mysql日志介绍
错误日志(log_error)
--记录启动关闭日常运行过程中,状态信息,警告,错误
默认就是开启的: /数据路径下/hostname.err 手工设定: Master [(none)]>select @@log_error; vim /etc/my.cnf log_error=/var/log/MySQL.log log_timestamps=system 重启生效 show variables like 'log_error';
默认位置：
DATADIR/hostname.err
配置方式
vim .etc/my.cnf
log_error=/data/mysql/data/mysql.log
查看：
mysql> select @@log_error;
+----------------------+
| @@log_error |
+----------------------+
| /data/3307/mysql.log |
+----------------------+
1 row in set (0.00 sec)
binlog(binary logs):二进制日志 *****
(1)备份恢复必须依赖二进制日志
(2)主从环境必须依赖二进制日志
binlog配置 (5.7必须加server_id)
注意：MySQL默认是没有开启二进制日志的。 基础参数查看: 开关: [(none)]>select @@log_bin; 日志路径及名字 [(none)]>select @@log_bin_basename; 服务ID号: [(none)]>select @@server_id; 二进制日志格式: [(none)]>select @@binlog_format; 双一标准之二: [(none)]>select @@sync_binlog;
log_error=
/data/mysql/data/mysql.log
log_bin=
/data/binlog/3306/mysql-bin #/data/binlog/3306/ ：目录 mysql-bin：文件名
[root@localhost 3306]# pwd
/data/binlog/3306
[root@localhost 3306]# ll
总用量 8
-rw-r----- 1 mysql mysql 154 9月 4 13:51 mysql-bin.000001
-rw-r----- 1 mysql mysql 35 9月 4 13:51 mysql-bin.index
查看配置
mysql> show variables like '%log_bin%';
+---------------------------------+-----------------------------------+
| Variable_name | Value |
+---------------------------------+-----------------------------------+
| log_bin | ON |
| log_bin_basename | /data/binlog/3306/mysql-bin |
| log_bin_index | /data/binlog/3306/mysql-bin.index |
| log_bin_trust_function_creators | OFF |
| log_bin_use_v1_row_events | OFF |
| sql_log_bin | ON |
+---------------------------------+-----------------------------------+
6 rows in set (0.00 sec)
binlog记录了什么?
-----binlog是SQL层的功能。记录的是变更SQL语句，不记录查询语句。
记录SQL语句种类
DDL ：原封不动的记录当前DDL(statement语句方式)。
DCL ：原封不动的记录当前DCL(statement语句方式)。
DML ：只记录已经提交的事务DML
DML三种记录方式binlog_format（binlog的记录格式）参数影响
（1）statement（5.6默认）SBR(statement based replication) ：语句模式原封不动的记录当前DML。
（2）ROW(5.7 默认值)RBR(ROW based replication) ：记录数据行的变化(用户看不懂，需要工具分析)
（3）mixed（混合）MBR(mixed based replication)模式 ：以上两种模式的混合
mysql> show variables like 'bnlog_format';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| binlog_format | ROW |
+---------------+-------+
1 row in set (0.00 sec)
面试题
SBR与RBR模式的对比
STATEMENT：可读性较高，日志量少，但是不够严谨
ROW ：可读性很低，日志量大，足够严谨
update t1 set xxx=xxx where id>1000 ? -->一共500w行，row模式怎么记录的日志
为什么row模式严谨？
id name intime
insert into t1 values(1,'zs',now())
我们建议使用：row记录模式
event（事件）是什么?
事件的简介
二进制日志的最小记录单元
对于DDL,DCL,一个语句就是一个event
对于DML语句来讲:只记录已提交的事务。
例如以下列子,就被分为了4个event
begin; 120 - 340
DML1 340 - 460
DML2 460 - 550
commit; 550 - 760
event的组成
三部分构成:
(1) 事件的开始标识
(2) 事件内容
(3) 事件的结束标识
Position:
开始标识: at 194
结束标识: end_log_pos 254
194? 254?
