1

I've dusted of a machine of which I want to repair an mdadm raid5 that I messed up. First the raid5 was 3 disks. A spare was added, just before one of the three started to fail. Spare one got used and failed disk was removed. Now months later, I cannot mount it correctly. The array is broken.

original build:

root# mdadm --create --metadata=1.0 --verbose /dev/md127 --chunk=512 --level=5 --raid-devices=3 /dev/sda1 /dev/sdb1 /dev/sdc1

current situation:

$ cat /proc/mdstat 
Personalities : 
md127 : inactive sdb1[1](S) sda1[0](S)
  3677730784 blocks super 1.0

unused devices: <none>  

mdadm -D /dev/md127

sudo mdadm -D /dev/md127
/dev/md127:
    Version : 1.0
 Raid Level : __raid0__
Total Devices : 1
Persistence : Superblock is persistent

      State : inactive

       Name : nas:127  (local to host nas)
       UUID : 71da073c:d1928293:6947fa19:92d8a7bd
     Events : 1

Number   Major   Minor   RaidDevice

   -       8       17        -        /dev/sdb1

output of examine for each drive

$ sudo mdadm -E /dev/sd{b,c,e}1
**/dev/sdb1**:
      Magic : a92b4efc
    Version : 1.0
Feature Map : 0x1
 Array UUID : 71da073c:d1928293:6947fa19:92d8a7bd
       Name : nas:127  (local to host nas)
Creation Time : Sun Dec 10 23:26:56 2017
 Raid Level : raid5
Raid Devices : 4

Avail Dev Size : 3677730784 (1753.68 GiB 1883.00 GB)
 Array Size : 5516594688 (5261.03 GiB 5648.99 GB)
Used Dev Size : 3677729792 (1753.68 GiB 1883.00 GB)
Super Offset : 3677730800 sectors
Unused Space : before=0 sectors, after=992 sectors
      State : clean
Device UUID : e1fdc3d2:b0f117a5:11856184:17db9522

Internal Bitmap : -16 sectors from superblock
Update Time : Mon Dec 18 11:48:12 2017
Bad Block Log : 512 entries available at offset -8 sectors
   Checksum : 54a1b1a7 - correct
     Events : **1**

     Layout : left-symmetric
 Chunk Size : 512K

Device Role : Active device 1
Array State : AA.. ('A' == active, '.' == missing, 'R' == replacing)

**/dev/sdc1**:
      Magic : a92b4efc
    Version : 1.0
Feature Map : 0x1
 Array UUID : a1498410:d13b2b4a:63379f8d:c821173f
       Name : fileserver:127
Creation Time : Mon Jan 19 15:35:41 2015
 Raid Level : raid5
Raid Devices : 4

Avail Dev Size : 3677730536 (1753.68 GiB 1883.00 GB)
 Array Size : 5516594688 (5261.03 GiB 5648.99 GB)
Used Dev Size : 3677729792 (1753.68 GiB 1883.00 GB)
Super Offset : 3677730800 sectors
Unused Space : before=0 sectors, after=992 sectors
      State : clean
Device UUID : 472f7a29:679e1f18:87ee0d4c:88b2a62b

Internal Bitmap : -16 sectors from superblock
Update Time : Sun Dec 10 21:09:34 2017
Bad Block Log : 512 entries available at offset -8 sectors
   Checksum : 68dd142f - correct
     Events : **1934728**

     Layout : left-symmetric
 Chunk Size : 512K

Device Role : Active device 1
Array State : AAAA ('A' == active, '.' == missing, 'R' == replacing)

**/dev/sde1**:
      Magic : a92b4efc
    Version : 1.0
Feature Map : 0x1
 Array UUID : 71da073c:d1928293:6947fa19:92d8a7bd
       Name : taknas:127  (local to host taknas)
Creation Time : Sun Dec 10 23:26:56 2017
 Raid Level : raid5
Raid Devices : 4

Avail Dev Size : 3677730784 (1753.68 GiB 1883.00 GB)
 Array Size : 5516594688 (5261.03 GiB 5648.99 GB)
Used Dev Size : 3677729792 (1753.68 GiB 1883.00 GB)
Super Offset : 3677730800 sectors
Unused Space : before=0 sectors, after=992 sectors
      State : clean
Device UUID : ebd3b12c:975c1a0b:4653f1ed:e9788e37

Internal Bitmap : -16 sectors from superblock
Update Time : Mon Dec 18 11:48:12 2017
Bad Block Log : 512 entries available at offset -8 sectors
   Checksum : 931a5e9d - correct
     Events : **1**

     Layout : left-symmetric
 Chunk Size : 512K

Device Role : Active device 0
Array State : AA.. ('A' == active, '.' == missing, 'R' == replacing)

thinking out loud:

It looks like the raid5 failed, went into raid0 with 2 disk, while sda1 isn't there. And in spare mode.
Also, the events counter reset for sdb1 and sde1. sdb1 seems to be in another array. I may have forgotten to remove the failing disk from the array correctly, as it thinks it consists of 4.

Not sure what to do here to repair the raid array and keep data intact.

1 Answer 1

0

There's probably something in sudo dmesg about sdb1 being assembled as part of a raid0 md127. That's curious, you should probably examine it.

Either way, that dysfunctional array needs to be stopped because it's hogging sdb1 now:

sudo mdadm --stop /dev/md127

Then try assembling it with what seem to be the right two out of three:

sudo mdadm --assemble /dev/md127 /dev/sdb1 /dev/sde1 --verbose

If that works out, then add the odd one:

sudo mdadm /dev/md127 --add /dev/sdc1
2
  • Wasn't able to find why it became part of raid0. I'm guessing because one drive went missing? I was able to reassamble first with just the 2, and add the 3rd. But I'm unable to mount it, as it gives the following error: $ sudo mount /dev/md127 /mnt/raid5/ mount: /mnt/raid5: mount(2) system call failed: File too large. (cant seem to get code block in comment)
    – nieweling
    Jun 13, 2018 at 23:47
  • @niewelimg weird, is there more detail in dmesg about that failure? Jun 15, 2018 at 5:43

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .