Is My Hard Drive Still Healthy?

[_Husskeyy]

Hi all!

 

I have a pretty old Western Digital 2TB HDD (WD20EARX). I bought it in 2011, so it's now 11 years old.

 

Hard Disk Sentinel shows its health is somehow still at 100%, despite it being used so much, for so long. For comparison, my main system SSD, which has a power on time of 16x less than the HDD, already has health of 86%.

 

Screenshot here.

 

Can the 100% health reading for the old HDD really be trusted? This thing is well into the age category of "generally the time HDDs fail". Should I just go ahead and ignore the reading and replace the thing before I lose data? I don't wanna get a

replacement unnecessarily.

 

Thanks in advance for any advice provided. :)

[MistahHaskins]

I would try CrystalDiskInfo, tends to show accurate reads on drive health.

 

Also you should just buy one even if you don't need it, when you HDD is well into the failing stage of life you should always have your data backed up onto a newer much safer drive. Then you can use it until it dies as secondary storage or ditch the drive altogether.

[_Husskeyy]

Just now, MistahHaskins said:

I would try CrystalDiskInfo, tends to show accurate reads on drive health.

 

Also you should just buy one even if you don't need it, when you HDD is well into the failing stage of life you should always have your data backed up onto a newer much safer drive. Then you can use it until it dies as secondary storage or ditch the drive altogether.

That's true. It's worth the peace of mind.

 

Here's the CrystalDiskInfo stuff for it.

 

How does that look?

[MistahHaskins]

Looks normal

 

You should also try your manufactures software

https://support.wdc.com/downloads.aspx?DL=

 

Should be Western Digital Dashboard, I use that to check my drives health, because mine are all Western Digital

[_Husskeyy]

10 minutes ago, MistahHaskins said:

Looks normal

 

You should also try your manufactures software

https://support.wdc.com/downloads.aspx?DL=

 

Should be Western Digital Dashboard, I use that to check my drives health, because mine are all Western Digital

Ya think I can just keep going with the drive despite 11 years? Maybe it's not as weird as I think it is that the drive still has such good health. Perhaps cause it's never been a boot drive?

 

Thanks! I didn't know about the Dashboard until just now. It says the drive health is "Normal".

 

Are those CrystalDiskInfo readings above the threshold values something of a concern?

[mariushm]

Mechanical hard drives store the data as magnetic fields on the platters inside the drive. 

As long as the motor doesn't fail, or the read/write heads don't hit the platter surfaces, there's not much reason for failure. 

Sometimes, super small areas of the platters can start to lose their magnetic properties, so the heads have harder time settings bits to 0 or 1, and that's when a "sector" (a group of bytes) is considered bad and the decision is made to not use it anymore and the contents of that sector is reallocated to another area of the platters, in a hidden area kept as reserve for precisely this reason. 

You can see how many such areas are detected by your hard drive in the "Reallocated Sector Count" entry in SMART : yours is 0, so the drive is fine. 

 

So, the drive is healthy.

 

SSDs work a bit differently. They don't have platters, nothing moving inside, they have memory chips. 

However, in order to squeeze so many bits into the memory chips, to make them cheap enough to produce, some decisions were taken that result in the memory chips "wearing out" over time. 

Basically, think of it like this... the data in memory chips are arranged like pages in a notebook, and each notebook page has a few lines where you can write stuff. 

The memory chips are designed in such a way that you can write a line of text in a page, but you can't erase that line or a single character in that line. If you want to erase something, you have to erase the whole page and each time you erase the page, there's a bit of "wear and tear" on this page. 

Once you erase the page too many times, the page becomes so "thin" you can't actually write stuff on the page anymore, the page becomes read only.

 

So the SSD controller manages these erases as best as it can, when you want to change a line the controller reads the line, makes the changes and writes the line in a totally different page and makes an internal note saying "line 100 on page 5 can be erased" and another internal note saying "if anyone asks for line 100 on page 5, I've moved it on page 500 line 123" ... and the controller delays erasing the whole page as much as possible because that hurts the page.

Only when most lines on that page are marked as "can be erased", or when pretty much all the other pages have just as many lines that are marked as erasable, the controller moves to erase the whole page. It does this by copying the still valid lines to other pages and then erases the whole page, wearing it out a bit, and now this empty page gets back in circulation, the lines on the page become available for writing on them. 

 

So what the SSD manufacturers do is hide some amount of memory from you, it's like selling you a book with 900 pages but the book actually has 1000 pages and 100 pages are hidden from you. 

