Hard Disk Monitoring

Hi,

I recently had an issue with my less than 1 month old 8TB HDD (Seagate ST8000AS0002).
I use this HDD as an offline mirror.

After a successfull backup, using the Windows "Safely Remove Hardware and Eject Media" and unplugging the disk, it made strange noises (as if the head was scratching the spinning platters while they were slowing down).
After that the HDD seemed dead and became unreachable. I tried many reboots / plugging / unplugging and at one point I could read, during boot, that it said "SMART failure". Hard Disk Sentinel just crashed (froze) when the HDD was inserted.

Next day I wanted to start the warrenty procedure, but I decided to try one more time.
Miraculously it suddenly worked again, as if nothing ever happened (no SMART failure, 100% health in Hard Disk Sentinel).

I ran a short self test, an extended self test, a SMART test. Nothing found...
Finally, to make sure that crashing head didn't destroy any sectors, I decided to do a Read-Write-Read Surface Test.

So it started and after calculating a bit, I figured out it would take a little more than 5 days.
Then 5 days later, after checking regularly, I suddenly noticed that around 90-95%, my PC had a blue screen (event logs don't mention why :/ ).
After cursing for about an hour, I went looking for logs in HD Sentinel to see how far it got exactly and if it found any bad sectors in those last few hours that I've missed.

I could not find such a log file...

So my questions:
Does a log file of a surface test exists?
Where can I find it?

Thanks for your message.

> Hard Disk Sentinel just crashed (froze) when the HDD was inserted.

Yes, this can happen if the hard disk is not responding at all. Hard Disk Sentinel is attempting to access the drive and collect the last known status information but if the drive may longer time to respond (or not responding at all) then the software trying to access it also seems not responding.
This is somehow similar as a RAM test software would even not show problems but crash when trying to test a faulty memory module.

No, I'm afraid currently no such logs / status information saved automatically, only if the user saves it any time (during or after the test).
Such feature is already planned and will be available in the next 5.00 version, which will periodcially and automatically save test status / log, exactly to help identifying problems or at least progress information when such issue happens.

May I ask what is the status of the hard disk now, after the disk test (at least partially) tested the hard disk drive?
If the hard disk is still available and working (detected by the system and Hard Disk Sentinel) can you pleas use Report menu -> Send test report to developer option?
The current status may help to identify problems (at least minor ones) and can help to find out how to improve things.

Maybe to save time and verify if the data sectors are all readable, accessible and there is no damaged / bad or slower areas, I'd start Disk menu -> Surface test -> Read test first. This saves time, as on a 8 TB hard disk, this complete surface test would take "just" about 16 hours....

The health status is still at 100%
I did notice that the log of the HDD now contains:
23/02/2016 7:31:51,#199 Ultra ATA CRC Error Count 0 -> 910

This time is actually when it was running an extended self test (which completed successfully strange enough)

In meanwhile I have "completed" the surface test by running a "Sequential Backward Test" (thanks for this option!) and rescanning (read-write-read) the last 10%.
Again it found nothing, so I guess the HDD has no damage at all.

Would it make sense to replace the SATA cable?

This sounds interesting.

Yes, generally SATA cable and/or its connection are the most frequent cause of such Ultra ATA CRC Errors (data communication errors).
But these usually happen during data communication between the hard disk and the controller.
But the extended self test runs "inside" the hard disk itself, without actual data communication, so it is weird if the counter increased during the extended self-test.
Usually Disk menu -> Surface test functions would reveal such errors, as these perform intensive data communication between the hard disk and the controller itself.

Anyway, these problems are independent from the hard disk, so these do not decrease of the health. And as they do not affect the actual operation of the hard disk, yes, the short / extended self tests usually complete without errors.

As the test completed successfully (and by the disk surface test) you verified that generally all sectors are perfect, so there is no problem with reading/writing them.

But personally I'd still worry about the original problem you wrote (strange noise and un-detected / un-responsive drive).

Personally I'd yes, verify and replace the SATA cable and also would check the power cable / connection too.
If possible (and if now used) I'd avoid power splitter / 4-pin -> SATA converters and similar devices.
Then I'd try to perform some power cycles, safe removal and disconnect / reconnect cycles as you tried previously, just to verify if the same behaviour happens again - or not, and the hard disk would be really safe to use.

Ideally there will be no problems and also the Ultra ATA CRC Error count will also remain stable, should not increase.
Then it is possible to clear the reported problems: then Hard Disk Sentinel would report only new issues (if there will be), as described at
http://www.hdsentinel.com/hard_disk_cas ... _error.php

Did you listen for the heads parking before actually removing power? Some drives do make scary noises whenever an emergency retract happens (pulling power before it's parked causes the spindle driver to forcefully retract to parking zone before the air pressure drop in the drive causes the head to hit the spinning platter), this can damage heads and the arms due to vibration and the force used.

It seems like it was the power cable...

I discovered that the recognizing of HDDs also had issues with other HDDs...
So I replaced the SATA cable first... no difference...
Then I replaced the switched the power cable (it indeed has a splitter, this is for the HDD bay which has an HDD LED)... this seems to have helped...

I've not had any issues since (as I use the HDD as offline backup, I do not test it 24/7 though )

Regarding the strange noise:
That indeed worries me a bit too... I was expecting the surface test to at least find something... but nothing...
Not sure if I can ever find the cause to this, but what you describe could have happened...

I safely removed the hardware and then 1-2 seconds later I pulled the HDD from the bay (as I've done 100s of times with other HDDs already)
Could it be that some HDDs require more time than others to park there head after a safe removal?

Yes, power cable adapters / splitters are relatively common cause of troubles...

Yes, some drives may require more time to spin down completely. This is especially true for large capacity hard disk drives and drives generally with high(er) number of heads / disk platters.