Recover files using Linux

The next steps depend on how bad you want the data back, and what risks you want to take.

Of course, you can always contact NETGEAR for Data Recovery, they offer this kind of service as a contract.

While there are some pending sectors on sdf and sdf3 is clearly out of sync (not by much), I don't understand why sdc3 doesn't get included in the RAID array, though it shows bad blocks (mdadm output).

You could try to backup the superblocks (if not already done), then recreate the RAID array. But this could result in irrevocable data loss.

(I'm not 100% sure of what is the best approach at this stage.)

Based on the outputs you provided, I think there are two possibilities.

- Again, this is dangerous territory -

• Either try to recreate the RAID as "--assume-clean".

https://raid.wiki.kernel.org/index.php/RAID_Recovery#Restore_array_by_recreating_.28after_multiple_device_failure.29

http://man7.org/linux/man-pages/man8/mdadm.8.html

• Or force the RAID array to assemble.

https://raid.wiki.kernel.org/index.php/RAID_Recovery#Trying_to_assemble_using_--force

For both:

• Either with sde3 sdd3 sdf3
• Or sde3 sdd3 sdc3

In theory, as you "--assume-clean" and only include three members, it shouldn't try to rewrite any block of data (but will overwrite the superblocks). So it shouldn't cause permanent damage. But it's a should.

 Parameters from the output you provided:

/dev/sdd3:
          Magic : a92b4efc
        Version : 1.2
    Feature Map : 0x0
     Array UUID : 9c92d78d:3fa2e084:a32cf226:37d5c3c2
           Name : 0e36f164:data-0
  Creation Time : Sun Feb  1 21:32:44 2015
     Raid Level : raid5
   Raid Devices : 4
 
Avail Dev Size : 615438961 (293.46 GiB 315.10 GB)
     Array Size : 923158272 (880.39 GiB 945.31 GB)
  Used Dev Size : 615438848 (293.46 GiB 315.10 GB)
    Data Offset : 262144 sectors
   Super Offset : 8 sectors
   Unused Space : before=262064 sectors, after=113 sectors
          State : active
    Device UUID : a51d6699:b35d008c:d3e1f115:610f3d0f
 
    Update Time : Tue Oct 31 08:32:47 2017
       Checksum : 3518f348 - correct
         Events : 14159
 
         Layout : left-symmetric
     Chunk Size : 64K
 
   Device Role : Active device 1
   Array State : AAA. ('A' == active, '.' == missing, 'R' == replacing)

/dev/sdc3:
   Device Role : Active device 2
/dev/sdd3:
   Device Role : Active device 1
/dev/sde3:
   Device Role : Active device 0
/dev/sdf3:
   Device Role : Active device 3

That's what I would try:

# Backup superblocks for each partition - if not already done
for partition in /dev/sd[a-f][0-9]; do echo "Backing up superblocks for $partition"; dd if=$partition of=/root/superblocks_$(basename $partition).mdsb bs=64k count=1; done
ls -lh /root/superblocks_*

# Backup of "mdadm --examine" for each partition - new
for partition in /dev/sd[a-f][0-9]; do echo "Backing up mdadm information for $partition"; mdadm --examine $partition > mdadm_-E_$(basename $partition).txt; done
ls -lh /root/mdadm_-E_*

# Start all healthy RAID arrays - if not already done
mdadm --assemble --verbose /dev/md126 /dev/sdc4 /dev/sdd4 /dev/sde5 /dev/sdf5
mdadm --assemble --verbose /dev/md125 /dev/sde4 /dev/sdf4
mdadm --assemble --verbose /dev/md127 /dev/sde6 /dev/sdf6

# Recreate the unhealthy RAID array - new
mdadm --create --verbose --assume-clean --level=5 --raid-devices=4 --size=461579136K --chunk=64K --data-offset=131072K /dev/md124 /dev/sde3 /dev/sdd3 missing /dev/sdf3

# Check the integrity - do it again
cat /proc/mdstat
btrfs device scan
btrfs filesystem show
btrfsck --readonly /dev/md127
mount -o ro /dev/md127 /mnt
btrfs filesystem usage /mnt