SSD life dropping fast

G

genadiyrool

n00b
Joined
May 10, 2020
Messages
24
Hello, everyone, im new here
i have a certain concern about my SSD drive

my laptop is HP Omen 15 came with SSD Toshiba KXG60ZNV512G preinstalled

laptop is currently around 7 months old, and after installing CrystalDiskInfo a few months ago, ive noticed that my SSD health is dropping fast for no apparent reason (when i installed it, it was already at 96%)

i thought it might be due to many write cycles, but since that time it dropped to 95% i wasnt installing-uninstalling pretty much of anything and it still dropped to 94% again recently

performance in benchmarks are also suffered somewhat (from 212% when i just purchased it to 195% now on average on userbenchmark)

what could be wrong with it, help?
Untitled-2.png
 
13TB/14TB read write... Do you have an issue where your PC is constantly writing, seems like a lot of data for only a few months old box that you don't constantly install/remove apps on.
Not familiar with userbench, but is that a static reference or does their average update as more users benchmark? I would verify with like crystaldiskmark or similar.
 
i was reinstalling windows quite a few times in the first months, i can imagine i rolled up a counter somewhat by doing that
but since i noticed health falling from 96% to 95%, i think i pretty much stopped installing/reinstalling anything just to check how it goes
seems that in the last month from april 1st to may 1st total system writes went up by 1TB still. i guess its a progress comparing to 13TB in 6 months, but still... is it suspicious?
and if those writes are legit, is it normal to lose such percentage of health with that many writes?

userbenchmark doesnt change its reference, as far as im aware. i think 100% is just some popular SSD and the rest are measured up against it (like RTX2070 similarly has 100% on that website, and the rest of the GPUs are measured in relation to it). it also gives the short verdict, like "performing above expectations" or "performing below expectations", from which and to which i went gradually

ps.it is a system drive. i also have a pagefile for ~2500mb set up on it
Untitled-1.png

and thats the results from crystaldiskmark, but sadly i dont have any tests from when ssd was ok, so i have nothing to compare it to
 
I just looked at my drive and it has 20TB written and still at 100%

Samsung-1TB-SSD-Smart.jpg
 
hmm, yeah, thats what i read on the internet, that ssd shouldnt go that fast, even with so many writes..
Zepher you do have 12000hours though compared to mine 1800. so i think your writes are justified... not sure about mine
why is your power on count so low though? isnt that means hom many times was ssd switched off/on?
 
genadiyrool said:
hmm, yeah, thats what i read on the internet, that ssd shouldnt go that fast, even with so many writes..
Zepher you do have 12000hours though compared to mine 1800. so i think your writes are justified... not sure about mine
why is your power on count so low though? isnt that means hom many times was ssd switched off/on?
Click to expand...

My machines are on 24/7 and only powered down during upgrades and maintenance.
 
I do have this old SSD in my Plex server, this might have been the first SSD I bought, if it is, then it's been installed in a bunch of different PC's and laptops since 2012.
Wish it showed the Read/Writes on this drive.
I do need to replace it since I don't want it to fail while it's running my Plex Server.

Plex-SSD-Life.jpg
 
ok, yeah, i figured, it seems that you have it on for 207 hours on average
so thats 1 parameter is ok at least, i was worried that is messed up too

have no idea whats wrong with it
trim is enabled, never done defragmentation on it

saw someone write that in a similar case they reinstalled firmware and it returned the health back to 100%
but i cant find any firmware for that ssd, except the one i downloaded from dell website, not sure it will fit, only dell model laptops are listed as compatable
 
and in second one it seems that you were losing 1% every 883 hours
while mine loosing 1% in every 306 hours
but with no writes it could mean anything i guess
 
At 1% per month, the drive still has 8 years to go.

As to why it’s dropping, Windows is constantly writing to the drive for stuff like the hibernation store, pagefile, updates, and telemetry data, Assuming you have good backups, I wouldn’t worry about it.
 
hmm, so it will work just fine until last percentage?? (or close to?)
if thats the case, i guess i shouldnt worry, i never expected it work beyond that
but performance still drops... or at least dropped

temped to try and install the firmware update, dell released, should i?
i found this topic where a guy seemed to have it installed on non dell machine and it went fine:
https://forums.servethehome.com/ind...cting-firmware-from-a-dell-update-file.26528/
 
if its a dell branded SSD then that's where you have to go to get the firmware updates

i have the Toshiba XG3 NVME dell branded and i sometimes check for firmware (you should apply them if there is new ones available) the firmware is for the SSD not the PC so does not matter what the SSD is in

my SSD has had 29TB Written and its at 93%

even at 0% it does not mean its suddenly going to fail most ssds can do easy 400-1000TB Written before errors happen on TLC NAND (samsung MLC stuff can go for +4PB of written data)
 
oh, thanks for the info
so, it is going out twice as fast still, but at least it looks realistic now..

