Health 26% - False alarm?

[Pluto46]

Hi,

Starting a few days ago, Hard Disk Sentinel flagged a warning about one of several 3GB hard drives on my home computer. According to the log the following had happened (latest event first)

03.11.2016 21.28.43 #5 Reallocated Sectors Count 144 -> 160

03.11.2016 21.53.43 #5 Reallocated Sectors Count 88 -> 144

03.11.2016 21.48.43 #5 Reallocated Sectors Count 8 -> 88

03.11.2016 21.43.43 #5 Reallocated Sectors Count 0 -> 8

Problems also occurred between the communication of the disk and the host 93 times

Health went from 100 to 26 on the 3rd and 160 bad sectors were moved

I checked the ingoing connections of all hard drives, and I've run two short self-tests and one extended 5-hour self-test. None of them found any problems.

I tried uninstalling and reinstalling Hard Disk Sentinel using Revo Uninstaller Pro to get rid of everything, thinking that would reset the readings, but it didn't. They're still the same.

Is it possible to reset the reading on this hard drive, and start from scratch again?

Thanks for any help!

Pluto46

[hdsentinel]

Thanks for your message.

No, I can confirm there is no false alarm.
And no, I can confirm that not Hard Disk Sentinel "flagged" the drive.

The hard disk itself detected the problems - and Hard Disk Sentinel read them, interpreted, displayed them exactly to increase the attention about the change: the degradation and allow time to prepare (for example with a backup).

Uninstallation / reinstallatin (even if done by any "magic" tool) does not affect this of course - because the status is read from the hard disk itself. The problems detected and logged by the hard disk (and displayed by Hard Disk Sentinel) completely independent from any software, application, OS - so a "clean" reinstallation does not clear these of course.


In general about the problems detected by your drive: the "bad sectors" reported in the text description are no longer used by the hard disk:
they are already reallocated. So the spare area is used for all reads and writes targeting those bad sectors. This means that disk surface tests (even the tests in Hard Disk Sentinel) does not access those sectors, but tests the remaining data area and the spare area.
This is good, as this way you can be sure that the original (bad) area does not contain important data and can't risk data loss.
This is why manufacturers allow some (or even more) of them - and offer replacement only when their count reaches an error-level threshold.

The problem is that re-allocations may risk data and cause system instability - and many times hard disks never reach the error-level threshold: they fail (or "just" complete data loss occur) long before that.
So having high number of reallocated sectors can be risky - as more problems may be expected with higher chance.

By using the surface tests, you can verify if the currently used data area is error-free, there are no further errors reported (no weak, damaged sectors, no further problems).

For more information about these "bad sectors", please click on the "?" next to the text description
(red area describing the issues) and check
http://www.hdsentinel.com/faq.php#health and
http://www.hdsentinel.com/faq.php#tests

So yes, it is very good idea to perform tests - to reveal any possible further problems, or confirm that the status is stable, all problems are now fixed.
If this happens, then yes, it is good idea to clear the error counters: to acknowledge problems related to the past and be notified about possible new issues and degradations (if there will be). This is possible with Hard Disk Sentinel - just not with uninstallation.

This is exactly described in Support -> Frequently Asked Questions -> How to repair hard disk drive? How to eliminate displayed hard disk problems?
http://www.hdsentinel.com/faq_repair_ha ... _drive.php

which describes the way to perform testing - and then use the Offset values on the S.M.A.R.T. page to clear the bad sector counter, remove the displayed problems and restore the health % value.

Additionally, Support -> Knowledge base -> Hard disk cases pages describe the most common problems, how to fix/improve the situation.
For example, http://www.hdsentinel.com/hard_disk_cas ... _error.php
shows the communication problems and again the use of "Offset" field to correct the detected errors.

If you use Report menu -> Send test report to developer option, I can check the actual, complete status of the hard disk drive and can send step-by-step instructions about how to clear the error counters, how to restore the hard disk drive health - and then Hard Disk Sentinel will show only possible new problems.

[Pluto46]

Hi, thanks for your answer. I didn't realize it was the hard drive itself that was giving the warning.

Here's the result of the surface read test

20161106_R_ST3000DM001-1CH166_W1F3GA5X_CC27-surface-full.jpg (564.73 KiB) Viewed 19317 times

and

20161106_R_ST3000DM001-1CH166_W1F3GA5X_CC27-surface-full-st.jpg (509.68 KiB) Viewed 19317 times

I've also sent you a report as requested - hope it's helpful.

[Pluto46]

Health down a percent today, to 25.

I've done some of the short tests in SeaTools (it's a 3GB Seagate hard drive)

S.M.A.R.T. test pass 100%
Short drive self-test pass100%
Short generic pass 100%

Currently on the long generic test, takes about 6 hours apparently so I'll report back later.

[Pluto46]

just a quickie to say that the Seatool's long generic test was completed without incident and a score of 100%.

[hdsentinel]

Yes, this is completely normal and expected.

The tools designed for completely different purpose than Hard Disk Sentinel.
Instead of showing possible issues, degradations, they designed to hide/minimize problems and minimize warranty requests.
Please check: http://www.hdsentinel.com/faq.php#difference

Until the error-level threshold is not reached, the tool is designed to show "100% perfect" status.
This is generally what you received from the test in Hard Disk Sentinel: as currently the status seems stable, there are no problems, just some minor performance degradations (on the beginning and the middle of the disk surface).
These tools only show problems when the error-level threshold reached (when we can ask for warranty replacement) - just the questions are: do we have the drive working at that point, will we experience data corruption / data loss before reaching that point?

