环境模拟：

1.删除掉数据目录下，ib*文件：

[root@jj-svn160_231 data]# rm -rf ib*

2.登录数据库插入数据：[发现数据可以正常插入]

mysql> use test

mysql> insert into a select * from a;

解决方法如下：

1.数据库不能停机

2.查看当前数据库进程号：从下面可以看出mysql的进程号为：20301

[root@jj-svn160_231 test]# ps -ef |grep mysql

root     14473  4261  0 19:27 tty1     00:00:00 ./mysql -uroot -px xxxxxxx

root     19918     1  0 20:14 pts/0    00:00:00 /bin/sh /opt/mysql/bin/mysqld_safe –datadir=/home/mysql/data –pid-file=/home/mysql/data/jj-   svn160_231.dacn.org.pid

mysql    20301 19918  0 20:14 pts/0    00:00:00 /opt/mysql/bin/mysqld –basedir=/opt/mysql –datadir=/home/mysql/data –plugin-dir=/opt/mysql/lib/plugin –user=mysql –log-error=/var/log/mysqld.log –pid-file=/home/mysql/data/jj-svn160_231.dacn.org.pid –socket=/tmp/mysql.sock –port=3306

root     25443 14404  0 21:27 pts/0    00:00:00 ./mysql -uroot -px xxxxxxx

root     25624 14450  0 21:29 pts/1    00:00:00 grep mysql

3.从内存中查看当前ibdatta和ib_logfile文件信息(看到这里我们放心多了，文件还在，只是被标记为删除状态)

[root@jj-svn160_231 test]# ls -la /proc/20301/fd/ |grep -e ibdata -e ib_

lrwx—— 1 root  root  64 Mar  6 21:30 10 -> /home/mysql/data/ib_logfile0 (deleted)

lrwx—— 1 root  root  64 Mar  6 21:30 11 -> /home/mysql/data/ib_logfile1 (deleted)

lrwx—— 1 root  root  64 Mar  6 21:30 4 -> /home/mysql/data/ibdata1 (deleted)

lrwx—— 1 root  root  64 Mar  6 21:30 9 -> /home/mysql/data/ibdata2 (deleted)

4.此时，直接将内存信息复制到数据目录是不行的，会造成空间损坏，因此我们应该采取加全局读锁操作，讲脏页全部刷入磁盘

mysql> flush tables with read lock;

5.查看当前innodb引擎的状态：[保证log sequence number的值和last checkpoint 值相等]

mysql> show  engine innodb  status\G;

—

LOG

—

Log sequence number 1838299

Log flushed up to   1838299

Last checkpoint at  1838299

注：有时脏页比较多，可以设置全局参数，加快速度：mysql> set global innodb_max_dirty_pages_pct=0;

保证一些后台线程完成它们的工作，如.ibuf=1

————————————-

INSERT BUFFER AND ADAPTIVE HASH INDEX

————————————-

Ibuf: size 1, free list len 0, seg size 2, 0 merges

purge thread 应该purge了全部的transcations：

————

TRANSACTIONS

————

Trx id counter 2B02

Purge done for trx’s n:o < 2503 undo n:o < 0

History list length 39

确保innodb不再进行写操作：

——–

FILE I/O

——–

I/O thread 0 state: waiting for completed aio requests (insert buffer thread)

I/O thread 1 state: waiting for completed aio requests (log thread)

I/O thread 2 state: waiting for completed aio requests (read thread)

I/O thread 3 state: waiting for completed aio requests (read thread)

I/O thread 4 state: waiting for completed aio requests (read thread)

I/O thread 5 state: waiting for completed aio requests (read thread)

I/O thread 6 state: waiting for completed aio requests (write thread)

I/O thread 7 state: waiting for completed aio requests (write thread)

I/O thread 8 state: waiting for completed aio requests (write thread)

I/O thread 9 state: waiting for completed aio requests (write thread)

Pending normal aio reads: 0 [0, 0, 0, 0] , aio writes: 0 [0, 0, 0, 0] ,

ibuf aio reads: 0, log i/o’s: 0, sync i/o’s: 0

Pending flushes (fsync) log: 0; buffer pool: 0

239 OS file reads, 15 OS file writes, 13 OS fsyncs

0.00 reads/s, 0 avg bytes/read, 0.00 writes/s, 0.00 fsyncs/s

6.以上数值如果都正常的话，则可以把内存的数据复制到数据目录，并修改数据的权限

[root@jj-svn160_231 test]# cp /proc/20301/fd/10 /home/mysql/data/ib_logfile0

[root@jj-svn160_231 test]# cp /proc/20301/fd/11 /home/mysql/data/ib_logfile1

[root@jj-svn160_231 test]# cp /proc/20301/fd/4 /home/mysql/data/ibdata1

[root@jj-svn160_231 test]# cp /proc/20301/fd/9 /home/mysql/data/ibdata2

[root@jj-svn160_231 data]# chown mysql. ib*

7.释放全局读锁，并重启数据库，验证数据：

mysql> unlock tables;

[root@jj-svn160_231 bin]# /etc/init.d/mysqld restart

至此，ib* 相关文件就已经全部恢复完成

Author: Michael