某个事件在binlog中的相对位置号
位置号的作用是什么？
为了方便我们截取事件
查看一共多少个binlog
Master [(none)]>show binary logs;
+------------------+-----------+
| Log_name | File_size |
+------------------+-----------+
| mysql-bin.000001 | 154 |
+------------------+-----------+
1 row in set (0.01 sec)
Master [(none)]>flush logs;
Query OK, 0 rows affected (0.03 sec)
Master [(none)]>flush logs;
Query OK, 0 rows affected (0.01 sec)
Master [(none)]>show binary logs;
+------------------+-----------+
| Log_name | File_size |
+------------------+-----------+
| mysql-bin.000001 | 201 |
| mysql-bin.000002 | 201 |
| mysql-bin.000003 | 154 |
+------------------+-----------+
3 rows in set (0.00 sec)
Master [(none)]>
通过执行flush logs命令之后，再次查询binary log信息，发现已经使用了一个新的bin log文件了。
flush logs命令的作用就是关闭当前使用的binary log，然后打开一个新的binary log文件，binlog文件的序号递增加1
查看mysql正在使用的日志文件
Master [(none)]>show master status;
+------------------+----------+--------------+------------------+-------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000003 | 154 | | | |
+------------------+----------+--------------+------------------+-------------------+
Master [(none)]>
file：当前MySQL正在使用的文件名
Position：最后一个事件的结束位置号
日志内容查看
event查看
Master [binlog]>show binlog events in 'mysql-bin.000003';
+------------------+-----+----------------+-----------+-------------+----------------------------------------+
| Log_name | Pos | Event_type | Server_id | End_log_pos | Info |
+------------------+-----+----------------+-----------+-------------+----------------------------------------+
| mysql-bin.000003 | 4 | Format_desc | 6 | 123 | Server ver: 5.7.20-log, Binlog ver: 4 |
| mysql-bin.000003 | 123 | Previous_gtids | 6 | 154 | |
| mysql-bin.000003 | 154 | Anonymous_Gtid | 6 | 219 | SET @@SESSION.GTID_NEXT= 'ANONYMOUS' |
| mysql-bin.000003 | 219 | Query | 6 | 319 | create database binlog |
| mysql-bin.000003 | 319 | Anonymous_Gtid | 6 | 384 | SET @@SESSION.GTID_NEXT= 'ANONYMOUS' |
| mysql-bin.000003 | 384 | Query | 6 | 486 | use `binlog`; create table t1 (id int) |
+------------------+-----+----------------+-----------+-------------+----------------------------------------+
Log_name：binlog文件名
Pos：开始的position *****
Event_type：事件类型
Format_desc：格式描述，每一个日志文件的第一个事件，多用户没有意义，MySQL识别binlog必要信息
Server_id：mysql服务号标识
End_log_pos：事件的结束位置号 *****
Info：事件内容*****
补充:
SHOW BINLOG EVENTS
[IN 'log_name']
[FROM pos]
[LIMIT [offset,] row_count]
[root@db01 binlog]# mysql -e "show binlog events in 'mysql-bin.000004'" |grep drop
binlog文件内容详细查看
mysqlbinlog /data/mysql/mysql-bin.000006
mysqlbinlog --base64-output=decode-rows -vvv /data/binlog/mysql-bin.000003
mysqlbinlog -d binlog /data/binlog/mysql-bin.000003
[root@db01 binlog]# mysqlbinlog --start-datetime='2019-05-06 17:00:00' --stop-datetime='2019-05-06 17:01:00' /data/binlog/mysql-bin.000004
基于Position号进行日志截取
核心就是找截取的起点和终点
--start-position=321
--stop-position=513
mysqlbinlog --start-position=219 --stop-position=1347 /data/binlog/mysql-bin.000003 >/tmp/bin.sql
案例: 使用binlog日志进行数据恢复
模拟:
1.
[(none)]>create database binlog charset utf8;
2.
[(none)]>use binlog;
[binlog]>create table t1(id int);
3.
[binlog]>insert into t1 values(1);
[binlog]>commit;
[binlog]>insert into t1 values(2);
[binlog]>commit;
[binlog]>insert into t1 values(3);
[binlog]>commit;
4.
