debian(squeeze)mdadm建立的raid崩溃求助

ljwsy

debian squeeze的系统，用mdadm建立raid1，跑有近两年无问题。前两天感觉不对去终端看发现报有fs inode方面的错误一大堆（具体没记下来），登陆不行，热键启动不行……只有强断电，然后就再也启不了。启动报错为文件系统错误。

这是另一台类似机上的环境：

1. root@s-ii:~# uname -a
   
2. Linux s-ii 2.6.32-5-686-bigmem #1 SMP Sun Sep 23 10:27:25 UTC 2012 i686 GNU/Linux
   
3. 
   
4. root@s-ii:~# df -h
   
5. Filesystem            Size  Used Avail Use% Mounted on
   
6. /dev/md0              942M  578M  317M  65% /
   
7. tmpfs                 4.0G     0  4.0G   0% /lib/init/rw
   
8. udev                  4.0G  280K  4.0G   1% /dev
   
9. tmpfs                 4.0G     0  4.0G   0% /dev/shm
   
10. /dev/md3              312G   25G  272G   9% /home
    
11. /dev/md5              110G   72G   34G  69% /mnt/nxp/disk
    
12. /dev/md8               36G  177M   34G   1% /mnt/nxp/swap
    
13. /dev/md7               37G  177M   35G   1% /mnt/nxp/user
    
14. /dev/md6               19G  253M   18G   2% /mnt/nxp/wks
    
15. /dev/md2              1.9G   35M  1.8G   2% /tmp
    
16. /dev/md4               19G  898M   17G   6% /usr
    
17. /dev/md1               12G  773M  9.7G   8% /var
    
18. 
    
19. 
    
20. root@s-ii:~# cat /proc/mdstat
    
21. Personalities : [raid0] [raid1]
    
22. md8 : active raid0 sda13[0] sdb13[3] sda14[2] sdb14[1]
    
23.       37302528 blocks 64k chunks
    
24.       
    
25. md7 : active raid0 sda11[0] sdb11[3] sda12[2] sdb12[1]
    
26.       39069696 blocks 64k chunks
    
27.       
    
28. md6 : active raid0 sda10[0] sdb10[1]
    
29.       19534848 blocks 64k chunks
    
30.       
    
31. md5 : active raid0 sda8[0] sdb8[3] sda9[2] sdb9[1]
    
32.       117177600 blocks 64k chunks
    
33.       
    
34. md4 : active raid1 sda7[0] sdb7[1]
    
35.       19534912 blocks [2/2] [UU]
    
36.       
    
37. md3 : active raid1 sda6[0] sdb6[1]
    
38.       332031296 blocks [2/2] [UU]
    
39.       
    
40. md2 : active raid1 sda5[0] sdb5[1]
    
41.       1951744 blocks [2/2] [UU]
    
42.       
    
43. md1 : active raid1 sda3[0] sdb3[1]
    
44.       11719296 blocks [2/2] [UU]
    
45.       
    
46. md0 : active raid1 sda1[0] sdb1[1]
    
47.       979840 blocks [2/2] [UU]
    
48.       
    
49. unused devices: <none>
    
50. 
    
51. 
    

复制代码

为了要出里面的数据，新做了一台debian _Jessie_，挂上一只坏盘，这是分区情况：

1. root@ljwsy:~# fdisk -l /dev/sdc
   
2. 
   
3. Disk /dev/sdc: 931.5 GiB, 1000204886016 bytes, 1953525168 sectors
   
4. Units: sectors of 1 * 512 = 512 bytes
   
5. Sector size (logical/physical): 512 bytes / 4096 bytes
   
6. I/O size (minimum/optimal): 4096 bytes / 4096 bytes
   
7. Disklabel type: dos
   
8. Disk identifier: 0x00017d93
   
9. 
   
10. Device     Boot     Start        End    Sectors   Size Id Type
    
11. /dev/sdc1  *           63    1959929    1959867   957M 83 Linux
    
12. /dev/sdc2         1959930   17591174   15631245   7.5G 82 Linux swap / Solaris
    
13. /dev/sdc3        17591175   56661254   39070080  18.6G 83 Linux
    
14. /dev/sdc4        56661255 1953520064 1896858810 904.5G  5 Extended
    
15. /dev/sdc5        56661318   60565049    3903732   1.9G 83 Linux
    
16. /dev/sdc6        60565113   78140159   17575047   8.4G 83 Linux
    
17. /dev/sdc7        78140223   97675199   19534977   9.3G 83 Linux
    
18. /dev/sdc8        97675263  839862134  742186872 353.9G 83 Linux
    
19. /dev/sdc9       839862198 1953520064 1113657867   531G 83 Linux
    
20. 
    
