WRITE test and READ test for a WD HDD

How, what, where and why - when using the software.
shawn
Posts: 14
Joined: 2021.02.17. 06:53

WRITE test and READ test for a WD HDD

Post by shawn »

I recently conducted WRITE and READ tests on my 3TB WD HDD to assess its health. Since HDS issued a 2% HDD health warning, I decided to perform a WRITE test first. The initial 2/3 of the test proceeded quickly, but the last 1/3 was notably slower, with the speed dropping from around 100MB/s to 30MB/s. This section also displayed numerous dark green sectors. Surprisingly, the health increased from 2% to 3% after the WRITE test. What a “significant” improvement. LOL

Out of curiosity, I proceeded to conduct another READ test. This time, there was no speed drop, and all sectors appeared light green. However, the health status remained at 3%.

I have a few questions regarding these results:
1. Which SMART attribute contributes to the 3% health rating? Is it the "Off-Line Uncorrectable Sector Count"? The current count is 198, and the Worst count is listed as 200.
2. Based on my understanding, the reading indicates that there are 200 spare sectors available for remapping bad sectors. With 198 bad sectors detected, only 2 spare sectors remain for this HDD. If I were to perform another WRITE surface test or perform a FULL format in Windows, would Windows try utilize those 198 bad sectors, or they are marked as “bad sectors” permanently?
3. The "5. Reallocated Sectors Count" displays a value of 447, but both the Value and Worst columns show 169. What is the significance of the Value and Worst columns in this context?
4. Does the READ test identify and trigger remapping of bad sectors? It seems that READ tests are much less likely to uncover problems compared to WRITE tests.
I appreciate your insights and clarification on these matters. Thank you for your assistance.

HDS.3T.4.Test.01.png
HDS.3T.4.Test.01.png (47.99 KiB) Viewed 1350 times
HDS.3T.4.Test.02.png
HDS.3T.4.Test.02.png (107.72 KiB) Viewed 1350 times
HDS.3T.4.Test.03.png
HDS.3T.4.Test.03.png (212.44 KiB) Viewed 1350 times
HDS.3T.4.Test.04.png
HDS.3T.4.Test.04.png (22.8 KiB) Viewed 1350 times
HDS.3T.4.Test.05.png
HDS.3T.4.Test.05.png (255.95 KiB) Viewed 1350 times
User avatar
hdsentinel
Site Admin
Posts: 3010
Joined: 2008.07.27. 17:00
Location: Hungary
Contact:

Re: WRITE test and READ test for a WD HDD

Post by hdsentinel »

First, generally I'd not expect anything from a hard disk drive with so low Health % value. It is expected to see new problems with time on the drive.
Always interesting to see how a such drive performs - but I'd no longer recommend for any real storage.

> Surprisingly, the health increased from 2% to 3% after the WRITE test. What a “significant” improvement. LOL

Yes, of course, this is completely normal and expected situation: as you see after the test completed, 248 pending sectors fixed, so the test could successfully force the hard disk drive to stabilize many problems. During this, 190 new bad sectors reallocated, so they will be surely never re-used.
This is why the Health % is also improved (on a such drive with low Health % we can't expect big improvement ;)

But in such situation, yes, exactly as you saw, a big part of the disk surface is MUCH slower than expected (represented by the darker green blocks) and we can expect new that sectors will fail with time. As the test stabilized the issue, it is possible that it may take some time: sometimes days, weeks or more - but in many cases even a following READ test will immediately reveal such issues.


> Out of curiosity, I proceeded to conduct another READ test. This time, there was no speed drop,

Yes. Ideally things should work exactly this way: the problems fixed, the drive itself stabilized: instead of the original sectors, the spare area used now. So ideally we'd expect that the drive is working as "error-free", perfect.


> and all sectors appeared light green. However, the health status remained at 3%.

Yes, exactly as expected. As the problems are recorded by the drive itself, the Health % will not improve back automatically until you manually acknowledge them: after the errors fixed, you can manually clear the error counters to "ignore" these issues and be notified about possible new problems only. This is common practice, but usually recommended on drives with generally higher Health %.

This solution (disk testing to stabilize/fix isues and then clear the error counters to restore Health %) is described at Support -> Frequently Asked Questions -> How to repair hard disk drive? How to eliminate displayed hard disk problems?
https://www.hdsentinel.com/faq_repair_hard_disk_drive.php


> 1. Which SMART attribute contributes to the 3% health rating? Is it the "Off-Line Uncorrectable Sector Count"?

