今天在群里看到有人说不熟悉innodb把ibdata(数据文件)和ib_logfile(事务日志)文件误删除了。不知道怎么解决。当时我也不知道怎么办。后来查阅相关资料。终找到解决方法。其实恢复也挺简单的。我们不知道的时候就觉得难了。谁说不是这样呢?
下面我们就来模拟生产环境下,人为删除数据文件和重做日志文件。然后详细说明恢复步骤。
1.用sysbench模拟数据的写入,如下所示:
[[email protected] ~]# sysbench --test=oltp --oltp-table-size=1000000 --oltp-read-only=off --init-rng=on --num-threads=16 --max-requests=0 --oltp-dist-type=uniform --max-time=1800 --mysql-user=root --mysql-socket=/tmp/mysqld.sock --mysql-password=123456 --db-driver=mysql --mysql-table-engine=innodb --oltp-test-mode=complex prepare
sysbench 0.4.10: multi-threaded system evaluation benchmarkCreating table ‘sbtest‘...
Creating 1000000 records in table ‘sbtest‘...
2.使用命令rm -f ib*删除数据文件和事务日志文件:
[[email protected] mysql]# ls
employees ib_logfile1 mysql-bin.000003 mysql-bin.000008 performance_schema world_innodb yayun-mysql-server.pid
general.log menagerie mysql-bin.000004 mysql-bin.000009 sakila world_myisam
host mysql mysql-bin.000005 mysql-bin.000010 sbtest xtrabackup_binlog_pos_innodb
ibdata1 mysql-bin.000001 mysql-bin.000006 mysql-bin.index slow-query.log yayun
ib_logfile0 mysql-bin.000002 mysql-bin.000007 percona test yayun-mysql-server.err
[[email protected]-mysql-server mysql]# rm -f ib*
[[email protected]-mysql-server mysql]#
下面我们来看看如何恢复:
若此时发现数据库还能正常工作,数据依然可读可写,切记:这个时候千万不要把mysqld进程杀死,否则真没法挽救了,你就等着哭吧。
(root@yayun 20:42:25pm> ) [yayun]>select * from t1;
+----+-------+
| id | name |
+----+-------+
| 1 | yayun |
| 2 | atlas |
| 3 | mysql |
+----+-------+
3 rows in set (0.00 sec)(root@yayun 20:42:28pm> ) [yayun]>insert into t1 select 4,‘python‘;
Query OK, 1 row affected (0.01 sec)
Records: 1 Duplicates: 0 Warnings: 0(root@yayun 20:42:48pm> ) [yayun]>select * from t1;
+----+--------+
| id | name |
+----+--------+
| 1 | yayun |
| 2 | atlas |
| 3 | mysql |
| 4 | python |
+----+--------+
4 rows in set (0.00 sec)(root@yayun 20:42:50pm> ) [yayun]>
我这里读写都正常。所以我们能够恢复成功的。
(1)首先,先找到mysqld进程的pid,如下所示:
[[email protected] ~]# netstat -nltp | grep mysqld
tcp 0 0 0.0.0.0:3306 0.0.0.0:* LISTEN 5725/mysqld
[[email protected]-mysql-server ~]#
可见mysqld的pid是5725,这一步是关键的一步。
(2)使用如下命令查看结果(非常重要)
[[email protected] ~]# ll /proc/5725/fd | egrep ‘ib_|ibdata‘
lrwx------. 1 root root 64 Apr 30 20:44 10 -> /data/mysql/ib_logfile1 (deleted)
lrwx------. 1 root root 64 Apr 30 20:44 4 -> /data/mysql/ibdata1 (deleted)
lrwx------. 1 root root 64 Apr 30 20:44 9 -> /data/mysql/ib_logfile0 (deleted)
[[email protected]-mysql-server ~]#
这里有相关很重要的知识,童鞋们自行查阅,删除一个文件时,并不是真正删除,而是打一个标记,同样在我们mysql数据库中,delete一条记录实际的删除操作也没有发生。
上面显示的结果中,其中10,4,9就是我们需要恢复的文件。
(3)在恢复文件前,需要执行flush tables with
read lock,确保数据库没有写入操作,以便我们完成恢复
(root@yayun 20:55:26pm> ) [(none)]>flush tables with read lock;
Query OK, 0 rows affected (0.00 sec)(root@yayun 20:55:31pm> ) [(none)]>
那么我们如何确定没有数据写入呢?分几个步骤查看
(1)设置脏页刷新比例(让脏页尽快刷新到磁盘)
(root@yayun 20:55:31pm> ) [(none)]>set global innodb_max_dirty_pages_pct=0;
