A story on the death of my SSD, and asking for advice on how to prevent future failures

[RanAwaySuccessfully]

Hi LTT community! I have a story to tell to y'all today.

 

The SSD in question is a WD GREEN 240GB 2.5".

 

So, yesterday it died. It wasn't a very good SSD mind you as it was the first one I ever bought, but I did not expect a sudden failure like that. I'm not going to ask for help in recovering any data, as I'm pretty sure it's dead and I have a backup from 3 months ago which isn't great but it'll do just fine. A backup misconfiguration on Timeshift meant that I didn't actually backup anything after April, I re-configured it again a few days ago but the next backup was scheduled to the beginning of August so I didn't bother doing it myself. I looked up online and I already have a good idea of what happened, apparently a specific part of the SSD is reserved for the firmware and other important system information, and this area may or may not be managed properly by the controller for wear and tear, meaning it can die prematurely after a lot of file writes.

 

Yesterday, I was in the middle of playing a game when suddenly the screen was replaced by a bunch of Linux system messages and EXT4 errors. I expected it to be /sda errors (hard drive) but instead it was complaining about /sdb (where both Linux Mint and my home partition, a.k.a user data are stored). "Oh no."

 

So next thing I try to do is power off the computer and get back in Linux as fast as humanly possible so I can backup any important data. But Linux did not want to boot up at all and kept giving me a load of errors during boot-up again. So, another restart, and this time it straight up didn't show up anything at all, as if it didn't even try to boot. This happened late at night so my brain wasn't exactly thinking as well as it should have, if I had plugged in a USB stick right after the first restart I might have been able to recover some stuff, but oh well.

 

The SSD is currently completely unrecognizable, even after trying to do a powercycle (I just learned about them yesterday, so I don't know exactly how they work but I tried to do it anyway). It died not only with less than 3 years since purchase, but also with less than 3 years since manufacturing (it was manufactured on "31 OCT 2018").

 

Before (Feb. 2019):

*-disk
     description: ATA Disk
     product: WDC WDS240G2G0A-
     vendor: Western Digital
     physical id: 0.0.0
     bus info: scsi@1:0.0.0
     logical name: /dev/sdb
     version: 0000
     serial: [REMOVED]
     size: 223GiB (240GB)
     capabilities: partitioned partitioned:dos
     configuration: ansiversion=5 logicalsectorsize=512 sectorsize=512 signature=3b430214

 

Now (Aug. 2021):

*-disk:1
    description: ATA Disk
    product: Sandisk Milpitas
    physical id: 1
    bus info: scsi@1:0.0.0
    logical name: /dev/sdb
    version: 0.00
    size: 16KiB (16KB)
    configuration: ansiversion=5 logicalsectorsize=512 sectorsize=512

 

I know now that Samsung EVO SSDs are the gold standard at the moment and I'll buy one of those as a replacement, probably a 870 EVO, but for now I wonder:

 

- Was this my own fault? Did I copy too many files to the SSD and wore it out prematurely?¹
- Is this something that can simply happen to any SSDs or is it specific to some types of SSDs?
- S.M.A.R.T. did not show up any errors just a few weeks prior. Is there any diagnostic I could've taken a look at so as to not get caught by surprise by this?
- Did I buy the wrong model, even while being in a budget? I had relied on Western Digital for HDDs, was it a mistake to rely on them for SSDs too?
- Do manufactures write somewhere in their manual or product page if the SSD does wear leveling of their system information?
- What can I do to prevent this in the future?

 

¹ I avoided copying in large files and games to it, but I have repartitioned the SSD a few times. I used to share it with Windows 10, then later on I split the root/OS partition and the home/user partition.

 

If anyone can answer these questions, it'd be of great help. Thank you for reading, and let this be reminder for y'all to always make frequent backups!

[da na]

It can happen to any SSD. WD Greens aren't the most high endurance drives out there, and it's not your fault from copying too many files to it. Technically the SSD wore out from lots of use and bytes written but it's definitely not your fault. Had a drive die on me a month or so ago to a controller failure, that's most likely how yours died too. Also, WD makes good SSDs, the Greens just aren't the best made SSDs out there.

[WintorialsLift]

3 minutes ago, RanAwaySuccessfully said:

Hi LTT community! I have a story to tell to y'all today.

 

The SSD in question is a WD GREEN 240GB 2.5".

 

So, yesterday it died. It wasn't a very good SSD mind you as it was the first one I ever bought, but I did not expect a sudden failure like that. I'm not going to ask for help in recovering any data, as I'm pretty sure it's dead and I have a backup from 3 months ago which isn't great but it'll do just fine. A backup misconfiguration on Timeshift meant that I didn't actually backup anything after April, I re-configured it again a few days ago but the next backup was scheduled to the beginning of August so I didn't bother doing it myself. I looked up online and I already have a good idea of what happened, apparently a specific part of the SSD is reserved for the firmware and other important system information, and this area may or may not be managed properly by the controller for wear and tear, meaning it can die prematurely after a lot of file writes.