[binlog]>drop database binlog;
恢复:
[(none)]>show master status ;
[(none)]>show binlog events in 'mysql-bin.000004';
[root@db01 binlog]# mysqlbinlog --start-position=1227 --stop-position=2342 /data/binlog/mysql-bin.000004 >/tmp/bin.sql
[(none)]>set sql_Log_bin=0;#设为0后，在Master数据库上执行的语句临时都不记录binlog，退出窗口可以恢复
[(none)]>source /tmp/bin.sql
-----上述通过pos位置恢复 需要关闭gitd----
经过测试发现 开起来gitd，通过pos位置无法恢复数据
mysqlbinlog --start-position=219 --stop-position=1012 /data/binlog/3306/mysql-bin.000002 >/tmp/bin.sql
binlog日志的GTID新特性
GTID 介绍
5.6 版本新加的特性,5.7中做了加强
5.6 中不开启,没有这个功能.
5.7 中的GTID,即使不开也会有自动生成
SET @@SESSION.GTID_NEXT= 'ANONYMOUS'
GTID(Global Transaction ID)
是对于一个已提交事务的编号，并且是一个全局唯一的编号。
它的官方定义如下：GTID = source_id ：transaction_id
7E11FA47-31CA-19E1-9E56-C43AA21293967:29
重要参数介绍：
vim /etc/my.cnf
gtid-mode=on
enforce-gtid-consistency=true
systemctl restart mysqld
Master [(none)]>create database gtid charset utf8;
Query OK, 1 row affected (0.01 sec)
Master [(none)]>show master status ;
+------------------+----------+--------------+------------------+----------------------------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+----------------------------------------+
| mysql-bin.000004 | 326 | | | dff98809-55c3-11e9-a58b-000c2928f5dd:1 |
+------------------+----------+--------------+------------------+----------------------------------------+
1 row in set (0.00 sec)
Master [(none)]>use gtid
Database changed
Master [gtid]>create table t1 (id int);
Query OK, 0 rows affected (0.01 sec)
Master [gtid]>show master status ;
+------------------+----------+--------------+------------------+------------------------------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+------------------------------------------+
| mysql-bin.000004 | 489 | | | dff98809-55c3-11e9-a58b-000c2928f5dd:1-2 |
+------------------+----------+--------------+------------------+------------------------------------------+
1 row in set (0.00 sec)
Master [gtid]>create table t2 (id int);
Query OK, 0 rows affected (0.01 sec)
Master [gtid]>create table t3 (id int);
Query OK, 0 rows affected (0.02 sec)
Master [gtid]>show master status ;
+------------------+----------+--------------+------------------+------------------------------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+------------------------------------------+
| mysql-bin.000004 | 815 | | | dff98809-55c3-11e9-a58b-000c2928f5dd:1-4 |
+------------------+----------+--------------+------------------+------------------------------------------+
1 row in set (0.00 sec)
Master [gtid]>begin;
Query OK, 0 rows affected (0.00 sec)
Master [gtid]>insert into t1 values(1);
Query OK, 1 row affected (0.00 sec)
Master [gtid]>commit;
Query OK, 0 rows affected (0.00 sec)
Master [gtid]>show master status ;
+------------------+----------+--------------+------------------+------------------------------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+------------------------------------------+
| mysql-bin.000004 | 1068 | | | dff98809-55c3-11e9-a58b-000c2928f5dd:1-5 |
+------------------+----------+--------------+------------------+------------------------------------------+
1 row in set (0.00 sec)
Master [gtid]>begin;
Query OK, 0 rows affected (0.00 sec)
Master [gtid]>insert into t2 values(1);
Query OK, 1 row affected (0.00 sec)
Master [gtid]>commit;
Query OK, 0 rows affected (0.01 sec)
Master [gtid]>show master status ;
+------------------+----------+--------------+------------------+------------------------------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+------------------------------------------+
| mysql-bin.000004 | 1321 | | | dff98809-55c3-11e9-a58b-000c2928f5dd:1-6 |
+------------------+----------+--------------+------------------+------------------------------------------+
1 row in set (0.00 sec)
基于GTID进行查看binlog
具备GTID后,截取查看某些事务日志:
--include-gtids
--exclude-gtids
mysqlbinlog --include-gtids='dff98809-55c3-11e9-a58b-000c2928f5dd:1-6' --exclude-gtids='dff98809-55c3-11e9-a58b-000c2928f5dd:4' /data/binlog/mysql-bin.000004
跳过多个事务id
mysqlbinlog --include-gtids='dff98809-55c3-11e9-a58b-000c2928f5dd:1-6' --exclude-gtids='dff98809-55c3-11e9-a58b-000c2928f5dd:4,dff98809-55c3-11e9-a58b-000c2928f5dd:5' /data/binlog/mysql-bin.000004
GTID的幂等性
开启GTID后,MySQL恢复Binlog时,重复GTID的事务不会再执行了
就想恢复?怎么办?