Query OK, 0 rows affected (0.00 sec)(root@yayun 20:57:42pm> ) [(none)]>
(2)查看binlog日志写入情况,确保File和Position的值没有发生变化
(root@yayun 20:57:42pm> ) [(none)]>show master status;
+------------------+----------+--------------+------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB |
+------------------+----------+--------------+------------------+
| mysql-bin.000010 | 61704130 | | |
+------------------+----------+--------------+------------------+
1 row in set (0.00 sec)(root@yayun 20:59:11pm> ) [(none)]>
(3)查看innodb状态信息,确保脏页已经刷新到磁盘
(root@yayun 20:59:11pm> ) [(none)]>show engine innodb status\G------------
TRANSACTIONS
------------
Trx id counter B9E0F
Purge done for trx‘s n:o < B9E0C undo n:o < 0 #确保后台线程purge把undo log全部清刷掉-------------------------------------
INSERT BUFFER AND ADAPTIVE HASH INDEX
-------------------------------------
Ibuf: size 1, free list len 2543, seg size 2545, 0 merges #确保合并插入缓存等于1---
LOG
---
Log sequence number 6173930288
Log flushed up to 6173930288 #确保这里三个值保持一致,并且不再变化
Last checkpoint at 6173930288Buffer pool size 65534
Free buffers 50513
Database pages 15020
Old database pages 5506
Modified db pages 0 #确保脏页数量为0--------------
ROW OPERATIONS
--------------
0 queries inside InnoDB, 0 queries in queue
1 read views open inside InnoDB
Main thread process no. 5725, id 140014471358208, state: waiting for server activity
Number of rows inserted 1000004, updated 1995, deleted 0, read 2008
0.00 inserts/s, 0.00 updates/s, 0.00 deletes/s, 0.00 reads/s #确保插入,更新,删除为0
上面我们都确认后,就可以进行恢复操作了。记得前面我们查看要恢复的文件的命令吧,我们的mysqld进程的pid是5725,我们再次看看需要恢复的文件
[[email protected] ~]# ll /proc/5725/fd | egrep ‘ib_|ibdata‘
lrwx------. 1 root root 64 Apr 30 20:44 10 -> /data/mysql/ib_logfile1 (deleted)
lrwx------. 1 root root 64 Apr 30 20:44 4 -> /data/mysql/ibdata1 (deleted)
lrwx------. 1 root root 64 Apr 30 20:44 9 -> /data/mysql/ib_logfile0 (deleted)
[[email protected]-mysql-server ~]#
把10,4,9文件cp到原来mysql的数据目录下:
[[email protected] ~]# cd /proc/5725/fd
[[email protected]-mysql-server fd]# cp 10 /data/mysql/ib_logfile1
[[email protected]-mysql-server fd]# cp 4 /data/mysql/ibdata1
[[email protected]-mysql-server fd]# cp 9 /data/mysql/ib_logfile0
[[email protected]-mysql-server fd]#
修改文件权限
[[email protected] ~]# cd /data/mysql
[[email protected]-mysql-server mysql]# chown -R mysql.mysql ib*
[[email protected]-mysql-server mysql]# ll | egrep ‘ib_|ibdata1‘
-rw-r--r-- 1 mysql mysql 866123776 Apr 30 21:13 ibdata1
-rw-r--r-- 1 mysql mysql 67108864 Apr 30 21:13 ib_logfile0
-rw-r--r-- 1 mysql mysql 67108864 Apr 30 21:11 ib_logfile1
[[email protected]-mysql-server mysql]#
重启mysql,修复完成
[[email protected] mysql]# /etc/init.d/mysqld restart
Shutting down MySQL... [ OK ]
Starting MySQL.... [ OK ]
[[email protected]-mysql-server mysql]#
注意:生产环境切勿测试,童鞋们可以自己开启虚拟机进行测试。
参考资料
《MySQL 管理之道,性能调优,高可用与监控》