Samsung SSD (870 evo 1tb) had multiple bad blocks/sectors, after a full format is now showing only good sectors. Is it still usable?

[Devilmoon]

Yesterday I embarked upon a journey because a game launcher started giving me disk i/o errors when trying to apply a patch. This was on an SSD I regularly use for some videogames.
From there I opened the Windows event viewer which showed me multiple errors related to disk, namely "This device, <ssd id>, has a bad block".
I then tried running `chkdsk` on the disk, which did not find any issues; however, relaunching it with `/f` and `/r` it found multiple bad blocks and appeared to be moving around data (from paths I haven't used in ~years).
While this was going on I also launched CrystalDiskInfo, however SMART was reporting good condition with a score of 98%.

After chkdsk finished, I tried relaunching the game that started this all and still got the same errors.

This being a Samsung SSD I also tried using their own utility for disk health (Samsung Magician), which reports the drive being in good health. The short scan/ short smart self-test options in the diagnostic scan menu also found no issues.
I then tried launching a full scan overnight, and when I woke up it had found some bad sectors/blocks/LBAs. It also told me that "Magician has found an error on the drive, it is recommended to recover" with an option to do "Recovery", which however failed with the message "Recovery failed, Check the disk connection and perform a diagnostic scan again".

Since I had to turn off the computer, I also clicked on a Windows notification saying something like "Problems found with the disk, reboot to fix the problems", which seemed to kick off a slightly longer reboot that printed some messages about "fixing" these issues.
After the reboot, I launched Magician and got the same result after another full scan;
I decided to the run a "full SMART self-test", which ended with another error message logged by Magician, "Defects have been detected from the device, Please check Help.", and after looking at the "Help" section of Magician I found a bit of text that just said "Some errors found during SMART self-test, we recommend replacing it with a new drive to protect your data".

Since no important data lives on this disk I proceeded to order a new replacement SSD, move out some of the stuff I wanted to keep, and then completely formatted the drive just for the sake of it.

After formatting I decided to relaunch the same tests, just to check if anything more could be done to maybe section of the bad sectors and keep the drive alive at reduced capacity, but to my complete surprise the "full scan" found no bad sectors whatsoever, and the "full SMART self-test" finished without any errors.

I am now wondering if the full formatting of the drive automagically did something that either fixed the problem the drive was having and/or removed those bad sectors I was seeing earlier, and I guess my question is: is this drive safe to keep using? all metrics and tests seem to indicate so, but given how much trouble it was having not even ~12 hours ago I am not sure if I can trust it.
As I said, no important data lives on here, basically only games and I don't really care about having to download them again, but I don't want to throw away a usable SSD if it can be salvaged.

I have attached a few screenshots taken throughout the process if they can somehow help pin this down.

[Timme]

12 minutes ago, Devilmoon said:

Yesterday I embarked upon a journey because a game launcher started giving me disk i/o errors when trying to apply a patch. This was on an SSD I regularly use for some videogames.
From there I opened the Windows event viewer which showed me multiple errors related to disk, namely "This device, <ssd id>, has a bad block".
I then tried running `chkdsk` on the disk, which did not find any issues; however, relaunching it with `/f` and `/r` it found multiple bad blocks and appeared to be moving around data (from paths I haven't used in ~years).
While this was going on I also launched CrystalDiskInfo, however SMART was reporting good condition with a score of 98%.

After chkdsk finished, I tried relaunching the game that started this all and still got the same errors.

This being a Samsung SSD I also tried using their own utility for disk health (Samsung Magician), which reports the drive being in good health. The short scan/ short smart self-test options in the diagnostic scan menu also found no issues.
I then tried launching a full scan overnight, and when I woke up it had found some bad sectors/blocks/LBAs. It also told me that "Magician has found an error on the drive, it is recommended to recover" with an option to do "Recovery", which however failed with the message "Recovery failed, Check the disk connection and perform a diagnostic scan again".

Since I had to turn off the computer, I also clicked on a Windows notification saying something like "Problems found with the disk, reboot to fix the problems", which seemed to kick off a slightly longer reboot that printed some messages about "fixing" these issues.
After the reboot, I launched Magician and got the same result after another full scan;
I decided to the run a "full SMART self-test", which ended with another error message logged by Magician, "Defects have been detected from the device, Please check Help.", and after looking at the "Help" section of Magician I found a bit of text that just said "Some errors found during SMART self-test, we recommend replacing it with a new drive to protect your data".

Since no important data lives on this disk I proceeded to order a new replacement SSD, move out some of the stuff I wanted to keep, and then completely formatted the drive just for the sake of it.

After formatting I decided to relaunch the same tests, just to check if anything more could be done to maybe section of the bad sectors and keep the drive alive at reduced capacity, but to my complete surprise the "full scan" found no bad sectors whatsoever, and the "full SMART self-test" finished without any errors.

I am now wondering if the full formatting of the drive automagically did something that either fixed the problem the drive was having and/or removed those bad sectors I was seeing earlier, and I guess my question is: is this drive safe to keep using? all metrics and tests seem to indicate so, but given how much trouble it was having not even ~12 hours ago I am not sure if I can trust it.
As I said, no important data lives on here, basically only games and I don't really care about having to download them again, but I don't want to throw away a usable SSD if it can be salvaged.