21. 
    
22. 
    

复制代码

然后加个新盘多方尝试不成功（下面的md0是拿来试水的，其他区没动）,sdb是新盘：

1. 
   
2. root@ljwsy:~# cat /proc/mdstat
   
3. Personalities : [raid1]
   
4. md0 : active raid1 sdc1[2] sdb1[0]
   
5.       979392 blocks super 1.2 [2/2] [UU]
   
6.       
   
7. md3 : inactive sdc6[2](S)
   
8.       8786499 blocks super 1.2
   
9.       
   
10. md6 : inactive sdc9[2](S)
    
11.       556827909 blocks super 1.2
    
12.       
    
13. md1 : inactive sdc3[2](S)
    
14.       19534016 blocks super 1.2
    
15.       
    
16. md2 : inactive sdc5[2](S)
    
17.       1950842 blocks super 1.2
    
18.       
    
19. md5 : inactive sdc8[2](S)
    
20.       371092412 blocks super 1.2
    
21.       
    
22. md4 : inactive sdc7[2](S)
    
23.       9766464 blocks super 1.2
    
24.       
    
25. unused devices: <none>
    
26. 
    
27. root@ljwsy:~# mdadm -E /dev/sdc1
    
28. /dev/sdc1:
    
29.           Magic : a92b4efc
    
30.         Version : 1.2
    
31.     Feature Map : 0x0
    
32.      Array UUID : f9d4b320:c4922ad9:17645e56:897e8cb5
    
33.            Name : ljwsy:0  (local to host ljwsy)
    
34.   Creation Time : Thu Nov 12 11:21:59 2015
    
35.      Raid Level : raid1
    
36.    Raid Devices : 2
    
37. 
    
38. Avail Dev Size : 1958811 (956.61 MiB 1002.91 MB)
    
39.      Array Size : 979392 (956.60 MiB 1002.90 MB)
    
40.   Used Dev Size : 1958784 (956.60 MiB 1002.90 MB)
    
41.     Data Offset : 1056 sectors
    
42.    Super Offset : 8 sectors
    
43.    Unused Space : before=968 sectors, after=27 sectors
    
44.           State : clean
    
45.     Device UUID : 4145857f:81c186e8:0a782966:e12dfc71
    
46. 
    
47.     Update Time : Thu Nov 12 11:22:40 2015
    
48.   Bad Block Log : 512 entries available at offset 72 sectors
    
49.        Checksum : f57c2b1 - correct
    
50.          Events : 19
    
51. 
    
52. 
    
53.    Device Role : Active device 1
    
54.    Array State : AA ('A' == active, '.' == missing, 'R' == replacing)
    
55. 
    
56. root@ljwsy:~# mdadm -E /dev/sdb1
    
57. /dev/sdb1:
    
58.           Magic : a92b4efc
    
59.         Version : 1.2
    
60.     Feature Map : 0x0
    
61.      Array UUID : f9d4b320:c4922ad9:17645e56:897e8cb5
    
62.            Name : ljwsy:0  (local to host ljwsy)
    
63.   Creation Time : Thu Nov 12 11:21:59 2015
    
64.      Raid Level : raid1
    
65.    Raid Devices : 2
    
66. 
    
67. Avail Dev Size : 1958880 (956.64 MiB 1002.95 MB)
    
68.      Array Size : 979392 (956.60 MiB 1002.90 MB)
    
69.   Used Dev Size : 1958784 (956.60 MiB 1002.90 MB)
    
70.     Data Offset : 1056 sectors
    
71.    Super Offset : 8 sectors
    
72.    Unused Space : before=968 sectors, after=96 sectors
    
73.           State : clean
    
74.     Device UUID : 8c21a15c:dad6feed:64f78efc:c6a07396
    
75. 
    
76.     Update Time : Thu Nov 12 11:22:40 2015
    
77.   Bad Block Log : 512 entries available at offset 72 sectors
    
78.        Checksum : acc8a6d7 - correct
    
79.          Events : 19
    
80. 
    
81. 
    
82.    Device Role : Active device 0
    
83.    Array State : AA ('A' == active, '.' == missing, 'R' == replacing)
    
84. 
    

复制代码

挂接时：

1. 
   
2. root@ljwsy:~# mount /dev/md0 /mnt/tmp
   
3. mount: wrong fs type, bad option, bad superblock on /dev/md0,
   
4.        missing codepage or helper program, or other error
   
5. 
   
6.        In some cases useful info is found in syslog - try
   
7.        dmesg | tail or so.
   
8. 
   

复制代码

按其提示dmesg没见有异常。

再来，针对没动过的区：

1. 
   
