Checking 'failed' hard disk

[SliMat]

Hi All

 

I have bought a 'failed' 6Tb Western Digital Red SATA disk, very cheap!! The seller advertised it as faulty, saying it was very slow booting up into Windows 10 - he said it would take up to 10 times longer than it should... but said that there was never any data loss or corruption and that CrystalDisk reports it as healthy and is unable to identify any errors.

 

The reason I bought it is that it has over 1yr left on the manufacturers warranty... I thought I would run 2 cycles of preclear to see if it shows any signs of failure and to see what the transfer rate shows as, before using it... obviously if it is failing I need to be able to prove this to get a warranty return started.

 

So, preclear has been running for nearly 4 hours and the log, so far, shows;

 

 

I wondered whether anyone can advise what sort of transfer rates should be seen as 'normal' and what to look for to see if there is a problem which I can prove to the manufacturer?

 

Any thoughts appreciated.

 

Thanks

Edited January 9 by SliMat

[trurl]

Looks normal to me. You might also run DiskSpeed docker on it.

[SliMat]

Thanks - I'll let it run a couple of cycles and see that the speed stays at about 200MB/s - hopefully if it can sustain 2 clear cycles it should be OK to add to the array. 

 

I havent seen DiskSpeed docker, so will have a look at that and try it too... 

 

I'm not sure what sort of transfer rate is normal, but 200-240 seems pretty OK from memory.

Edited January 9 by SliMat

[SliMat]

3 hours ago, trurl said:

Looks normal to me. You might also run DiskSpeed docker on it.

 

I just checked again and its still progressing, although has slowed quite a lot from the early speeds... one thing I have noticed, which was the same on the first check above but I didnt spot, or include it in the screen shot... the CRC Error Count is listed as 17130... I am not entirely sure what this means, but am guessing 17130 errors is not a good thing? Although my understanding is that this was probably caused by a connection error in the last owners machine. I think I'm right in saying that as long as this doesnt increase in my installation it will mean everything is good?

 

Edited January 10 by SliMat
updated image

[trurl]

47 minutes ago, SliMat said:

has slowed quite a lot from the early speeds

Longer outer tracks have more data than shorter inner tracks for the same RPM, and it starts with the outer, so normal for the data rate to decrease somewhat as it progresses.

 

46 minutes ago, SliMat said:

probably caused by a connection error in the last owners machine

Probably explains the slow performance they were getting.

[itimpi]

9 hours ago, SliMat said:

I think I'm right in saying that as long as this doesnt increase in my installation it will mean everything is good?

 

Yes (as this count never resets).

[SliMat]

So, preclear has finished - do you think this is enough to prove that this dosk is OK to add into my array? I've pasted the results below and as far as I can tell it all looks good and I think the problem the last owner suffered was because of a cable/controller problem rather than a disk problem...

 

 

 

The top line (sdh) relates to the 6Tb drive

Edited January 11 by SliMat

[trurl]

3 hours ago, SliMat said:

OK to add into my array?

As long as everything else is working well, if this single disk does fail it shouldn't be much trouble to recover from anyway.

 

I seldom preclear, but I only introduce one disk at a time into my array, usually with months or more before another disk is used. And I have good (enough) backups.

[SliMat]

9 hours ago, trurl said:

I only introduce one disk at a time into my array, usually with months or more before another disk is used. And I have good (enough) backups.

 

Good advice... and I do the same. The server I want to add this to has 6 x 6Tb Purple disks and has been running for 2 months now... one observation from the DiskSpeed screen above... I noticed on the server with 6 x 6Tb disks the speed trace of all the disks is identical... which is reassuring... however the screen shot above is from a server at home which has 4 identical 4Tb RED disks plus a few unassigned devices, such as the 6Tb disk I am testing. 

 

However on the trace above, the 3 x 4Tb RED disks in the array have identical traces, BUT the exact same 4Tb RED disk which is assigned parity disk is the blue line immediately below the trace for the 3 array disks. Is it normal for the parity drive to DiskSpeed slower, or should I be concerned about imminent failure?

 

I've never used the DiskSpeed docker until your recommendation above... so am not sure what to expect.

 

Thanks