本篇内容主要讲解“MySQL慢SQL采集方案分析”,感兴趣的朋友不妨来看看。本文介绍的方法操作简单快捷,实用性强。下面就让小编来带大家学习“MySQL慢SQL采集方案分析”吧!
作为一名MySQL DBA,首要的任务是维持数据库的可用性和稳定性,在生产中,有时候一条慢SQL会拖垮整个系统的响应和体验,所以慢SQL治理至关重要。
首先我们采用pt-query-digest
来进行慢sql采集和分析
mysql> show global variables where variable_name in ('slow_query_log','long_query_time','slow_query_log_file','log_queries_not_using_indexes','log_throttle_queries_not_using_indexes','min_examined_row_limit','log_slow_admin_statements','log_slow_slave_statements','log_output'); +----------------------------------------+----------------------------+ | Variable_name | Value | +----------------------------------------+----------------------------+ | log_output | FILE | | log_queries_not_using_indexes | ON | | log_slow_admin_statements | OFF | | log_slow_slave_statements | ON | | log_throttle_queries_not_using_indexes | 100 | | long_query_time | 0.500000 | | min_examined_row_limit | 0 | | slow_query_log | ON | | slow_query_log_file | slowquery_201908161156.log | +----------------------------------------+----------------------------+ 9 rows in set (0.01 sec)
该mysql版本为percona5.7.21版本,控制slow log的参数主要为这几个:
log_output: 控制慢sql是记录在文件还是记录在table,FILE|TABLE可选择
log_queries_not_using_indexes:控制没有使用索引的sql也将被记录到慢查询日志中;
log_slow_admin_statements:管理语句执行时间大于阈值也将写入到慢查询日志中,管理语句包括alter table, check table等;
log_slow_slave_statements:从库应用binlog,如果binlog格式是statement,执行时间超过阈值时,将写入从库的慢查询日志, 对于ROW格式binlog,不管执行时间有没有超过阈值,都不会写入到从库的慢查询日志;
log_throttle_queries_not_using_indexes:如果log_queries_not_using_indexes打开,没有使用索引的sql将会写入到慢查询日志中,该参数将限制每分钟写入的sql数量;
long_query_time:慢查询阈值,单位秒,执行时间超过这个值的将被记录为慢查询日志中;
min_examined_row_limit:对于查询扫描行数小于此参数的sql,将不会记录到慢查询日志中;
slow_query_log:控制是否打开慢查询;
slow_query_log_file:慢查询日志记录文件;
这些参数我们目前主要使用log_output
,slow_query_log
,slow_query_log_file
,min_examined_row_limit
,log_throttle_queries_not_using_indexes
这几个
mysql> set global log_output = 'FILE'; Query OK, 0 rows affected (0.00 sec) mysql> set global slow_query_log = 1; Query OK, 0 rows affected (0.00 sec) mysql> set global min_examined_row_limit = 0; Query OK, 0 rows affected (0.00 sec) mysql> set global log_throttle_queries_not_using_indexes = 1; Query OK, 0 rows affected (0.00 sec) mysql> set global slow_query_log_file = 'slowquery_201908161203.log'; Query OK, 0 rows affected (0.00 sec)
设置参数后开始安装pt工具
我们这里直接安装个全家桶(percona-toolkit
)吧,首先安装依赖包:
yum install -y perl perl-IO-Socket-SSL perl-DBD-MySQL perl-Time-HiRes perl-Digest-MD5 perl-ExtUtils-MakeMaker
然后下载percona-toolkit:
https://www.percona.com/downloads/percona-toolkit/LATEST/
这里是cenos7环境,如果是其他环境另行下载
我们直接安装
rpm -ivh percona-toolkit-3.0.13-1.el7.x86_64.rpm
pt-query-digest [OPTIONS] [FILES] [DSN] --create-review-table 当使用--review参数把分析结果输出到表中时,如果没有表就自动创建。 --create-history-table 当使用--history参数把分析结果输出到表中时,如果没有表就自动创建。 --filter 对输入的慢查询按指定的字符串进行匹配过滤后再进行分析 --limit 限制输出结果百分比或数量,默认值是20,即将最慢的20条语句输出,如果是50%则按总响应时间占比从大到小排序,输出到总和达到50%位置截止。 --host mysql服务器地址 --user mysql用户名 --password mysql用户密码 --history 将分析结果保存到表中,分析结果比较详细,下次再使用--history时,如果存在相同的语句,且查询所在的时间区间和历史表中的不同,则会记录到数据表中,可以通过查询同一CHECKSUM来比较某类型查询的历史变化。 --review 将分析结果保存到表中,这个分析只是对查询条件进行参数化,一个类型的查询一条记录,比较简单。当下次使用--review时,如果存在相同的语句分析,就不会记录到数据表中。 --output 分析结果输出类型,值可以是report(标准分析报告)、slowlog(Mysql slow log)、json、json-anon,一般使用report,以便于阅读。 --since 从什么时间开始分析,值为字符串,可以是指定的某个”yyyy-mm-dd [hh:mm:ss]”格式的时间点,也可以是简单的一个时间值:s(秒)、h(小时)、m(分钟)、d(天),如12h就表示从12小时前开始统计。 --until 截止时间,配合—since可以分析一段时间内的慢查询。
