SpinRite

[CrazyTeeka]

What if SMART just doesn't feel like sharing much on this drive?
Maybe the answer is to stop trusting SMART.

Screenshot 2023-11-15 164841.png (68.54 KiB) Viewed 30649 times

[fzabkar]

I think SMART reports logical sectors for some attributes and physical for others, and this varies between models.

As for those partial sectors, I would think that a drive would/should preemptively reallocate any weak sector if it can get one good read. This means that the next attempt at each logical sector should succeed.

[fzabkar]

Drive specifications...

This ST9500423AS drive has "AF" on the label (FW 0005DEM1):

https://www.harddrivesforsale.com/media ... 6_1_67.jpg

This one doesn't (FW 0002SDM1):

https://i.ebayimg.com/images/g/VEcAAOSw ... -l1600.jpg

This version of the product manual specifies "Bytes per sector (logical/physical) -- 512/4K":

https://www.seagate.com/files/www-conte ... 27937f.pdf

[CrazyTeeka]

There's no AF anywhere on the label on this drive. But I assume they don't put AF on the label unless its 4k logical/physical.

[fzabkar]

There's no AF anywhere on the label on this drive. But I assume they don't put AF on the label unless its 4k logical/physical.

I wouldn't assume that. I have noticed that manufacturers often prefer to sneak new tech into the marketplace so as not to alarm their customers. For example, I saw one tech site which reviewed a Seagate drive without mentioning that it was a helium model.

If you can dump the 512-byte Identify Device data, I could check whether the drive actually reports its physical sector size. CrystalDiskInfo is one tool that can do this.

Alternatively ...

Code: Select all

fsutil fsinfo ntfsinfo X:

where X: is your drive letter in Windows.

Samsung 860 Evo SSD

Code: Select all

NTFS Volume Serial Number :        0xnnnnnnnnnnnnnnnn
NTFS Version      :                3.1
LFS Version       :                2.0
Total Sectors     :                1,758,172,591  (838.4 GB)
Total Clusters    :                  219,771,573  (838.4 GB)
Free Clusters     :                  139,200,504  (531.0 GB)
Total Reserved Clusters :                  8,123  ( 31.7 MB)
Reserved For Storage Reserve :                 0  (  0.0 KB)
Bytes Per Sector  :                512
Bytes Per Physical Sector :        512
Bytes Per Cluster :                4096

Crucial MX500

Code: Select all

NTFS Volume Serial Number :        0xnnnnnnnnnnnnnnnn
NTFS Version      :                3.1
LFS Version       :                2.0
Total Sectors     :                1,758,172,591  (838.4 GB)
Total Clusters    :                  219,771,573  (838.4 GB)
Free Clusters     :                  149,969,486  (572.1 GB)
Total Reserved Clusters :                  4,895  ( 19.1 MB)
Reserved For Storage Reserve :                 0  (  0.0 KB)
Bytes Per Sector  :                512
Bytes Per Physical Sector :        4096
Bytes Per Cluster :                4096

[fzabkar]

@CrazyTeeka, can you examine those bad sectors after SpinRite "repaired" them? Would it be OK to upload the entire physical sector, assuming the data are not sensitive?

[CrazyTeeka]

@CrazyTeeka, can you examine those bad sectors after SpinRite "repaired" them? Would it be OK to upload the entire physical sector, assuming the data are not sensitive?

I'll do that on the next drive. This drive's already been wiped.

[Joep]

What if SMART just doesn't feel like sharing much on this drive?
Maybe the answer is to stop trusting SMART.

Screenshot 2023-11-15 164841.png

Yes, that's a typical answer from Steve G, half way between I don't know and a conspiracy theory. It's bullshit, and in my experience pending and reallocated are usually accurate regardless the drive brand.

He also contradicts himself, when SR6 was introduced he sort of promised SR would make SMART data useful (also mixed with some vague claims of drives hiding important stuff), reality is that he gets half of the SMART stats wrong. For example, SR hasn't got a clue about SMART stats as reported by Seagate drives.

When I pushed him for an answer how het arrived at this theory: ..

Writing to NAND flash memory requires the generation of high voltage, but reading requires low voltage. Before any NAND chip can write data into NAND flash memory it must internally generate a high voltage. This uses a “voltage pump” which takes time to charge up, and lower-end solid state storage does this much more slowly. Then, before reading, that high voltage must be dumped.

.. he basically brushed it off by telling me "experience". It's a load of mumbo jumbo and I explained him why. 9 out 10 times when he talks about NAND based drives it reveals a total LACK of experience and he's making stuff up as he goes ..

[fzabkar]

For example, SR hasn't got a clue about SMART stats as reported by Seagate drives.

Steve distrusts SMART because he doesn't understand it. As you say, the one time that he tried to make use of SMART, he made a complete mess of things. His mistake was to apply intuitive logic to a counterintuitive case, specifically a Seagate drive. Essentially, SpinRite took a perfectly healthy drive and just kept spewing out thousands of bogus errors.

https://www.hddoracle.com/viewtopic.php?p=21743#p21743

[fzabkar]

It would be interesting to hear what DeepSpar have to say about this "partial physical sector" anomaly.