Yes and many others too which can indicate problems with the disk drive. Attribute 5 Reallocated Sectors Count, 196 Reallocation Event Count, 197 Current Pending Sector Count and many-many others.

The Help describes how the Health % calculated by the first version of Hard Disk Sentinel (which attributes used and how). Things are much more sophisticated since that so the whole procedure can't be explained but you'll see the idea: to detect/reveal real error count and generally disk drive issues which are affecting MORE attributes. Seeing/checking one single attribute is usually a WRONG approach.


> The current count is 198, and the Worst count is listed as 200.

The Data field shows the actual count (198). The Worst field is not relevant at all in this case, it is never related to the actual number of problems.


> Based on my understanding, the reading indicates that there are 200 spare sectors available for remapping bad sectors.

No, this is absolutely wrong.
There are no 200 spare sectors available. The Worst = 200 does not mean that.

First of all, not this but the 5 Reallocated Sectors Count attribute is more important in this case. It shows 447 bad sectors which are already reallocated, the spare sector used instead of these sectors.
The Value field is decreasing based on this amount - and its count, the original value and the "speed" of the decrease can indicate the size of the spare area which is now approximately 3000 sectors.
(The Value field starts from 200, decreasing as new bad sectors detected. It is now 169 because of the 447 bad sectors reported there)

Please refer to www.hdsentinel.com/smart page which explains the general relation between the real error counters (in the Data column) and the Value / Worts / Threshold fields - and this explains why checking the Value / Threshold combination alone is a wrong approach.

> With 198 bad sectors detected, only 2 spare sectors remain for this HDD.

No, sorry, but this is completely wrong assumption. There are still more spare sectors left (approx. 2400) but the drive will show 0% FAILURE PREDICTED status even before that (while there are still spare sectors).


> If I were to perform another WRITE surface test or perform a FULL format in Windows, would Windows try
> utilize those 198 bad sectors, or they are marked as “bad sectors” permanently?

The currently detected and reported BAD SECTORS will be never re-used. You can do read test, write test, Reinitialise Disk Surface test, FULL format in Windows, create/delete partitions - the BAD SECTORS will be never re-used again. This is why some (very few) bad sectors are acceptable (really) as described at
https://www.hdsentinel.com/faq.php#health
(and in the Help too, exactly to give more information/explanation)

The disk test (exactly as should) forced the hard disk drive to detect and stabilize the problems - so then ideally the drive should work as error-free.
In real life, lower Health % means higher chances for possible further problems and data corruption / data loss - this is why it is important to note the
- real error count
- possible new issues (even minor) which can quickly lead to data corruption /data loss


> 3. The "5. Reallocated Sectors Count" displays a value of 447, but both the Value and Worst columns show 169.
> What is the significance of the Value and Worst columns in this context?

Please refer to the link above which explains that the Value field is calculated by the drive internally based on the actual error count (447). Decreasing the Value field means degradation of the drive and if it drops below the Threshold, then the drive is considered as FAILING. Then Hard Disk Sentinel will show 0 % Health and this is the point when the BIOS may also show S.M.A.R.T. failure predicted, Windows (or Linux) may show drive failure warning. Usually then no new OS installation is possible.
(note: the Worst field shows the lowest Value ever recorded as in some rare cases the Value may increase back for other problems).

In real life, most drives fail long before reaching this point. This is why it is wrong approach to verify Value/Threshold pairs - but better to focus on real, actual number of problems and degradations.


> 4. Does the READ test identify and trigger remapping of bad sectors?
> It seems that READ tests are much less likely to uncover problems compared to WRITE tests.

It is absolutely correct and expected.
Please refer to the description of the tests: the READ test designed to REVEAL (identify) the problems but it does not force the drive to reallocate bad sectors.
The WRITE type tests (Disk menu -> Surface test -> Write test and even more intensive Disk menu -> Surface test -> Reinitialise Disk Surface test) designed to stabilize the problems.

If you use Report menu -> Send test report to developer option, I'd be happy to check the actual status of the drive. If you run additional test(s) and/or try to use the drive and later send new report(s) it is possible to examine changes, follow how the values degrade further - and when the drive may fail completely. Investigation of such drives always help development (eg. by adjusting reporting of errors, calculation of Health, changing disk tests).
shawn
Posts: 14
Joined: 2021.02.17. 06:53

Re: WRITE test and READ test for a WD HDD

Post by shawn »

I really appreciate your detailed answers! You are the best!
Post Reply