1、分析慢查询文件:
pt-query-digest slow-query.log > slow-query-report.log
2、分析最近1小时内的查询:
pt-query-digest --since=1h slow-query.log > slow-before_1h.log
3、分析指定时间范围内的查询:
pt-query-digest slow-query.log --since '2019-08-08 10:30:00' --until '2019-08-08 10:35:00' > slow-query-5min.log
4、针对某个用户的慢查询
pt-query-digest --filter '($event->{user} || "") =~ m/^dba/i' slow-query.log > slow-query-fordbauser.log
5、查询所有所有的全表扫描或full join的慢查询
pt-query-digest --filter '(($event->{Full_scan} || "") eq "yes") ||(($event->{Full_join} || "") eq "yes")' slow-query.log> slow_query-full.log
6、把查询保存到t_slowq_review表
pt-query-digest --user=dbms –password=admin --review h=localhost,D=test,t=t_slowq_review --create-review-table slow-query.log
7、把查询保存到t_slowq_details表
pt-query-digest --user=dbms –password=admin --history h=localhost,D=test,t=t_slowq_details --create-history-table slow-query.log
8、分析指含有select语句的慢查询
pt-query-digest --filter '$event->{fingerprint} =~ m/^select/i' slow-query.log> slow-query-justselect.log
9、通过tcpdump抓取mysql的tcp协议数据,然后再分析
tcpdump -s 65535 -x -nn -q -tttt -i any -c 1000 port 3306 > mysql.tcp3306.txt pt-query-digest --type tcpdump mysql.tcp3306.txt> slow-query-3306.log
10、分析binlog
mysqlbinlog mysql-bin.000011 > mysql-bin.000011.sql pt-query-digest --type=binlog mysql-bin.000011.sql > mysql-bin.000011.log
11.分析general log
pt-query-digest --type=genlog general_3306.log > generallog_3306.log
这里采用直接入库的方式,然后会采用脚本去切割,可以使用时间间隔切割,然后删除,尝试过since util,发现文本越大,每次的解析时间都会很长,不太适合生产部署,可以采用切割日志,如果需要保留本地的话,可以拷贝走
pt-query-digest \ --user=user --password=password --port=port \ --review h=ip,D=dbname,t=t_slowq_review \ --history h=ip,D=dbname,t=t_slowq_details \ --no-report --limit=100% --charset=utf8 \ --filter="\$event->{Bytes}=length(\$event->{arg}) and \$event->{instanceid}=15 and \$event->{hostname}='idc-mysql18' and \$event->{client}=\$event->{ip}" \ slow-query.log
在生产中由于可能存在分库,所以sql的checksum值可能会一致,所以我们加了每个实例的id进去来标识不同的数据库,完全靠checksum值来标记可能混淆
CREATE TABLE `t_slowq_details` ( `instanceid_max` int(11) NOT NULL, `hostname_max` varchar(64) NOT NULL, `client_max` varchar(64) DEFAULT NULL, `user_max` varchar(64) NOT NULL, `db_max` varchar(64) DEFAULT NULL, `checksum` char(32) NOT NULL, `sample` longtext NOT NULL, `ts_min` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, `ts_max` datetime(6) NOT NULL, `ts_cnt` float DEFAULT NULL, `Query_time_sum` float DEFAULT NULL, `Query_time_min` float DEFAULT NULL, `Query_time_max` float DEFAULT NULL, `Query_time_pct_95` float DEFAULT NULL, `Query_time_stddev` float DEFAULT NULL, `Query_time_median` float DEFAULT NULL, `Lock_time_sum` float DEFAULT NULL, `Lock_time_min` float DEFAULT NULL, `Lock_time_max` float DEFAULT NULL, `Lock_time_pct_95` float DEFAULT NULL, `Lock_time_stddev` float DEFAULT NULL, `Lock_time_median` float DEFAULT NULL, `Rows_sent_sum` float DEFAULT NULL, `Rows_sent_min` float DEFAULT NULL, `Rows_sent_max` float DEFAULT NULL, `Rows_sent_pct_95` float DEFAULT NULL, `Rows_sent_stddev` float DEFAULT NULL, `Rows_sent_median` float DEFAULT NULL, `Rows_examined_sum` float DEFAULT NULL, `Rows_examined_min` float DEFAULT NULL, `Rows_examined_max` float DEFAULT NULL, `Rows_examined_pct_95` float DEFAULT