Storage Solutions Dying WD Hard Disk - Slow Transfer - Recovery Possible?

[varkey]

My Western Digital WD15EADS is almost dead. It still gets detected and I am able to see the files in it, but when I try to copy it to another hard disk the transfer rate is in Kilo Bytes

SMART shows high read errors which is increasing rapidly, and also Current Pending Sectors about 1388.

There is about 500 GB of data in it, is there any way I can recover the data? If the max speed I get is in kBs there is no point as I'd rather download the files again as it will be much faster.

 

[Gaurish]

you can try cloning the complete HDD. I had done that on a sever using dd command and later copied the image onto new replacement HDD of same size. Day Saved without any hassles

http://www.backuphowto.info/linux-backup-hard-disk-clone-dd

 

[varkey]

^ Will try that, but will it make any difference in the data transfer rates?

Edit : Creating a disc image using dd. So far so good, its making the image with good speed. Hope it works!!

 

[Gaurish]

^ Will try that, but will it make any difference in the data transfer rates?

Edit : Creating a disc image using dd. So far so good, its making the image with good speed. Hope it works!!

Yes, transfer speed with dd are fast because dd does not care about files, it copies byte by byte. even if you have serious fs errors, dd will still create a image from dying hdd. It has saved my a$$ on number of occasions

Best of luck! I hope your hdd does not give up

 

[varkey]

Hopefully it will work.

6884393+0 records in

6884393+0 records out

3524809216 bytes (3.5 GB) copied, 416.541 s, 8.5 MB/s

Its a 1.5TB hard disk, if it goes at that speed it will take 2 days to complete. But still better than the 50kB speed.

 

[nfsnfs]

So make disk clone , partion by partion or entire 1.5 TB at once ?

 

[varkey]

I bought one more 1.5TB to store the recovered data.

The dd command stopped at an error, so now trying with ddrescue. Made for data recovery, can be resumed very easily too. Will leave it running the whole night and see.

Right now its making an exact clone of the old hard disk to the new one.

rescued: 9521 MB, errsize: 0 B, current rate: 24117 kB/s

ipos: 9521 MB, errors: 0, average rate: 22830 kB/s

opos: 9521 MB

Copying data...

Edit : After 10 hours, hope it will finish in another 8-9 hours.

rescued: 889570 MB, errsize: 0 B, current rate: 27000 kB/s

ipos: 889570 MB, errors: 0, average rate: 25429 kB/s

opos: 889570 MB

Copying data...

 

[Gaurish]

Till date, hdd never failed on me while running dd, seems, bad luck is on your side!

how are things going with ddrescue? more bad luck or what

 

[blr_p]

My Western Digital WD15EADS is almost dead. It still gets detected and I am able to see the files in it, but when I try to copy it to another hard disk the transfer rate is in Kilo Bytes

I'm at a loss to explain why your drive is acting so slow :huh:

Actually your smart readings are pretty good, zero reallocation of sectors etc. I bet if you did an error scan there would no red blocks at all.

If this was in windows i'd have thought somehow your drive got stuck in PIO mode instead of being in ultraDMA or your BIOS setting somehow toggled away from ultraDMA for some reason.

 

[varkey]

ddrescue just finished, after about 32 hours. I increased the block size so it will be faster.

root@ubuntu:~# ddrescue -n -b 2048 -v /dev/sda /dev/sdc log.txt

About to copy 1500 GBytes from /dev/sda to /dev/sdc

Starting positions: infile = 0 B, outfile = 0 B

Copy block size: 32 hard blocks

Hard block size: 2048 bytes

Max_retries: 0 Split: no Truncate: no

Press Ctrl-C to interrupt

Initial status (read from logfile)

rescued: 1377 GB, errsize: 13475 kB, errors: 294

Current status

rescued: 1500 GB, errsize: 30035 kB, current rate: 24313 kB/s

ipos: 1500 GB, errors: 624, average rate: 4270 kB/s

opos: 1500 GB

Till about 1370GB it went fast, after which is was damn slow. It took 18 hours to copy from 1370 GB to 1500 GB. Anyway now that its finished, let me try mounting the partition. Hopefully will be able to read the data.

blr_p - According to SMART there are 1388 Weak Sectors which maybe causing the problem. Actually initially ddrecue was going at really good speeds.

Edit : I just mounted the partition as read only and the files were there and most of them were opening fine. Right now its checking the file system for errors, shoule take few hours I guess.

Gaurish - Thank you very much for suggesting dd, else I would have never recovered the files.

 

[Gaurish]

Gaurish - Thank you very much for suggesting dd, else I would have never recovered the files.

You should Say Thanks to GNU project folks specially Antonio Diaz the main coder behind the tool who choose to give this for free instead charging dollars for it so data recovery tools can reach to masses.almost all good data recovery tools are paid

PS: if you are feeling generous, you can also donate or even plain thank you email sent to the project mailing list or author directly would be appreciated

Cheers!

 

[blr_p]

blr_p - According to SMART there are 1388 Weak Sectors which maybe causing the problem. Actually initially ddrecue was going at really good speeds.

Disagree, acc to SMART there are no weak sectors on your HDD at all. The cable connected to the drive is also good which means the drive controller is fine as well. Beyond that is where the problem is.

Try and run the error scan, that is the tab after SMART.

Otherwise, hook this drive up on a USB or on a different PC and then try copying to it. This way you can elliminate the mobo its attached too and know for sure whether it really is the HDD or not.