but how do i check if my SSD is dell branded?
or are all toshiba XG ssds are dell branded?
it should say "Dell" somewhere in the info? i couldnt find this word anywhere

edit: and its definitely writing something alot. counter is already at 14381GB. so 171GB in less than 1 day? i mean i did the CrystalDiskMark test, could that be it? i feel something is wrong
 
Last edited:
hmm, installation failed
i guess its not dell after all

guess ill just have to live with it
thanks for trying to help, everybody
 
I know I have an OCZ SSD which was also dropping health fast, after downloading the latest firmware and tools from OCZ the drive health returned to what it should have been and is behaving in line with the rest of my SSD's now. It's currently at 98% while it was down to 84% b4 the updates which was basically impossible for a drive used only to store and play games from.
 
hmm, so firmware install can help
i cant find one for XG6 though, anywhere
except on that dell website, which doesnt install
i think it might just not exist as an install package
 
genadiyrool said:
oh, thanks for the info
so, it is going out twice as fast still, but at least it looks realistic now..

but how do i check if my SSD is dell branded?
or are all toshiba XG ssds are dell branded?
it should say "Dell" somewhere in the info? i couldnt find this word anywhere

edit: and its definitely writing something alot. counter is already at 14381GB. so 171GB in less than 1 day? i mean i did the CrystalDiskMark test, could that be it? i feel something is wrong
Click to expand...
In your original post you said, "my laptop is HP Omen 15 came with SSD Toshiba KXG60ZNV512G preinstalled". If this is true you should check the HP site for firmware not Dell. I you don't find are firmware updates at HP then check Toshiba for updates. Hope this helps.
 
Chuklr said:
In your original post you said, "my laptop is HP Omen 15 came with SSD Toshiba KXG60ZNV512G preinstalled". If this is true you should check the HP site for firmware not Dell. I you don't find are firmware updates at HP then check Toshiba for updates. Hope this helps.
Click to expand...
yeah man, i checked all of it
hp site doesnt have anything, only thing i found is update for XG5 (https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX_008fc83badf748258c12df6386). other official site doesnt have anything related at all
their forums are also silent, no response
found an old topic on hp forums also, where someone advised the guy with hp laptop and similar ssd to download from dell (https://h30434.www3.hp.com/t5/Noteb...iba-xg4-256gb-nvme-thnsn5256gpuk/td-p/6085218) but he is a new member, so i guess he just googled it, and posted the link

toshiba website itself has only OCZ SSD firmware, as far as i understood, my SSD is different, since i cant find my model (or any XG models) when trying to download from their website
i wrote to their support, hope it will have some response

edit: also SSD keep on writing something, even though i havent installed or downloaded anything, except those dell firmwares, which are 12mb
counter is already up by 3GB since 8:22am message here, dont know, if its normal or not
 
Like I said, Windows is very chatty with the disks. Plus any other programs you've got going have their own various logs and caches. Though there is a bug in the Windows store that cause the drive to fill up with temp files, so you might check that. Otherwise, to get an idea of what's writing to your disk, open the resource monitor, click on the disk tab and watch the disk activity pane to see what files are being written to on your SSD (you can sort this be written bytes per second).

I imagine that the only thing that firmware changes do to improve remaining life is to switch to a nonlinear formula for bytes written, but I could be wrong.
 
  • Like
Reactions: raz-0
like this
yeah, i guess ~10GB a day is not so much
i was checking task manager for some time now also, it seems that its usually 1 to 4 tasks writing at 0,1mb/s speed while idle (it only goes up to 1mb/s or past in total when im watching videos on youtube or something)
i checked my Temp folders, and surprisingly they are also completely empty, 0 bytes

i would still want to try and reinstall the firmware, as a last try, i mean, if it could change something for the better, life, or performance, and if not, i guess i will just have accept it as inevitable
hope toshiba will reply something

edit: although, in last month it was 1TB in 1 month, it mean it averaging at 33GB per day, considering i havent been installing or deleting anything, its kinda too much i feel
 
Last edited:
Run Resource Monitor for a while and see what's writing to the SSD the most.

I had a similar issue (albeit not as extreme as yours) where Plex was constantly writing to its database due to it constantly updating metadata for a library on a different drive.

