两台服务器同时挂载共享存储盘，导致文件索引损坏

 

 

 

估计是两边一起挂盘，导致/dev/sdb1文件索引被覆盖，

210.76.211.64 上发现mysql总是启动失败， 挂载到210.76.211.88 mysql也失败

最终导致mysq无法启动

 

用fsck.ext4对/dev/sdb1进行扫描检查，虽修复，但数据库无法启动

必须在操作时，避免两边同时挂，用脚本避免

[root@localhost ~]# fsck.ext4 /dev/sdb1

e2fsck 1.41.12 (17-May-2010)

/dev/sdb1 is mounted.

e2fsck: 无法继续, 中止.

 

[root@localhost ~]# umount /dev/sdb1

[root@localhost ~]# fsck.ext4 /dev/sdb1

e2fsck 1.41.12 (17-May-2010)

/dev/sdb1 contains a file system with errors, 强制检查.

第一步: 检查inode,块,和大小

 

Running additional passes to resolve 块s claimed by more than one inode...

Pass 1B: Rescanning for multiply-claimed 块s

Multiply-claimed 块(s) in inode 2883617: 34065 34065

Pass 1C: Scanning directories for inodes with multiply-claimed 块s

Pass 1D: Reconciling multiply-claimed 块s

(There are 1 inodes containing multiply-claimed 块s.)

 

File /mysql/localhost.localdomain.err (inode #2883617, mod time Mon May 25 19:22:14 2015)

  has 2 multiply-claimed 块(s), shared with 0 file(s):

Clone multiply-claimed blocks? 是

 

第二步: 检查目录结构

入口 'mysql-relay-bin.000024' in /mysql (2883585) has 删除/unused inode 2883811.  清除? 是

第3步: 检查目录连接性

/lost+found未找到.创建? 是

Pass 4: Checking reference counts

第5步: 检查簇概要信息

块位图差异:  -(34304--34321) -11578881 -11578883

处理? 是

Free 块s count wrong for 簇 #0 (23516, counted=23515).

处理? 是

Free 块s count wrong for 簇 #1 (741, counted=759).

处理? 是

Free 块s count wrong for 簇 #353 (26461, counted=26462).

处理? 是

Free 块s count wrong (24426642, counted=24426660).

处理? 是

Inode位图差异:  -2883811

处理? 是

Free inodes count wrong for 簇 #352 (7948, counted=7949).

处理? 是

Free inodes count wrong (6553345, counted=6553346).

处理? 是

/dev/sdb1: ***** 文件系统已修改 *****

/dev/sdb1: 254/6553600 files (2.0% non-contiguous), 1787395/26214055 blocks