 

[ahit]

I also feel like this is PIO mode stuck issue

 

[varkey]

Gaurish : GNU folks saved the day!

Disagree, acc to SMART there are no weak sectors on your HDD at all. The cable connected to the drive is also good which means the drive controller is fine as well. Beyond that is where the problem is.

Try and run the error scan, that is the tab after SMART.

Otherwise, hook this drive up on a USB or on a different PC and then try copying to it. This way you can elliminate the mobo its attached too and know for sure whether it really is the HDD or not.

The current pending sectors parameter is high according to SMART, which corresponds to unstable or weak sectors. Also the Raw Read Error parameter at the time of taking the screenshot was 41k and now its 80k.

Anyway I tried it on two different machines, and a Trascend eSata case and the results were the same.

Also as I said earlier the ddrecue transfer rate initially was very good and slowed down only when it reached the error areas. I had pasted the ddrescue output in my previous post.

 

[blr_p]

The current pending sectors parameter is high according to SMART, which corresponds to unstable or weak sectors. Also the Raw Read Error parameter at the time of taking the screenshot was 41k and now its 80k.

And no where are those parameters flagged by any warnings whatsoever.

Anyway I tried it on two different machines, and a Trascend eSata case and the results were the same.

You still seem unwilling to run the error scan on that drive. why is that ?

Also do a benchmark test so we know what the avg transfer speeds are like.

I'd recommend another tool to also check that drive out, but if you're not interested we'll leave it at that.

 

[varkey]

And no where are those parameters flagged by any warnings whatsoever.

It may not be flagged, but a very high read error rate correspond to a drive problem? Also according to HDD Sentinel the drives health is only 52% as it contained some 1388 Weak Sectors.

You still seem unwilling to run the error scan on that drive. why is that ?

Also do a benchmark test so we know what the avg transfer speeds are like.

I'd recommend another tool to also check that drive out, but if you're not interested we'll leave it at that.

I tried doing an error scan in HD Tune and left it for about two hours, but didnt reach anywhere. It was going very slow.

I also tried the Western Digital Diagnostic Tool, both the short and long test failed. Tried the Short and Extended Self Test using smartmontools in Ubuntu, both failed, the extended test failed at about 90% remaining.

From the above I think its somewhat certain that the issue is with the drive, isn't it?

 

[blr_p]

It may not be flagged, but a very high read error rate correspond to a drive problem? Also according to HDD Sentinel the drives health is only 52% as it contained some 1388 Weak Sectors.

How do you know its high

100 or 1000 or 10000 or more.

What is a high number in this context ?

I tried doing an error scan in HD Tune and left it for about two hours, but didnt reach anywhere. It was going very slow.

ok

I also tried the Western Digital Diagnostic Tool, both the short and long test failed. Tried the Short and Extended Self Test using smartmontools in Ubuntu, both failed, the extended test failed at about 90% remaining.

From the above I think its somewhat certain that the issue is with the drive, isn't it?

All the above was not present in your first post, had you mentioned it then things would have been more clear. But it indicates there was a problem with that HDD on that particular rig. If you were successful duplicating those tests on another rig then I would agree that the HDD is bad.

Insitead a shorter way would be when you put your new drive in and see whether it makes a difference. If the new one works in the same rig then you made your case.

Its interesting to see that even tho SMART values appear to be almost normal that the drive isn't altogether ok

 

[varkey]

How do you know its high

100 or 1000 or 10000 or more.

What is a high number in this context ?

80000 seemed high to me. lol

All the above was not present in your first post, had you mentioned it then things would have been more clear. But it indicates there was a problem with that HDD on that particular rig. If you were successful duplicating those tests on another rig then I would agree that the HDD is bad.

I never considered the situation where the drive was not the culprit, as I had already tried it in two machines, external case etc. So I didn't think it was necessary to include that info in my first post.

My only thought when creating this thread was to find a way to recover the data somehow.

Its interesting to see that even tho SMART values appear to be almost normal that the drive isn't altogether ok

I didn't find the SMART info normal, RAW Read Error, Write Error, Current Pending Sectors, all these should ideally be zero right? It is zero for my new drive and two other WD drives.

 

[blr_p]

My only thought when creating this thread was to find a way to recover the data somehow.

Agreed

I didn't find the SMART info normal, RAW Read Error, Write Error, Current Pending Sectors, all these should ideally be zero right? It is zero for my new drive and two other WD drives.

The one in bold is the only one that is suspicious. It means those sectors had trouble but the drive did not reallocate them as your count was zero. So maybe they go down with time or not.

Your write error is 71, is that bad ? no idea

Your read error rate is vendor specific acc to wiki RAW read happens to be WD specific, i don't have that on any of my drives.

What is true is you noticed it became very slow all of a sudden, that was your only indicator. Next time this similar topic comes up it would be good to see transfer graphs on the benchmark. Thats usually a good sign something is up.

So how is the new drive on this rig working then ?

Trust you've installed it already.

 

[varkey]

RAW Read Error value started increasing rapidly which got me worried. It was 40k at the time I created this post and increased to 80k. The absolute integer value may not be of good use, but the rapid increase in the value does indicate an issue, doesn't it?

It became really slow, a simple directory listing will take 4-5 seconds.

The drive was in my download rig / NAS actually, an Atom box. Its back to normal working fine.