Unless you start monitoring for writes either using Resource Monitor, Performance Monitor, or something similar, guessing won't really be able to help narrow it down.
 
oh, ok, resource monitor, i see it now
i was thinking its just a tab in a task manager, now i see its a separate thing
first time i opened this thing
will keep it running for 1 day to see whats happening, thanks guys
 
The size of the drive and the NAND, controller and firmware play a part in how quickly you can wear a drive.


I think the Power on count is a large part the problem meaning you are likely hibernating to many times.
 
Some drives will vary the health information based on the TBW regardless of actual health. I believe this is an OEM drive (XG6) which has no official TBW, however their retail drive (RC500) is rated for 200TB. (200)(1 - .94) = 12TB of writes, or the range would be 10-14TB (5%-7% wear) in which you happen to fall (you are at 13.874TB). This means nothing about actual endurance - in fact the flash on the XG6 would likely survive in the PB range.
 
seems that i found what making so many writes
chrome has almost 20 different processes open in resource monitor (all in appdata/local/google/chrome/user data/default/cache), all writing, ranging from 30kb/s to 2mb/s constantly, adding up to around 5mb/s. while watching im some livestream or something
task manager oddly shows that chrome only uses 0,1-0,5mb/s at the exact same time

checked laptop in idle state, nothing unusual, check during all my 5 video games, nothing unusual as well, pagefile used very rarely, only for several seconds, without much writes
 
drescherjm said:
The size of the drive and the NAND, controller and firmware play a part in how quickly you can wear a drive.





I think the Power on count is a large part the problem meaning you are likely hibernating to many times.
Click to expand...

power on count seems to be around 2,5 per day, i dont think its too much frankly
but i want to try and reinstall the firmware, yeah

Maxx said:
Some drives will vary the health information based on the TBW regardless of actual health. I believe this is an OEM drive (XG6) which has no official TBW, however their retail drive (RC500) is rated for 200TB. (200)(1 - .94) = 12TB of writes, or the range would be 10-14TB (5%-7% wear) in which you happen to fall (you are at 13.874TB). This means nothing about actual endurance - in fact the flash on the XG6 would likely survive in the PB range.
Click to expand...

well, so it seems its legit %? just a low projected lifespan? thats a bummer.. but the part that it means nothing, thats inspiring) performance definitely dropped though

Denpepe said:
Did you actually check the HP support page for SSD firmware for your exact laptop model?
Click to expand...

of course man, first place i went to
 
The Health Status percentage does seem to be somewhat arbitrary. I have a Samsung 850 Pro with 33k+ power-on hours and 25TB+ written, still showing 100% good. Then I have a Samsung 960 Pro with only 12TB written, less than 10K power-on hours, but now only showing 98%. Both are MLC drives, and I'd like to think that I've barely scratched the surface of what they are each capable of. I think that this stat is more for the manufacturer's benefit than the consumer's benefit. The drive can still work fine even at "0%", but at that point the companies would have something to point to (to assuage themselves of blame) just in-case it doesn't.

850Pro.png
960Pro.png
 
that makes me feel better
but why performance dropped, it definitely dropped, im not delusional

edit: and +20GB of writes at the end of the day again
havent spotted anything except chrome, but even with it, it shouldnt be even nearly so much, i have no idea

edit edit: wait, its mb per second, dammit its exactly how much it should be, its getting late here i counted wrong
 
Last edited:
Zepher said:
I do have this old SSD in my Plex server, this might have been the first SSD I bought, if it is, then it's been installed in a bunch of different PC's and laptops since 2012.
Wish it showed the Read/Writes on this drive.
I do need to replace it since I don't want it to fail while it's running my Plex Server.
Click to expand...

You can tell how worn it is by the erase block count raw value, 48A = 1162 P/E. This determines the reported health value (which is why its value is "62") with a target of 3000 P/E. This implies you're already well beyond the TBW (I believe it was 72TB on that model - you're already at least twice that in writes) but...it's MLC with DRAM. It's rated for 3000 P/E and it's likely to survive at least 10K. Chances are the flash won't wear out, but something else might.
 
GotNoRice said:
The Health Status percentage does seem to be somewhat arbitrary. I have a Samsung 850 Pro with 33k+ power-on hours and 25TB+ written, still showing 100% good. Then I have a Samsung 960 Pro with only 12TB written, less than 10K power-on hours, but now only showing 98%. Both are MLC drives, and I'd like to think that I've barely scratched the surface of what they are each capable of. I think that this stat is more for the manufacturer's benefit than the consumer's benefit. The drive can still work fine even at "0%", but at that point the companies would have something to point to (to assuage themselves of blame) just in-case it doesn't.
Click to expand...

