ddrescue recovers 0 byte from HDD

[bobby_yan]

1 of the 3 HDDs (Western Digitial WD3003FZEX Black 3TB SATA 6GB/S 7200RPM 64MB Cache 3.5IN Hard Drive) fails in a RAID 0 array. At first, it disappears from the OS. Then, it re-appears after re-plug the cables. However, it is not stable, sometime even the serial number cannot be read out (see below).

I try to clone it first with ddrescue, but even cannot even read a byte.

Code: Select all

sudo ddrescue -f  /dev/sdc Port1.img Port1.log
GNU ddrescue 1.22
     ipos:    3000 GB, non-trimmed:        0 B,  current rate:       0 B/sB/s
     opos:    3000 GB, non-scraped:        0 B,  average rate:       0 B/s
non-tried:        0 B,  bad-sector:    3000 GB,    error rate:  23256 kB/s
  rescued:        0 B,   bad areas:        1,        run time: 14h 18m 13s
pct rescued:    0.00%, read errors:5906318682,  remaining time:         n/a
                              time since last successful read:         n/a
Finished

sudo ddrescue -f -d /dev/sdc Port1.img Port1.log
GNU ddrescue 1.22
     ipos:   38402 MB, non-trimmed:   38408 MB,  current rate:       0 B/s
     opos:   38402 MB, non-scraped:        0 B,  average rate:       0 B/s
non-tried:    2962 GB,  bad-sector:        0 B,    error rate:    875 MB/s
  rescued:        0 B,   bad areas:        0,        run time:         17s
pct rescued:    0.00%, read errors:   586172,  remaining time:         n/a
                              time since last successful read:         n/a
Copying non-tried blocks... Pass 5 (forwards)^C
  Interrupted by user

sudo ddrescue -f -d -R /dev/sdc Port1.img Port1.log
GNU ddrescue 1.22
     ipos:    2895 GB, non-trimmed:  105045 MB,  current rate:       0 B/s
     opos:    2895 GB, non-scraped:        0 B,  average rate:       0 B/s
non-tried:    2895 GB,  bad-sector:        0 B,    error rate:   2350 MB/s
  rescued:        0 B,   bad areas:        0,        run time:         46s
pct rescued:    0.00%, read errors:  1602973,  remaining time:         n/a
                              time since last successful read:         n/a
Copying non-tried blocks... Pass 5 (backwards)

After ddrescue, I try dd as well, but it copies out more than 16TB data and is continuing, which is obviously wrong.

Code: Select all

sudo dd if=/dev/sdc conv=sync,noerror bs=64K | gzip -c  > ./Port1.img.gz
dd: error reading '/dev/sdc': Input/output error
0+248815925 records in
248815925+0 records out
16306400460800 bytes (16 TB, 15 TiB) copied, 91157.6 s, 179 MB/s
dd: error reading '/dev/sdc': Input/output error
0+248815926 records in
248815926+0 records out
16306400526336 bytes (16 TB, 15 TiB) copied, 91157.6 s, 179 MB/s
dd: error reading '/dev/sdc': Input/output error
0+248815927 records in
248815927+0 records out
16306400591872 bytes (16 TB, 15 TiB) copied, 91157.6 s, 179 MB/s
^C0+248815928 records in
248815927+0 records out
16306400591872 bytes (16 TB, 15 TiB) copied, 91157.6 s, 179 MB/s

Here is the HDD information, the one without serial number of the broken one. After reboot, there is a chance to read the serial number successfully, and it is WD-WMC5D0D61CYD.

