometer提供web界面。

anemometer介绍参见：https://github.com/box/Anemometer/wiki

1. 安装

 

# cd /data/www/my.ttlsa.com # git clone https://github.com/box/Anemometer.git anemometer # cd anemometer

1 2 3

# cd /data/www/my.ttlsa.com # git clone https://github.com/box/Anemometer.git anemometer # cd anemometer

 

2. 创建表和用户名

 

mysql < install.sql mysql -e "grant ALL ON slow_query_log.* to 'anemometer'@'localhost' IDENTIFIED BY 'my.ttlsa.com';" mysql -e "grant SELECT ON *.* to 'anemometer'@'localhost' IDENTIFIED BY 'my.ttlsa.com';"

1 2 3

mysql < install . sql mysql - e "grant ALL ON slow_query_log.* to 'anemometer'@'localhost' IDENTIFIED BY 'my.ttlsa.com';" mysql - e "grant SELECT ON *.* to 'anemometer'@'localhost' IDENTIFIED BY 'my.ttlsa.com';"

 

3. 分析MySQL慢日志

 

# pt-query-digest --user=anemometer --password=superSecurePass \ --review D=slow_query_log,t=global_query_review \ --review-history D=slow_query_log,t=global_query_review_history \ --no-report --limit=0% --filter=" \$event->{Bytes} = length(\$event->{arg}) and \$event->{hostname}=\"$HOSTNAME\"" /data/log/mysql/slow.log

1 2 3 4

# pt-query-digest --user=anemometer --password=superSecurePass \                    -- review D = slow_query_log , t = global_query _review \                    -- review - history D = slow_query_log , t = global_query_review _history \                    -- no - report -- limit = 0 % -- filter = " \$event->{Bytes} = length(\$event->{arg}) and \$event->{hostname}=\"$HOSTNAME\""    / data / log / mysql / slow .log

 

4. 配置anemometer

 

# cp conf/sample.config.inc. php conf/config.inc.php # vi conf/config.inc.php $conf['datasources']['localhost'] = array( 'host' => 'localhost', 'port' => 3306, 'db' => 'slow_query_log', 'user' => 'anemometer', 'password' => 'my.ttlsa.com', 'tables' => array( 'global_query_review' => 'fact', 'global_query_review_history' => 'dimension' ), 'source_type' => 'slow_query_log' ); $conf['plugins'] = array( 'visual_explain' => '/usr/bin/pt-visual-explain', 'query_advisor' => '/usr/bin/pt-query-advisor', #... other lines $conn['user'] = 'anemometer'; $conn['password'] = 'my.ttlsa.com'; return $conn; },

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

# cp conf/sample.config.inc.php conf/config.inc.php # vi conf/config.inc.php $conf [ 'datasources' ] [ 'localhost' ] = array (          'host'    = > 'localhost' ,          'port'    = > 3306 ,          'db'      = > 'slow_query_log' ,          'user'    = > 'anemometer' ,          'password' = > 'my.ttlsa.com' ,          'tables' = > array (                  'global_query_review' = > 'fact' ,                  'global_query_review_history' = > 'dimension'          ) ,          'source_type' = > 'slow_query_log' ) ;   $conf [ 'plugins' ] = array (   'visual_explain' = > '/usr/bin/pt-visual-explain' , 'query_advisor' = > '/usr/bin/pt-query-advisor' ,   #... other lines   $conn [ 'user' ] = 'anemometer' ; $conn [ 'password' ] = 'my.ttlsa.com' ;   return $conn ; } ,

 

5. 自动化处理

 

# vi /etc/logrotate.d/mysql postrotate pt-query-digest --user=anemometer --password=superSecurePass \ --review D=slow_query_log,t=global_query_review \ --review-history D=slow_query_log,t=global_query_review_history \ --no-report --limit=0% --filter=" \$event->{Bytes} = length(\$event->{arg}) and \$event->{hostname}=\"$HOSTNAME\"" /data/log/mysql/slow.log.1 endscript

1 2 3 4 5 6 7

# vi /etc/logrotate.d/mysql postrotate pt - query - digest -- user = anemometer -- password = superSecurePass \                    -- review D = slow_query_log , t = global_query _review \                    -- review - history D = slow_query_log , t = global_query_review _history \                    -- no - report -- limit = 0 % -- filter = " \$event->{Bytes} = length(\$event->{arg}) and \$event->{hostname}=\"$HOSTNAME\"" / data / log / mysql / slow .log . 1 endscript

这样就可以通过my.ttlsa.com/anemometer来访问查看慢查询了。

参考文档：

https://www.box.com/blog/optimizing-mysql-performance-at-scale-with-anemometer-2/

http://isadba.com/?p=655