NULL, `Rows_examined_stddev` float DEFAULT NULL, `Rows_examined_median` float DEFAULT NULL, `Rows_affected_sum` float DEFAULT NULL, `Rows_affected_min` float DEFAULT NULL, `Rows_affected_max` float DEFAULT NULL, `Rows_affected_pct_95` float DEFAULT NULL, `Rows_affected_stddev` float DEFAULT NULL, `Rows_affected_median` float DEFAULT NULL, `Rows_read_sum` float DEFAULT NULL, `Rows_read_min` float DEFAULT NULL, `Rows_read_max` float DEFAULT NULL, `Rows_read_pct_95` float DEFAULT NULL, `Rows_read_stddev` float DEFAULT NULL, `Rows_read_median` float DEFAULT NULL, `Merge_passes_sum` float DEFAULT NULL, `Merge_passes_min` float DEFAULT NULL, `Merge_passes_max` float DEFAULT NULL, `Merge_passes_pct_95` float DEFAULT NULL, `Merge_passes_stddev` float DEFAULT NULL, `Merge_passes_median` float DEFAULT NULL, `InnoDB_IO_r_ops_min` float DEFAULT NULL, `InnoDB_IO_r_ops_max` float DEFAULT NULL, `InnoDB_IO_r_ops_pct_95` float DEFAULT NULL, `InnoDB_IO_r_ops_stddev` float DEFAULT NULL, `InnoDB_IO_r_ops_median` float DEFAULT NULL, `InnoDB_IO_r_bytes_min` float DEFAULT NULL, `InnoDB_IO_r_bytes_max` float DEFAULT NULL, `InnoDB_IO_r_bytes_pct_95` float DEFAULT NULL, `InnoDB_IO_r_bytes_stddev` float DEFAULT NULL, `InnoDB_IO_r_bytes_median` float DEFAULT NULL, `InnoDB_IO_r_wait_min` float DEFAULT NULL, `InnoDB_IO_r_wait_max` float DEFAULT NULL, `InnoDB_IO_r_wait_pct_95` float DEFAULT NULL, `InnoDB_IO_r_wait_stddev` float DEFAULT NULL, `InnoDB_IO_r_wait_median` float DEFAULT NULL, `InnoDB_rec_lock_wait_min` float DEFAULT NULL, `InnoDB_rec_lock_wait_max` float DEFAULT NULL, `InnoDB_rec_lock_wait_pct_95` float DEFAULT NULL, `InnoDB_rec_lock_wait_stddev` float DEFAULT NULL, `InnoDB_rec_lock_wait_median` float DEFAULT NULL, `InnoDB_queue_wait_min` float DEFAULT NULL, `InnoDB_queue_wait_max` float DEFAULT NULL, `InnoDB_queue_wait_pct_95` float DEFAULT NULL, `InnoDB_queue_wait_stddev` float DEFAULT NULL, `InnoDB_queue_wait_median` float DEFAULT NULL, `InnoDB_pages_distinct_min` float DEFAULT NULL, `InnoDB_pages_distinct_max` float DEFAULT NULL, `InnoDB_pages_distinct_pct_95` float DEFAULT NULL, `InnoDB_pages_distinct_stddev` float DEFAULT NULL, `InnoDB_pages_distinct_median` float DEFAULT NULL, `QC_Hit_cnt` float DEFAULT NULL, `QC_Hit_sum` float DEFAULT NULL, `Full_scan_cnt` float DEFAULT NULL, `Full_scan_sum` float DEFAULT NULL, `Full_join_cnt` float DEFAULT NULL, `Full_join_sum` float DEFAULT NULL, `Tmp_table_cnt` float DEFAULT NULL, `Tmp_table_sum` float DEFAULT NULL, `Tmp_table_on_disk_cnt` float DEFAULT NULL, `Tmp_table_on_disk_sum` float DEFAULT NULL, `Filesort_cnt` float DEFAULT NULL, `Filesort_sum` float DEFAULT NULL, `Filesort_on_disk_cnt` float DEFAULT NULL, `Filesort_on_disk_sum` float DEFAULT NULL, `Bytes_sum` float DEFAULT NULL, `Bytes_min` float DEFAULT NULL, `Bytes_max` float DEFAULT NULL, `Bytes_pct_95` float DEFAULT NULL, `Bytes_stddev` float DEFAULT NULL, `Bytes_median` float DEFAULT NULL, PRIMARY KEY (`instanceid_max`,`checksum`,`ts_min`,`ts_max`) USING BTREE, KEY `idx_hostname_max_ts_min` (`instanceid_max`,`ts_min`) ) ENGINE=InnoDB DEFAULT CHARSET=utf8 /*!50100 PARTITION BY RANGE (UNIX_TIMESTAMP(ts_min)) (PARTITION p201908 VALUES LESS THAN (1567267200) ENGINE = InnoDB, PARTITION p201909 VALUES LESS THAN (1569859200) ENGINE = InnoDB, PARTITION p201910 