SSD Reliability degraded but still has 94% of life

Topic

New Reply

My 3 yr old WDC 512 GB NVME SSD has generated a Windows 11 message that says “reliability degraded – back up in case of drive failure”.  However, it then says the estimated remaining life of my drive is 94% and available spare is 100%.  I am puzzled at the difference between the degraded reliability and the 94% remaining life.

Crystal Disk info also reports a problem, picture attached below.

I made a backup with Macrium and all my data is in the cloud anyway.  I suppose I will shop for a new PC or at least a new SSD.  Any other advice?  Risk it for a while and see what happens?

 

Reply | Quote

Replies

reliability degraded – back up in case of drive failure

With a current Macrium image at the ready I suggest opening an administrators CMD window (DOS) and running “chkdsk /f” (without the quotes).  You’ll be instructed to restart you computer before “chddsk /f” can run.

Desktop mobo Asus TUF X299 Mark 1, CPU: Intel Core i7-7820X Skylake-X 8-Core 3.6 GHz, RAM: 32GB, GPU: Nvidia GTX 1050 Ti 4GB. Display: Four 27" 1080p screens 2 over 2 quad.

Reply | Quote

Does CrystalDiskInfo have a place where it shows why it says the drive is bad? The view there shows only a handful of stats, and the color codes where it shows which stats are bad are not there.

The 94% refers to the rated life of the NAND cells on the drive in terms of write cycles. Each cell is rated for a certain number of write-erase cycles, and the 94% means you still have 94% of that rated life remaining. When the drive reaches 0%, that does not mean it is dead… the drive will probably continue to function normally for some time to come even after the NAND cells have reached their rated limit. If the NAND begins to wear out, but everything else is still good, you should see the available spare threshold count begin to decrease. If it continues steadily at that point, the drive is near the end.

There are other failure modes an SSD can have other than the media wearing out. There could be a problem with one of the NAND chips other than media wearout, or the controller could be failing. It could also be a cracked solder joint somewhere on the unit. These failure modes can cause the drive to fail catastrophically, but they will not show up in the drive health percentage figure. The 94% figure does not contradict the idea that the drive is bad, but without knowing why the utility and Windows think the drive is bad, it is hard to say.

One thing looks immediately wrong, though. The power on count shows over 462,000 power cycles. This is insanely high for a drive with 5172 power-on hours… I can’t imagine it is correct. This would be the number of times that the drive has been powered down and back up again, like when the PC is turned off or enters S3 sleep. I don’t believe it will register as a power cycle when the unit enters S0ix/S2idle/modern standby (all names for the same thing). If you are not putting the PC to sleep or shutting it down 100 times for every hour of use, ever since the drive was brand new, this number is not realistic.

I wonder if the drive is being flagged as bad because the cycle count is so high.

Why it is showing as that high is a mystery. I would guess a firmware bug, where the number is being artificially inflated for some reason, similar to how the Samsung 990 Pro SSDs recently were showing incorrectly high write cycle info, leading some to be reported as near the wearout limit when they were really only a fraction of the way there. The fix was to install the latest drive firmware. That corrected the erroneously high rate of write cycles being reported, but it would not correct the bad number of cycles the drive already thought had taken place.

The “bad” status from Windows and CrystalDiskInfo may or may not mean the drive is about to fail (if it is just a bug in the reporting of the stat, the drive could well be fine), but by backing up up at the first sign of danger, you did the right thing for sure.

Western Digital has tools on their site that can be used to diagnose the drive and either give it a clean bill of health or to verify it is bad. I would give that a try and see what it says. Perhaps the actual power cycle count reported by SMART is not 462,000, but somehow Windows and CrystalDiskMark are reading it as such.

Dell XPS 13/9310, i5-1135G7/16GB, KDE Neon 6.2
XPG Xenia 15, i7-9750H/32GB & GTX1660ti, Kubuntu 24.04
Acer Swift Go 14, i5-1335U/16GB, Kubuntu 24.04 (and Win 11)

4 users thanked author for this post.

Microfix, gtd12345, satrow, BobbyB

Reply | Quote

I agree with Arcaris. The power cycle is wrong.

My 5 years old Samsung Nvme .PC is 24/7/365 on.

1 user thanked author for this post.

gtd12345

Reply | Quote

The disk is rated bad because it has the Critical Warning value above 0.
This is set by the disk firmware.

I would assume the disk is about to fail and should be replaced – an easy thing to do if you have made a boot USB with Macrium.

As suggested by Ascaris, run the manufacturer’s disk diags to be sure.

cheers, Paul

1 user thanked author for this post.

gtd12345

Reply | Quote

Thanks for all the help.

1. I ran chkdsk /r, which completed successfully, but that did not clear the msg.
2. CrystalDiskInfo: The rest of the SMART data it reported was all green, only the first entry about critical warnings was red.
3. I downloaded the WD dashboard, which said a) the NVM system was degraded; b) agreed with CDI that there were 463K power cycles (!); and c) reported a phenomenal 91587755866489574258136806589572 media and data integrity errors.
4. I have asked for an estimate to replace the SDD from a local computer repair shop (I do not want to open up the laptop myself).  Once I get the estimate I’ll have to decide whether it is worth it to fix a 3-yr-old computer or to buy new.

Reply | Quote

gtd12345 wrote:

Once I get the estimate I’ll have to decide whether it is worth it to fix a 3-yr-old computer or to buy new.

SSDs are cheap and you can order a compatible SSD for your laptop from crucial.com.
You can also replace the SSD your self. Look for tutorials on iFixit.com.
There is no need to replace a 3-years-old laptop unless you need more powerful one.

Reply | Quote