As you can see your other drives of the same model has no problems at all (they're still 100%) - but this one drive is producing more and more issues.

Ps. the Disk menu -> Short self test / extended self test functions are the same as the hardware self tests available in the tools provided by the manufacturer.

[Pluto46]

Thanks for taking all the trouble, hdsentinal, to explain things and for pointing to the right stuff for me to read up on. I'll devote time to it all in the coming days. I'd still like to know though how to reset (if poss) the readings (168 bad sectors, 93 communication failures), and start from scratch again.

best,
Pluto46

[hdsentinel]

To clear the error counters in Hard Disk Sentinel, please do the following:

1) open the S.M.A.R.T. page of the appropriate hard disk.
2) locate the "5 Reallocated Sectors Count" attribute. In the "Offset" column, click on the 0 (between the + and - ) and acknowledge the warning message.
3) specify the amount of bad sectors with negative sign:
-168
to clear the error counter.

Soon the status and the reported health should improve, the problems will be removed from the text description.

Then proceed similarly to clear the error data communication errors:

4) locate the "199 Ultra ATA CRC Error Count" attribute. In the "Offset" column, click on the 0 (between the + and - ) and acknowledge the warning message.
3) specify the amount of communication errors with negative sign:
-93
to clear the error counter.


Then (as these issues are already fixed according the tests, the hard disk status is now stable - so they should not appear there) Hard Disk Sentinel will not show them. The status will improve, the health should be 100% again and then only possible new hard disk problems will be reported.

[Pluto46]

Brilliant, thanks a ton for your help so far. Invaluable.

I offset the reallocated sectors count and the spin up times a couple of days ago, and did the Ultra ATA CRC error count today. As regards the first two, I haven't noticed any changes in two days, apart from a one point change in the spin up times from the number I offset (93) to the current value today of 94.

I suppose the most important indicator here anyway is the reallocated sectors count, though.

I tried doing a Write + Read test on the drive, but was prevented by error 5: Access denied, which I've learned from reading other entries here means the drive is being used by other programs or processes. I had a look at all running processes in Task Manager, but - being a total amateur - I couldn't see which programs or processes I needed to terminate to allow the test to go ahead. Is it possible to unmount the disk via Control Panel > Computer Management > Disk Management for example, and then do the test?

best,
Pluto46 (learning curve rising slowly )

[hdsentinel]

Me thanks so much - glad if I could help

The "spin up time" value may fluctuate, the number 93-94 does not require correction/attention now.
The "spin retry count" would suggest problems (if any) but according the report, this was perfect, there were no spin up errors/problems.
So you may only need to check/correct the other attributes (5 Reallocated Sectors Count and 199 Ultra ATA CRC Error Count only).

Yes, the reallocated sector count is interesting now. The best would be if it would stay constant, so no new problems should be detected, reported.


> I tried doing a Write + Read test on the drive, but was prevented by error 5: Access denied,
> which I've learned from reading other entries here means the drive is being used by other programs or processes.

Yes, that's true: some software / application (or the OS itself) may have open files/folders on the logical drive on that hard disk and this prevents write testing to be performed. Windows indexing service, antivirus software, automatic defragmentation etc... may cause this.

> Is it possible to unmount the disk via Control Panel > Computer Management > Disk Management for example, and then do the test?

Yes, this can help: if you open Control Panel > Computer Management > Disk Management and there manually delete the partition first, it should work.
Windows may also refuse to delete the partition if any files/folders opened on the volume - but may perform on next restart.
Please note that the Write + Read test is destructive, it clears all data on the selected hard disk drive - so use carefully and after backup only (which is required if you attempt to delete the partition before starting the test).

[Pluto46]

Thanks, I'll try the method you suggest later today or tomorrow. Everything's backed up in three different places (another hard drive, a Qnap backup system, and a cloud subscription).

[Pluto46]

So I deleted the volume on the hard drive, and set Hard Drive Sentinel up to do a Read + Write test over night. For some reason it stopped half way through (50.36%) when the disk stopped responding. There are two errors in the bottom bar of the test:

Read + Write test 2016.11.11.jpg (263.19 KiB) Viewed 19196 times

and health has declined to 5% with 3840 bad sectors!

The vendor said I could return it and get my money back, so I'll probably return it today. I've restarted the Read/Write test again, though there's probably not much point at this stage. I've obviously got a denial problem. Only a couple of weeks ago, the power supply unit conked out, just two years old.

anyway, if anything emerges from the test being run as I speak, I'll let you know.

thanks again for being generous with your time. Much appreciated!

[hdsentinel]

Thanks for the image and the information. Me thanks too, glad to hear if I could help

Yes, the situation is relatively common.
The drive already started showing degradation when you noticed the 26% health - and that point problems were otherwise remain unnoticeable.
But yes, some further using, the intensive testing revealed more and more problems - and exactly as you see, when such new problems found, the drive (and the software attempting to use it) may stops responding, risking both data loss and generic system instability.

I'm sorry about the situation - but it is good if the seller of the hard disk offers refund in this case.

Seems what Hard Disk Sentinel displayed was not a false alarm - but (as we'd expect) an early warning which allowed time to prepare and backup important data before complete failure.