As pages are worn out, the SSD controller replaces those worn out pages with some of those hidden pages and extends the life of the SSD, but eventually you're running out of those reserve pages as well. 

 

So the health on your SSD basically just tells you the average amount of wear on the pages ... for example, if each page can be erased a minimum of 1000 times and the average erase count is 100, then the SSD would tell you that 10% of drive's health is gone and that now the Health is at 90%  because 100 is 10% of 1000.

But there's nothing bad, nothing is damaged on the SSD, it's still perfectly fine and can hold data reliably,  it's just an indicator that as you get close to 5-10%, you should consider replacing the SSD because the drive would no longer be able to write stuff reliably in the memory chips. 

Edited February 10, 2022 by mariushm
minor corrections and improving of text clarity to make it easier to read hopefully

[_Husskeyy]

9 hours ago, mariushm said:

Mechanical hard drives store the data as magnetic fields on the platters inside the drive. 

As long as the motor doesn't fail, or the read/write heads don't hit the platter surfaces, there's not much reason for failure. 

Sometimes, super small areas of the platters can start to lose their magnetic properties, so the heads have harder time settings bits to 0 or 1, and that's when a "sector" (a group of bytes) is considered bad and the decision is made to not use it anymore and the contents of that sector is reallocated to another area of the platters, in a hidden area kept as reserve for precisely this reason. 

You can see how many such areas are detected by your hard drive in the "Reallocated Sector Count" entry in SMART : yours is 0, so the drive is fine. 

 

So, the drive is healthy.

 

SSDs work a bit differently. They don't have platters, nothing moving inside, they have memory chips. 

However, in order to squeeze so many bits into the memory chips, to make them cheap enough to produce, some decisions were taken that result in the memory chips "wearing out" over time. 

Basically, think of it like this... the data in memory chips are arranged like pages in a notebook, and each notebook has a few lines where you can write stuff. 

The memory chips are designed in such a way that you can write a line of text, but you can't erase that line or a single character in that line. If you want to erase something, you have to erase the whole page and each time you erase the page, there's a bit of "wear and tear" on this page. 

Once you erase the page too many times, the page becomes so "thin" you can't actually write stuff on the page anymore, the page becomes read only.

 

So the SSD controller manages these erases as best as it can, when you want to change a line the controller reads the line, makes the changes and writes the line in a totally different page and makes an internal note saying "line 100 on page 5 can be erased"  ... only when most lines on that page are marked as "can be erased", the controller then copies the lines with valid content in other pages and then erases the whole page, wearing it out a bit. 

 

So what the SSD manufacturers do is hide some amount of memory from you, it's like selling you a book with 900 pages but the book actually has 1000 pages and 100 pages are hidden from you. 

As pages are worn out, the SSD controller replaces those worn out pages with some of those hidden pages and extends the life of the SSD, but eventually you're running out of those reserve pages as well. 

 

So the health on your SSD basically just tells you the average amount of wear on the pages ... if each page can be erased 1000 times and the average erase count is 100, then the SSD would tell you the Health is at 90%  because 100 is 10% of 1000.

But there's nothing bad, nothing is damaged on the SSD, it's still perfectly fine and can hold data reliably,  it's just an indicator that as you get close to 5-10%, you should consider replacing the SSD because the drive would no longer be able to write stuff reliably in the memory chips. 

This is a fantastic explanation, and despite being as into tech as I am, you've still taught me some things I didn't know. Thank you for explaining things so thoroughly yet concisely.  Good to know both drives are still fine. Just to confirm, what you've said about SSDs can be applied to the NVMe type as well, right? Just cause they're PCIe-based doesn't mean the rules about data storage or health are different, correct? I assume the chips are still similar if not the same.

[MistahHaskins]

2 hours ago, _Husskeyy said:

This is a fantastic explanation, and despite being as into tech as I am, you've still taught me some things I didn't know. Thank you for explaining things so thoroughly yet concisely.  Good to know both drives are still fine. Just to confirm, what you've said about SSDs can be applied to the NVMe type as well, right? Just cause they're PCIe-based doesn't mean the rules about data storage or health are different, correct? I assume the chips are still similar if not the same.

That's true, and going back to your other statement, I would personally buy the secondary drive. Even if you don't need it, its a back storage, so I would move everything you NEED and cannot afford to lose onto that drive and stow it away until the original does eventually give.