VALUES LESS THAN (1572537600) ENGINE = InnoDB, PARTITION p201911 VALUES LESS THAN (1575129600) ENGINE = InnoDB, PARTITION p201912 VALUES LESS THAN (1577808000) ENGINE = InnoDB) */ CREATE TABLE `t_slowq_review` ( `checksum` char(32) NOT NULL, `fingerprint` longtext NOT NULL, `sample` longtext NOT NULL, `first_seen` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, `last_seen` datetime(6) DEFAULT NULL, `reviewed_by` varchar(20) DEFAULT NULL, `reviewed_on` datetime(6) DEFAULT NULL, `comments` longtext, `reviewed_status` varchar(24) DEFAULT NULL, PRIMARY KEY (`checksum`) USING BTREE, KEY `idx_last_seen` (`last_seen`) USING BTREE ) ENGINE=InnoDB DEFAULT CHARSET=utf8
这个脚本只是提供一个思路,逻辑就是,会自动去切割slowlog,然后解析上传,切割后解析玩会删除前一个日志,如果需要保留本地日志可以把代码里面的删除改成拷贝走,其中我提出的instance_id是我们觉得生产中只有checksum没法和系统进行结合和日常使用,各位可以参考使用,并进行改造,其中有啥问题可以留言????
#!/usr/bin/env python # -*- coding:utf-8 -*- # create_time: 2019-08-08 11:20 import sys import os import socket import time from multiprocessing import Process import subprocess import pymysql as MySQLdb DBMS_HOST = "" # 存储数据库的ip DBMS_NAME = "" # 存储数据库的dbname DBMS_PORT = # 存储数据库的端口 DBMS_PWD = "" # 存储数据库的账号密码 DBMS_USER = "" # 存储数据库的账号 DBA_HOST = "" # 被采集实例的ip,一般默认为 127.0.0.1 DBA_USER = "" # 被采集实例的 用户 DBA_PASSWORD = "" # 被采集实例 用户密码 # 一般数据实例上建立一个 dba管理账号来使用 class MySQLConnection(object): def __init__(self, host, port, user, passwd, db=None, charset="utf8"): self.host = host self.port = port self.user = user self.passwd = passwd self.db = db self.charset = charset self.error = "" def Success(self): return self.error == "" def Connect(self): self.error = "" try: self.conn = MySQLdb.connect(host=self.host, port=self.port, user=self.user, passwd=self.passwd, charset=self.charset) if self.db: self.conn.select_db(self.db) self.cursor = self.conn.cursor() except Exception as e: self.error = str(e) def Close(self): self.error = "" try: self.cursor.close() self.conn.close() except Exception as e: self.error = str(e) def Fileds(self): res = [] try: for r in self.cursor.description: res.append(r[0]) except: pass return res def Query(self, sql, params=None, rownum=[0], fetch=0): self.error = "" data = () try: if params: rownum[0] = self.cursor.execute(sql, params) else: rownum[0] = self.cursor.execute(sql) if rownum[0] > 0: if fetch == 0: data = self.cursor.fetchall() elif fetch == 1: data = self.cursor.fetchone() except Exception as e: self.error = str(e) return data def Execute(self, sql, params=None): self.error = "" try: if params: self.cursor.execute(sql, params) else: self.cursor.execute(sql) self.conn.commit() except Exception as e: self.error = str(e) return False return True def RunCommand(c): p = subprocess.Popen(c, stdout=subprocess.PIPE, stderr = subprocess.PIPE, shell = True) (stdout, stderr) = p.communicate() return (stdout.decode("utf-8"), stderr.decode("utf-8")) def log_out(instance_id, start = time.time(), error = None): """ 这里是用来记录慢sql解析情况的,会上报相关状态 """ dbms_con = MySQLConnection(DBMS_HOST, DBMS_PORT, DBMS_USER, DBMS_PWD, DBMS_NAME) dbms_con.Connect() if dbms_con.error: return start_time = time.strftime('%Y-%m-%d %T', time.localtime(start)) dbms_con.Execute("replace into t_slowq_message(instance_id, error, start_time) values(%s, %s, %s)", (instance_id, error, start_time)) dbms_con.Close() def