--skip-gtids
mysqlbinlog --skip-gtids --include-gtids='3ca79ab5-3e4d-11e9-a709-000c293b577e:4' /data/binlog/mysql-bin.000004 > /tmp/binlog.sql
set sql_log_bin=0;
source /tmp/binlog.sql
set sql_log_bin=1;
使用二进制日志恢复数据案例
故障环境介绍
创建了一个库 db, 导入了表t1 ,t1表中录入了很多数据
一个开发人员,drop database db;
没有备份,日志都在.怎么恢复?
思路:找到建库语句到删库之前所有的日志,进行恢复.(开启了GTID模式)
故障案例模拟:
(0) drop database if exists db ;
(1) create database db charset utf8;
(2) use db;
(3) create table t1 (id int);
(4) insert into t1 values(1),(2),(3);
(5) insert into t1 values(4),(5),(6);
(6) commit
(7) update t1 set id=30 where id=3;
(8) commit;
(9) delete from t1 where id=4;
(10)commit;
(11)insert into t1 values(7),(8),(9);
(12)commit;
(13)drop database db;
========================
drop database if exists db ;
create database db charset utf8;
use db;
create table t1 (id int);
insert into t1 values(1),(2),(3);
insert into t1 values(4),(5),(6);
commit;
update t1 set id=30 where id=3;
commit;
delete from t1 where id=4;
commit;
insert into t1 values(7),(8),(9);
commit;
drop database db;
=======
运行以上语句，模拟故障场景
需求：将数据库恢复到以下状态（提示第9步和第13步是误操作，其他都是正常操作）
恢复过程(无GTID时的恢复)
1，查看当前使用的 binlog文件
oldguo [db]>show master status ;
+------------------+----------+--------------+------------------+-------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000006 | 1873 | | | |
+------------------+----------+--------------+------------------+-------------------+
2.查看事件：
第一段：
| mysql-bin.000006 | 813 | Query | 1 | 907 | use `db`; create table t1 (id int) |
| mysql-bin.000006 | 907 | Query | 1 | 977 | BEGIN |
| mysql-bin.000006 | 977 | Table_map | 1 | 1020 | table_id: 77 (db.t1) |
| mysql-bin.000006 | 1020 | Write_rows | 1 | 1070 | table_id: 77 flags: STMT_END_F |
| mysql-bin.000006 | 1070 | Table_map | 1 | 1113 | table_id: 77 (db.t1) |
| mysql-bin.000006 | 1113 | Write_rows | 1 | 1163 | table_id: 77 flags: STMT_END_F |
| mysql-bin.000006 | 1163 | Xid | 1 | 1194 | COMMIT /* xid=74 */ |
| mysql-bin.000006 | 1194 | Query | 1 | 1264 | BEGIN |
| mysql-bin.000006 | 1264 | Table_map | 1 | 1307 | table_id: 77 (db.t1) |
| mysql-bin.000006 | 1307 | Update_rows | 1 | 1353 | table_id: 77 flags: STMT_END_F |
| mysql-bin.000006 | 1353 | Xid | 1 | 1384 | COMMIT /* xid=77 */
mysqlbinlog --start-position=813 --stop-position=1384 /data/mysql/mysql-bin.000006 >/tmp/bin1.sql
第二段：
| mysql-bin.000006 | 1568 | Query | 1 | 1638 | BEGIN |
| mysql-bin.000006 | 1638 | Table_map | 1 | 1681 | table_id: 77 (db.t1) |
| mysql-bin.000006 | 1681 | Write_rows | 1 | 1731 | table_id: 77 flags: STMT_END_F |
| mysql-bin.000006 | 1731 | Xid | 1 | 1762 | COMMIT /* xid=81 */