 

Yesterday, I was in the middle of playing a game when suddenly the screen was replaced by a bunch of Linux system messages and EXT4 errors. I expected it to be /sda errors (hard drive) but instead it was complaining about /sdb (where both Linux Mint and my home partition, a.k.a user data are stored). "Oh no."

 

So next thing I try to do is power off the computer and get back in Linux as fast as humanly possible so I can backup any important data. But Linux did not want to boot up at all and kept giving me a load of errors during boot-up again. So, another restart, and this time it straight up didn't show up anything at all, as if it didn't even try to boot. This happened late at night so my brain wasn't exactly thinking as well as it should have, if I had plugged in a USB stick right after the first restart I might have been able to recover some stuff, but oh well.

 

The SSD is currently completely unrecognizable, even after trying to do a powercycle (I just learned about them yesterday, so I don't know exactly how they work but I tried to do it anyway). It died not only with less than 3 years since purchase, but also with less than 3 years since manufacturing (it was manufactured on "31 OCT 2018").

 

Before (Feb. 2019):

*-disk
     description: ATA Disk
     product: WDC WDS240G2G0A-
     vendor: Western Digital
     physical id: 0.0.0
     bus info: scsi@1:0.0.0
     logical name: /dev/sdb
     version: 0000
     serial: [REMOVED]
     size: 223GiB (240GB)
     capabilities: partitioned partitioned:dos
     configuration: ansiversion=5 logicalsectorsize=512 sectorsize=512 signature=3b430214

 

Now (Aug. 2021):

*-disk:1
    description: ATA Disk
    product: Sandisk Milpitas
    physical id: 1
    bus info: scsi@1:0.0.0
    logical name: /dev/sdb
    version: 0.00
    size: 16KiB (16KB)
    configuration: ansiversion=5 logicalsectorsize=512 sectorsize=512

 

I know now that Samsung EVO SSDs are the gold standard at the moment and I'll buy one of those as a replacement, probably a 870 EVO, but for now I wonder:

 

- Was this my own fault? Did I copy too many files to the SSD and wore it out prematurely?¹
- Is this something that can simply happen to any SSDs or is it specific to some types of SSDs?
- S.M.A.R.T. did not show up any errors just a few weeks prior. Is there any diagnostic I could've taken a look at so as to not get caught by surprise by this?
- Did I buy the wrong model, even while being in a budget? I had relied on Western Digital for HDDs, was it a mistake to rely on them for SSDs too?
- Do manufactures write somewhere in their manual or product page if the SSD does wear leveling of their system information?
- What can I do to prevent this in the future?

 

¹ I avoided copying in large files and games to it, but I have repartitioned the SSD a few times. I used to share it with Windows 10, then later on I split the root/OS partition and the home/user partition.

 

If anyone can answer these questions, it'd be of great help. Thank you for reading, and let this be reminder for y'all to always make frequent backups!

The exact same thing happened to me with a very similar SSD (only that instead of the 240GB 2.5'' version it's the 120GB M.2 version). Sitting on the login screen, shut off instantly, and the SSD was unrecognized.

Same thing for me too, SMART did not show any errors months, weeks, even days before it happened. No signs of bad things happening, no visible signs of disk check on startup, etc

For your questions - I really am not sure, I have the same questions.

I think it's just something that happens to SSDs at some point, but even then usually not this fast

Western Digital is one of the most reliable companies for storage media, whether HDDs or SSDs, but I particularly do not trust their WD Green products anymore after my personal experience and hearing yours.

It depends on the manufacturer. I got my SSD bundled with my laptop so I don't know

Maybe just try to reduce the load (i.e. don't write/read to/from it too much), but other than that I really have no idea either.

[WintorialsLift]

3 minutes ago, Mel0nMan said:

It can happen to any SSD. WD Greens aren't the most high endurance drives out there, and it's not your fault from copying too many files to it. Technically the SSD wore out from lots of use and bytes written but it's definitely not your fault. Had a drive die on me a month or so ago to a controller failure, that's most likely how yours died too. Also, WD makes good SSDs, the Greens just aren't the best made SSDs out there.

Yes, I asked this question a few days ago and someone also replied with "probably a controller failure" which is what I suspect happened to this person too

[Biohazard777]

9 minutes ago, RanAwaySuccessfully said:

- Was this my own fault? Did I copy too many files to the SSD and wore it out prematurely?¹

Do you have TRIM enabled?
For some reason when I installed latest Manjaro it was disabled by default... it should have been enabled. Dunno which distro you are running.
to check: systemctl status fstrim.timer

to enable: systemctl enable --now fstrim.timer

[WintorialsLift]

2 minutes ago, WintorialsLift said:

The exact same thing happened to me with a very similar SSD (only that instead of the 240GB 2.5'' version it's the 120GB M.2 version). Sitting on the login screen, shut off instantly, and the SSD was unrecognized.

Same thing for me too, SMART did not show any errors months, weeks, even days before it happened. No signs of bad things happening, no visible signs of disk check on startup, etc

For your questions - I really am not sure, I have the same questions.