The 512GB 960 Pro has a TBW of 400TB so 98% might simply mean between 8-12TB of writes. CDI is reading the health of your 850 Pro incorrectly - your wear-leveling count is 165 = 357 P/E and the 850 Pro is rated at 6000 P/E which comes out to ~6% worn, which is why the value of that count is shown as "94" (94% health).
 
Zepher said:
I just looked at my drive and it has 20TB written and still at 100%
Click to expand...

CDI is reporting it wrong. The 1TB 860 EVO is rated for 600TB TBW and you have 19.89TB of writes, which should be reported at 96% health. Which it actually is under the wear leveling count.
 
Last edited:
Maxx said:
CDI is reading the health of your 850 Pro incorrectly - your wear-leveling count is 165 = 357 P/E and the 850 Pro is rated at 6000 P/E which comes out to ~6% worn, which is why the value of that count is shown as "94" (94% health).
Click to expand...

Cool, thanks for pointing that out. I don't recall having issues with CDI before but I guess I'll have to start double-checking it's numbers on drives where it reports 100%.

It does bother me on a certain level once a drive starts to slowly tick downward in percentage, even though I know that it does not matter. I guess that's probably why other software like Samsung Magician simply reports that the drive is "Good" instead of giving more detail.
 
The reported health % is very rarely accurate. Usually you can tell when a drive is dying when the spare block count goes down - typically it starts at 64 in hex, which is 100 in decimal, although that doesn't always mean 100 spare blocks. Once the drive starts using spare blocks it's at the tail-end of the bathtub curve such that you should at that point back it up and replace it. The average block erase count is therefore a pretty good metric, moreso than writes as they don't always take write amplification into account, with modern 3D TLC being in to 3K-10K P/E range. (usually rated for less, e.g. 2500 for BiCS3 with an expected WAF of 1.5 as with the TBW of E12-based drives, however I'd expect it to survive at least twice that)
 
drescherjm said:
Your performance looks fine to me unless I am reading this wrong your crystaldiskmark scores seem to be a little better than the submitted benchmarks at this site:

https://www.notebookcheck.net/Toshiba-XG6-KXG60ZNV512G-SSD-Benchmarks.458858.0.html
Click to expand...

hmm, interesting. well, i think something is wrong somewhere

the proclaimed read/write speeds for this one are up to 3,100 MB/s and 2,800 MB/s, so its my read what is definitely lower right now
https://business.kioxia.com/en-us/ssd/client-ssd/xg6.html

also userbenchmark definitely showed 212% with "performing above expectations" (which basically means above average collected from all users, which is 200%) before
drop to 195% "performing below expectations"

i mean im not really bothered as long as its not continuing to drop further

but at the same time it would be just nice to restore it to previous levels, why not

so i guess that health % should be no concern, judging from all the info, it probably was dropping since the beginning, when ssd was outputting 212%, i just didnt see it
 
Last edited:
When you're surrounded by overflow of data monitoring, you're just going to find something to go worrywart about.

If the percentages are taking a month to reduce by one, that would be fairly normal (one-hundred months total life, or 8 years!)

I wouldn't worry about it unless it's decreasing daily. And it sound like you've only run this thing a few times recently (a couple test runs does not yield DATA with monitoring programs!)

Give it a couple weeks , then come back and complain about daily falls in percentage. TLC Flash has a finite write lifetime
 
Last edited:
defaultluser said:
When you're surrounded by overflow of data monitoring, you're just going to find something to go worrywart about.

If the percentages are taking a month to reduce by one, that would be fairly normal (one-hundred months total life, or 8 years!)

I wouldn't worry about it unless it's decreasing daily. And it sound like you've only run this thing a few times recently (a couple test runs does not yield DATA with monitoring programs!)

Give it a couple weeks , then come back and complain about daily falls in percentage. TLC Flash has a finite write lifetime
Click to expand...
yeah, thats true
im not really worrying about health percentage, thanks to this topic

its just my initial thoughts were that performance dropped suddenly and health dropped suddenly (since i wasnt monitoring health until i saw performance drop)
so i thought something happened, and continues to happen and was perfect before

userbenchmark i run many many times, it was consistent 210-212%, and then it dropped to 195-198%, with average from other users being the same (BUT maybe it actually gradually dropped, maybe i just wasnt paying attention until it went past below 200%, now when i think about it)
so while im settled down about health % dropping, im still not really sure about that performance drop
could it be natural consequence of health % dropping? or something did change, which dropped performance once?
 
You must log in or register to reply here.
Back
Top