Code: Select all

sudo mdadm --detail-platform
mdadm: imsm capabilities not found for controller: /sys/devices/pci0000:00/0000:00:11.4 (type SATA)
       Platform : Intel(R) Rapid Storage Technology
        Version : 14.8.0.2377
    RAID Levels : raid0 raid1 raid10 raid5
    Chunk Sizes : 4k 8k 16k 32k 64k 128k
    2TB volumes : supported
      2TB disks : supported
      Max Disks : 7
    Max Volumes : 2 per array, 4 per controller
 I/O Controller : /sys/devices/pci0000:00/0000:00:1f.2 (SATA)
          Port2 : /dev/sdd (WD-WMC5D0D9X6D8)
          Port3 : /dev/sde (WD-WMC1F0EARDW0)
          Port1 : /dev/sdc ()
          Port0 : - no device attached -
          Port4 : - no device attached -
          Port5 : - no device attached –

The motherboard is ASUS X99-E USB 3.1, the RAID 0 is setup with Intel RST, and the I/O Controller is still left in RAID mode in BIOS. However, the raid array only consists of WD-WMC5D0D9X6D8 and WD-WMC1F0EARDW0 in BIOS now, the broken is now a non-member drive. The system must somehow detect the drive failure and remove it from the array automatically.

Any further effort I should try before given if I would like to clone the failed drive? Change the PCB? Thank you.

Edit Jan. 21, 2021: I find the HDD works a few hours after reboot, and ddrescue reads about 90GB data out before the above mentioned 0-byte problem, at an average speed of 2 MB/s. Moreover, mdadm is able to read out the serial number just after reboot, but not any more after 0-byte problem reappears. Shall I keep rebooting the machine? Is there any command I can just reboot/reset the HDD rather than the whole system? Thank you.

[lcoughey]

It seems that the one bad drive has some internal issues that are preventing you from getting a clone of the drive. I don't believe a PCB swap with a ROM transfer will do much to remedy the situation. If the data is valuable to you, it is best to have it looked at by a professional data recovery lab before further damage is caused.

At the very least, you should get a full backup clone of the good drives too.

[ddrecovery]

Also posted on Reddit and HDDGuru.

[lcoughey]

Also posted on Reddit and HDDGuru.

..and data-medics.

[bobby_yan]

Thanks. I'm cloning other two drives. The data loss will lead to some extra work of several weeks, it's acceptable in the worse case. But I would like to try my best before giving up. After ddrescue, I tried dd as well, but it copies out more than 16TB data and is continuing, which is obviously wrong.

Code: Select all

sudo dd if=/dev/sdc conv=sync,noerror bs=64K | gzip -c  > ./Port1.img.gz
dd: error reading '/dev/sdc': Input/output error
0+248815925 records in
248815925+0 records out
16306400460800 bytes (16 TB, 15 TiB) copied, 91157.6 s, 179 MB/s
dd: error reading '/dev/sdc': Input/output error
0+248815926 records in
248815926+0 records out
16306400526336 bytes (16 TB, 15 TiB) copied, 91157.6 s, 179 MB/s
dd: error reading '/dev/sdc': Input/output error
0+248815927 records in
248815927+0 records out
16306400591872 bytes (16 TB, 15 TiB) copied, 91157.6 s, 179 MB/s
^C0+248815928 records in
248815927+0 records out
16306400591872 bytes (16 TB, 15 TiB) copied, 91157.6 s, 179 MB/s

[bobby_yan]

I find the HDD works a few hours after reboot, and ddrescue reads about 90GB data out before the above mentioned 0-byte problem, at an average speed of 2 MB/s. Moreover, mdadm is able to read out the serial number just after reboot, but not any more after 0-byte problem reappears.

Shall I keep rebooting the machine? Is there any command I can just reboot/reset the HDD rather than the whole system?

Thank you.

[lcoughey]

Repowering the drive is fine, but pay attention to the drive path assigned when it redetects. It may not be the same and you will need to modify you command accordingly.

[bobby_yan]

Repowering the drive is fine, but pay attention to the drive path assigned when it redetects. It may not be the same and you will need to modify you command accordingly.

I wonder if reboot causes repowering. I do not cutoff the power during reboot. Thank you.

[lcoughey]

Carefully pull the power from the drive or use an external power source with a switch.

[bobby_yan]

Carefully pull the power from the drive or use an external power source with a switch.

Thank you for the suggestion. Since it's failing drive, would power cycle hurt it and make it die faster? Is there any milder way without power cycle? Reboot machine does not lead to power cycle but it's not very convenience. Thank you.