I have attached a few screenshots taken throughout the process if they can somehow help pin this down.

It is fine to use. Bad sectors are the worn-out NAND cells. Reformatting the drive removes the dead cells from the addressable storage map.

Some read, if you care.

[porina]

Can you run CrystalDiskInfo and screenshot the full output of that?

 

I have come across similar situations with low end SSDs before. This Samsung isn't low end, but the mechanism might be similar. If the flash is degrading then there might be more errors than can be corrected. The drive knows where these are, so when you "format" it, it can point the space to free good spots. Given the quantity of bad areas it found, the drive might just be failing in general. Even if it looks good, you might find more bad areas growing in not too distant future.

[Devilmoon]

19 minutes ago, Timme said:

It is fine to use. Bad sectors are the worn-out NAND cells. Reformatting the drive removes the dead cells from the addressable storage map.

Some read, if you care.

I believe I read that same article yesterday while trying to troubleshoot on my own. I kinda came to the same logical conclusion, however I don't know if my intuition can be fully trusted
 

 

19 minutes ago, porina said:

Can you run CrystalDiskInfo and screenshot the full output of that?

 

I have come across similar situations with low end SSDs before. This Samsung isn't low end, but the mechanism might be similar. If the flash is degrading then there might be more errors than can be corrected. The drive knows where these are, so when you "format" it, it can point the space to free good spots. Given the quantity of bad areas it found, the drive might just be failing in general. Even if it looks good, you might find more bad areas growing in not too distant future.

Sure, here you go:

[Tetras]

With 870 Evos that are consuming reserved blocks, they're normally dead or read only by the point it is noticed, but if they're not dead, once you update the firmware and reformat it will remap and act like they never existed. When was the drive manufactured?

[Timme]

2 minutes ago, Devilmoon said:

I believe I read that same article yesterday while trying to troubleshoot on my own. I kinda came to the same logical conclusion, however I don't know if my intuition can be fully trusted

Like, you doubt your sanity and comprehension, or the ability to match basic patterns?JK, you do right smart.

[Devilmoon]

12 minutes ago, Tetras said:

With 870 Evos that are consuming reserved blocks, they're normally dead or read only by the point it is noticed, but if they're not dead, once you update the firmware and reformat it will remap and act like they never existed. When was the drive manufactured?

I don't know, I can tell you that I bought it in April of 2021 so it must've been earlier than that

[porina]

20 minutes ago, Devilmoon said:

Sure, here you go:

It's been powered on for over 2.5 years, but the total writes don't seem excessive in that time. The bad blocks are reported 3 times, totally 636 if it is a direct conversion. This drive doesn't seem to report how many spares are remaining.

 

If I had this drive, I'd put it in "use with caution" category. Only put data on it you can replace, such as game installs. Especially if they're on a client that can repair e.g. Steam. Do frequent scans. If more new bad areas appear, ditch it.

[Devilmoon]

7 minutes ago, porina said:

It's been powered on for over 2.5 years, but the total writes don't seem excessive in that time. The bad blocks are reported 3 times, totally 636 if it is a direct conversion. This drive doesn't seem to report how many spares are remaining.

 

If I had this drive, I'd put it in "use with caution" category. Only put data on it you can replace, such as game installs. Especially if they're on a client that can repair e.g. Steam. Do frequent scans. If more new bad areas appear, ditch it.

not sure if 1 block == 1 LBA, because if so Samsung Magician at the very least tells me what is the maximum number of LBAs:


>If I had this drive, I'd put it in "use with caution" category.
Would you turn on RAPID and/or Over Provisioning?

[porina]

8 minutes ago, Devilmoon said:

not sure if 1 block == 1 LBA, because if so Samsung Magician at the very least tells me what is the maximum number of LBAs:

I think that just reports the drives max usable capacity, regardless of the condition.

 

8 minutes ago, Devilmoon said:

Would you turn on RAPID and/or Over Provisioning?

Rapid uses a ram cache that makes some benchmarks look better, but I'm not sure it makes much real world difference outside that. Let Windows handle it.

 

Over Provisioning I wouldn't bother with personally. As long as you don't fill up the SSD, and TRIM is working, you have a kind of overprovisioning there. Drive will know what blocks are unused and can use them to optimise for future writes. There are no exact numbers, but I would aim not to fill past 75%.

[Tetras]

36 minutes ago, Devilmoon said:

I don't know, I can tell you that I bought it in April of 2021 so it must've been earlier than that

I believe that fits the timeframe of the bad drives. When did you update the firmware? It is possible that most of the bad blocks emerged prior to the firmware update, in which case you're probably good to keep using it (while making sure no more are being reported).

[Devilmoon]

Just now, Tetras said:

I believe that fits the timeframe of the bad drives. When did you update the firmware? It is possible that most of the bad blocks emerged prior to the firmware update, in which case you're probably good to keep using it (while making sure no more are being reported).

>When did you update the firmware?
yesterday when I started investigating the issue

[Tetras]

24 minutes ago, Devilmoon said:

>When did you update the firmware?
yesterday when I started investigating the issue

I'd make a note of the number of bad blocks and reserved blocks used and then continue to use the drive. If the bad blocks continue to increase then your drive is bad. If not, then they all happened prior to the firmware update.