2. root@ljwsy:~# mdadm -E /dev/sdc3
   
3. /dev/sdc3:
   
4.           Magic : a92b4efc
   
5.         Version : 1.2
   
6.     Feature Map : 0x0
   
7.      Array UUID : 70b21eb6:3fb9ddaa:e06afa2a:236955e2
   
8.            Name : wb:1
   
9.   Creation Time : Tue Jan 15 20:25:33 2013
   
10.      Raid Level : raid1
    
11.    Raid Devices : 2
    
12. 
    
13. Avail Dev Size : 39068032 (18.63 GiB 20.00 GB)
    
14.      Array Size : 19533944 (18.63 GiB 20.00 GB)
    
15.   Used Dev Size : 39067888 (18.63 GiB 20.00 GB)
    
16.     Data Offset : 2048 sectors
    
17.    Super Offset : 8 sectors
    
18.    Unused Space : before=1968 sectors, after=144 sectors
    
19.           State : clean
    
20.     Device UUID : ec182df4:009d6cb4:b71fa16a:a5d02009
    
21. 
    
22.     Update Time : Wed Nov 11 09:03:16 2015
    
23.        Checksum : 10eec90d - correct
    
24.          Events : 1804
    
25. 
    
26. 
    
27.    Device Role : Active device 0
    
28.    Array State : AA ('A' == active, '.' == missing, 'R' == replacing)
    
29. root@ljwsy:~# mdadm -D /dev/md1
    
30. /dev/md1:
    
31.         Version : 1.2
    
32.      Raid Level : raid0
    
33.   Total Devices : 1
    
34.     Persistence : Superblock is persistent
    
35. 
    
36.           State : inactive
    
37. 
    
38.            Name : wb:1
    
39.            UUID : 70b21eb6:3fb9ddaa:e06afa2a:236955e2
    
40.          Events : 1804
    
41. 
    
42.     Number   Major   Minor   RaidDevice
    
43. 
    
44.        -       8       35        -        /dev/sdc3
    
45. 
    
46. root@ljwsy:~# mount /dev/md1 /mnt/tmp
    
47. mount: wrong fs type, bad option, bad superblock on /dev/md1,
    
48.        missing codepage or helper program, or other error
    
49. 
    
50.        In some cases useful info is found in syslog - try
    
51.        dmesg | tail or so.
    
52. 
    
53. root@ljwsy:~# dmesg
    
54. ……
    
55. [ 3691.999867] EXT4-fs (md1): unable to read superblock
    
56. [ 3691.999955] EXT4-fs (md1): unable to read superblock
    
57. [ 3692.000533] EXT4-fs (md1): unable to read superblock
    
58. root@ljwsy:~# fsck.ext3 /dev/md1
    
59. e2fsck 1.42.12 (29-Aug-2014)
    
60. fsck.ext3: Invalid argument while trying to open /dev/md1
    
61. 
    
62. The superblock could not be read or does not describe a valid ext2/ext3/ext4
    
63. filesystem.  If the device is valid and it really contains an ext2/ext3/ext4
    
64. filesystem (and not swap or ufs or something else), then the superblock
    
65. is corrupt, and you might try running e2fsck with an alternate superblock:
    
66.     e2fsck -b 8193 <device>
    
67. or
    
68.     e2fsck -b 32768 <device>
    
69. root@ljwsy:~# e2fsck -b 8193 /dev/md1
    
70. e2fsck 1.42.12 (29-Aug-2014)
    
71. e2fsck: Invalid argument while trying to open /dev/md1
    
72. 
    
73. The superblock could not be read or does not describe a valid ext2/ext3/ext4
    
74. filesystem.  If the device is valid and it really contains an ext2/ext3/ext4
    
75. filesystem (and not swap or ufs or something else), then the superblock
    
76. is corrupt, and you might try running e2fsck with an alternate superblock:
    
77.     e2fsck -b 8193 <device>
    
78. or
    
79.     e2fsck -b 32768 <device>
    
80. 
    
81. 
    

复制代码

没思路了，哪位同学指点一二，非常感谢你们的帮助。

EeeLo

我感觉文件系统崩溃主要跟你强制断电原因关系比较大，我一次也是强制断电，文件系统就崩溃了。
具体怎么恢复不太懂

ljwsy

回复 2# EeeLo


    是跟断电相关。昨天的０点左右机房断电，６:１０又来电，问电工及其他人，回答是没人动过。心中暗想：机房有女鬼，这是第二次了，吓死宝宝！！！中午正在做工突然UPS又叫，打电话给电工，他们也是没有什么动作，开灯一看，有电。晕，闸没跳呀。拉闸再上，电正常了。原来没有女鬼，失望之中……