mysqlbinlog --start-position=1568 --stop-position=1762 /data/mysql/mysql-bin.000006 >/tmp/bin2.sql
3，恢复
set sql_log_bin=0;
source /tmp/bin1.sql
source /tmp/bin2.sql
set sql_log_bin=1;
oldguo [db]>select * from t1;
+------+
| id |
+------+
| 1 |
| 2 |
| 30 |
| 4 |
| 5 |
| 6 |
| 7 |
| 8 |
| 9 |
有GTID的恢复:
mysqlbinlog --skip-gtids --include-gtids='3ca79ab5-3e4d-11e9-a709-000c293b577e:7-12' mysql-bin.000004> /tmp/bin.sql
恢复
set sql_log_bin=0;
source /tmp/bin.sql
二进制日志其他操作
自动清理日志
show variables like '%expire%';
expire_logs_days 0
自动清理时间,是要按照全备周期+1
set global expire_logs_days=8;
永久生效:修改配置文件
vim my.cnf
expire_logs_days=15;
企业建议,至少保留两个全备周期+1的binlog
手工清理（危险）
PURGE BINARY LOGS BEFORE now() - INTERVAL 3 day;
PURGE BINARY LOGS TO 'mysql-bin.000010';
注意:不要手工 rm binlog文件
1. my.cnf binlog关闭掉,启动数据库
2.把数据库关闭,开启binlog,启动数据库
删除所有binlog,并从000001开始重新记录日志
reset master; 主从关系中，主库执行此操作，主从环境必崩
日志是怎么滚动
flush logs;
重启mysql也会自动滚动一个新的
日志文件达到1G大小(max_binlog_size)
| max_binlog_size | 1073741824
备份时,加入参数也可以自动滚动
slow_log 慢日志
记录慢SQL语句的日志,定位低效SQL语句的工具日志
mysql> show variables like '%slow%';
+---------------------------+-------------------------------------+
| Variable_name | Value |
+---------------------------+-------------------------------------+
| log_slow_admin_statements | OFF |
| log_slow_slave_statements | OFF |
| slow_launch_time | 2 |
| slow_query_log | ON |
| slow_query_log_file | /data/mysql/data/localhost-slow.log |
+---------------------------+-------------------------------------+
5 rows in set (0.00 sec)
开启慢日志(默认没开启)
开关:
slow_query_log=1
文件位置及名字
slow_query_log_file=/data/mysql/slow.log
设定慢查询时间:
long_query_time=0.1
没走索引的语句也记录:
log_queries_not_using_indexes #不走索引
vim /etc/my.cnf
slow_query_log=1
slow_query_log_file=/data/mysql/slow.log
long_query_time=0.1
log_queries_not_using_indexes
systemctl restart mysqld
mysqldumpslow 分析慢日志
mysqldumpslow -s c -t 10 /data/mysql/slow.log
# 第三方工具(自己扩展)
https://www.percona.com/downloads/percona-toolkit/LATEST/
yum install perl-DBI perl-DBD-MySQL perl-Time-HiRes perl-IO-Socket-SSL perl-Digest-MD5
toolkit工具包中的命令:
./pt-query-diagest /data/mysql/slow.log
Anemometer基于pt-query-digest将MySQL慢查询可视化
常用的参数
-s：是表示按照何种方式排序
|
参数值 |
说明 |
|
c |
访问计数 |
|
l |
锁定时间 |
|
r |
返回记录 |
|
t |
查询时间 |
|
al |
平均锁定时间 |
|
ar |
平均返回记录数 |
|
at |
平均查询时间 |
-t：是top n的意思，即为返回前面多少条的数据
-g：后边可以写一个正则匹配模式，大小写不敏感的
Mysql慢查询日志分析及percona-toolkit的使用
相关文献：
https://blog.csdn.NET/qq_41600067/article/details/112952251
https://naibawu.com/1250.html
https://www.cnblogs.com/yaomaomao/p/12672283.html
本文来自博客园，作者：孙龙-程序员，转载请注明原文链接：
https://www.cnblogs.com/sunlong88/p/16655005.html