I think it's just something that happens to SSDs at some point, but even then usually not this fast

Western Digital is one of the most reliable companies for storage media, whether HDDs or SSDs, but I particularly do not trust their WD Green products anymore after my personal experience and hearing yours.

It depends on the manufacturer. I got my SSD bundled with my laptop so I don't know

Maybe just try to reduce the load (i.e. don't write/read to/from it too much), but other than that I really have no idea either.

Oh yeah if you're wondering, my WD Green was manufactured date was somewhere around February 2020, only 15 months before failure

[RanAwaySuccessfully]

4 minutes ago, Biohazard777 said:

Do you have TRIM enabled?
For some reason when I installed latest Manjaro it was disabled by default... it should have been enabled. Dunno which distro you are running.
to check: systemctl status fstrim.timer

to enable: systemctl enable --now fstrim.timer

this is on my fresh install on my hard drive, so I assume it's enabled by default. i haven't actually restored from the Timeshift backups yet to check:

● fstrim.timer - Discard unused blocks once a week
     Loaded: loaded (/lib/systemd/system/fstrim.timer; enabled; vendor preset: >
     Active: active (waiting) since Sun 2021-08-01 06:43:43 -03; 1h 39min ago
    Trigger: Mon 2021-08-02 00:00:00 -03; 15h left
   Triggers: ● fstrim.service
       Docs: man:fstrim

 

4 minutes ago, WintorialsLift said:

Oh yeah if you're wondering, my WD Green was manufactured date was somewhere around February 2020, only 15 months before failure

jeez that's even worse. did their 3-year warranty cover that?

[WintorialsLift]

Just now, RanAwaySuccessfully said:

jeez that's even worse. did their 3-year warranty cover that?

Didn't bother contacting WD about it, since it just so happened that my laptop's LCD was also acting up and I sent it to HP, told them about both problems, and they installed Windows on the already existing hard drive + fixed the LCD.

 

Currently I'm eyeing a few SSDs, since I only recently knew that this laptop also supports NVMe SSDs lol.

 

Maybe I'm just unlucky I don't know

[Biohazard777]

5 minutes ago, RanAwaySuccessfully said:

this is on my fresh install on my hard drive, so I assume it's enabled by default. i haven't actually restored from the Timeshift backups yet to check:

● fstrim.timer - Discard unused blocks once a week
     Loaded: loaded (/lib/systemd/system/fstrim.timer; enabled; vendor preset: >
     Active: active (waiting) since Sun 2021-08-01 06:43:43 -03; 1h 39min ago
    Trigger: Mon 2021-08-02 00:00:00 -03; 15h left
   Triggers: ● fstrim.service
       Docs: man:fstrim

 

That looks fine.
Runaway logging or excessive use of hibernation maybe, but those would have to be really extreme to cause it to die within 3 years... (or worse in @WintorialsLiftcase).
Guess I am gonna stay away from WD Green line of products heh.

[WintorialsLift]

2 minutes ago, Biohazard777 said:

That looks fine.
Runaway logging or excessive use of hibernation maybe, but those would have to be really extreme to cause it to die within 3 years... (or worse in @WintorialsLiftcase).
Guess I am gonna stay away from WD Green line of products heh.

For information - I never really write big files per day, nothing that I think of that could've crapped the SSD, etc

I rarely let it hibernate/sleep

TRIM was on

I guess I was just... EXTREMELY unlucky.

[RanAwaySuccessfully]

I'm guessing WD Greens are just a bad product. I may try contacting WD to see if they'll either get a replacement or repair it, either way there's no way I'm using a WD Green as a main SSD anymore.

[Vishera]

Cases like that are one of the reasons that hard drives can have better longevity than SSDs.

 

You ask how to prevent future failures like that: You can't,It will eventually happen.

But you can delay it by writing as less as possible to the drive - for maximum longevity just barely write to the drive.

Do most of the writing on a hard drive,disable or move the page file to another drive and disable hibernation.

And buy higher quality + capacity drives - That's the most important factor here,because you can write more on higher quality + capacity drives

 

Make sure to backup your drives regularly.

[xg32]

2 hours ago, RanAwaySuccessfully said:

.

i tend to get 2 or 3 bit drives and never 4bit, i've had about 17-18 since 2011 due to different builds and the only one that failed on me so far was a crucial m4 made in 2011, the samsung one on my ps4 that keeps getting messed with by power outages is still going, doesn't mean they are bad drives, just RNG. Since then i've been using nothing but samsung ssds, 960, 970 evo evo plus pro, no pcie4 yet. Even if none has died the premium isn't worth it over some of the other options with the similar controllers. I believe these forums has a ssd tier list. Samsung's TV, monitors, and even fridges might be sus, but their ssds are good.

[Daethz]

All i know is

Don't get budget SSD's

keep firmware updated

 

The only SSD's I have ever seen fail have been cheap ssd's.

Plus 250gb isn't going to last long compared to a 1tb+, more space means it has more space to write stuff and prevent nand from being worn out.