GetSlowlogfile(instanceid,port): """ slowlog = datadir+slowlog :param port: :return: """ local_conn = MySQLConnection(host=DBA_HOST, port=port, user=DBA_USER, passwd=DBA_PASSWORD) local_conn.Connect() if not local_conn.error: data = local_conn.Query("show global variables where variable_name in ('long_query_time', 'slow_query_log_file','datadir')") if data: res = { data[0][0]: data[0][1], data[1][0]: data[1][1], data[2][0]: data[2][1], } local_conn.Close() return res local_conn.Close() error = "get slowlog_file paremeters failed" log_out(instance_id=instanceid, error=error) return None else: log_out(instance_id=instanceid, error=str(local_conn.error)) local_conn.Close() return None def SwitchSlowlogfile(port): local_conn = MySQLConnection(host=DBA_HOST, port=port, user=DBA_USER, passwd=DBA_PASSWORD) local_conn.Connect() if local_conn.error: return False tmp_log = "slowquery_%s.log" % (time.strftime('%Y%m%d%H%M', time.localtime(time.time())),) res = local_conn.Execute("set global slow_query_log_file = %s", tmp_log) return True def CollectSlowlog(instanceid,port,hostname): res = GetSlowlogfile(instanceid,port) if res: slowlog_time = eval(res["long_query_time"]) slowlog_file = res["slow_query_log_file"] data_dir = res['datadir'] slow_log = data_dir+slowlog_file if SwitchSlowlogfile(port): try: pt_query_digest = RunCommand("whereis pt-query-digest")[0].split()[1] except: pt_query_digest = "/usr/bin/pt-query-digest" cmd = "%s --user=%s --password=%s --port=%s --review h=%s,D=%s,t=%s --history h=%s,D=%s,t=%s --no-report --limit=100%% --charset=utf8 " \ "--filter=\"\\$event->{Bytes} = length(\\$event->{arg}) and \\$event->{instanceid}=%s and \\$event->{hostname}='%s'" \ " and \$event->{client}=\$event->{ip}\" %s " % ( pt_query_digest, DBMS_USER, DBMS_PWD, DBMS_PORT, DBMS_HOST, DBMS_NAME, "t_slowq_review", DBMS_HOST, DBMS_NAME, "t_slowq_details", instanceid, hostname,slow_log) out, error = RunCommand(cmd) if not error: os.remove(slow_log) log_out(instance_id=instanceid) else: error = "parse slowlog_file failed" log_out(instance_id=instanceid,error=error) else: error = "switch slowlog_file failed" log_out(instance_id=instanceid, error=error) def Main(): """ 该部分可以自己进行改造,我这边的逻辑是我有一张记录实例的元数据库,来查询当前主机上面的实例,然后获取实例id和port,各位可以自行改造 """ try: hostname = socket.gethostname() dbms_con = MySQLConnection(DBMS_HOST, DBMS_PORT, DBMS_USER, DBMS_PWD, DBMS_NAME) dbms_con.Connect() if dbms_con.error: return sql1 = "SELECT a.instance_id,a.port FROM t_info_instance AS a JOIN t_info_machine AS b ON a.m_id = b.m_id WHERE b.hostname = %s" res1 = dbms_con.Query(sql1, (hostname,)) if res1 != None and len(res1) > 0: for instance in res1: instance_id = instance[0] port = instance[1] p = Process(target=CollectSlowlog,args=(instance_id,port,hostname)) p.start() dbms_con.Close() except Exception as e: print (e) if __name__ == "__main__": Main()
另一张状态表表结构:
CREATE TABLE `t_slowq_message` ( `instance_id` int(11) NOT NULL, `error` text, `start_time` datetime NOT NULL, `end_time` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, PRIMARY KEY (`instance_id`) ) ENGINE=InnoDB DEFAULT CHARSET=utf8
到此,相信大家对“MySQL慢SQL采集方案分析”有了更深的了解,不妨来实际操作一番吧!这里是亿速云网站,更多相关内容可以进入相关频道进行查询,关注我们,继续学习!
免责声明:本站发布的内容(图片、视频和文字)以原创、转载和分享为主,文章观点不代表本网站立场,如果涉及侵权请联系站长邮箱:is@yisu.com进行举报,并提供相关证据,一经查实,将立刻删除涉嫌侵权内容。