NVMe failures

[cmdr milesred]

that report looks fine, only thing I see bad is the temperature and that's pretty hot. Hot enough to cause issues occasionally
your windows disk is able to repair the install for windows.
for Linux, there are many tools available to restore the disk
if the MBR is healthy and it is not booting and the bios is set correctly to boot to the Nvme, then something is preventing the boot
remove any usb or cd/dvd
make sure all boot options are null except the one you want.
If you need a better program to do disk stuff, Paragon makes great stuff and they keep up with the tech
you can get free programs like that and much more on Hirens disks. or image for usb
that aside
If I were you I would check Tom's hardware for some help to this.
and I would do my best to keep the temp down, normal s.b. around 70C
I have 2 such devices running in RAID config. in a micro pc
It took 4 returns to vendor before they sent me 2 that worked properly - this pc has no fans, it has to be placed properly for the venting to work well is all.
the problem for me was they sent at least one piece of garbage each time
I asked them to open them up and make them work before sending to me, stop wasting my time.
They are ram and are far more susceptible to damage than a regular moving parts hard drive
like a memory card or a usb stick, the ram can get damaged very easily.

I had the same issue when ssd's were new to us all, every one I bought for the first 2 years was pure garbage with life expectancy of 5 minutes IMO.
It was like ram chips not dying was new to all manufacturers...
Many died from very poor housings and heat killed them soooo fast.


but for you, you need to make sure the innards of your pc are not hostile to additions.
Temperature is important
my pc has 1 fan on each hard drive
2 on the back
2 on the top
4 in the front
I change them when they start to rattle (2-3 years each)
I clean them every month with a C02 pistol
My laptops(I have a few) have extra fans around them
my main laptop MSI GT70 has great fans, I keep it on a flat hard stable surface
and when I run ED on it, it has 2 fans I added to it , 1 on each side blowing 1 way to cool it more than its internal ones can.
For ED, this is a must.

Once a year I take my laptop and my pc apart and make sure all connections are clean(I remove any oxidation) and all heatsinks are clean.


I hope you get it sorted.
I like your mb btw, nothing wrong there.

 

[Fnord Velkor]

that report looks fine, only thing I see bad is the temperature and that's pretty hot. Hot enough to cause issues occasionally
....

Thing is those temps are the stick's set levels for Warning and Critical conditions (I assume) since there is no reporting of the actual SMART data. I admit to not actually knowing what the rest means though. (Technology has moved on since my day.) LOL

 

[Morbad]

As noted, the temperatures were thresholds set for the controller, not current temperature. Anyway, though that is not a complete SMART report, no warnings means the drive doesn't have any failure or pre-failure conditions of note that SMART would be able to detect.

While SMART can't catch nearly everything, chances are it's not something physically/electrically wrong with the drive.

Are you using secure boot or something that could cause issues?

 

[Fnord Velkor]

As noted, the temperatures were thresholds set for the controller, not current temperature. Anyway, though that is not a complete SMART report, no warnings means the drive doesn't have any failure or pre-failure conditions of note that SMART would be able to detect.

While SMART can't catch nearly everything, chances are it's not something physically/electrically wrong with the drive.

Are you using secure boot or something that could cause issues?

Your remark made me go into BIOS and mess with secure boot - made no difference. At the moment (running UBUNTU) and by default it sets to "other OS" - changing boot order and selecting secure boot to "Windows UEFI" - makes no difference because I don't have secure boot key etc. So in either case I still get blue screen on failure to load W10 with the "critical process died" message when the boot order has the NVMe stick at the top.

So basically Windows is just not seeing the NVMe stick, it thinks a different SSD is drive C and sice that has no OS on it, windows fails and can't repair. I have tried the system restore option in the startup repair sequence and that just has drive C greyed-out with the "you need to enable system restore on this drive ... ". I even tried the repair windows (save your files) option and that won't proceed since "drive C" - has no OS.

I am not up to speed on PC stuff any more so I assume that the BIOS only showing the NVME in the boot order and not listing it under the "PCH Storage Configuration" is normal (since it is not in SATA mode).

I have a new stick arriving tomorrow so this weekend will be strip the PC and play with the physical aspect.

Thanks again to you and everyone that has chipped-in, it is appreciated.

 

[cmdr milesred]

interesting thing about ram and s.m.a.r.t.
one hit of too hot is all it takes to make both the s.m.a.r.t. not effective and some ram no longer existing.
there is no s.m.a.r.t. that is capable of doing this.
only by monitoring from another 'control' computer can you see damage being done.
hit it with static and test it yourself, it cannot tell if it is damaged. its only clue is it no longer works the way it was supposed to.

like RAID, these things in penny computers are not the same as in a pc built for actual tolerance.

also, I want to say, you had it working once, so it does/did work.
this is not new tech

even really old laptops from dell and others have been using m2 for years with no issues.
sharing sata ports is the cheap MB and requires knowing what ports 'SHARE' and disable.
make sure the one you buy IS compatable with your MB.

Use the warranty, don't waste time on faulty product.
I would be interested in knowing where(what store you are buying these from)
Not all stores buy from good suppliers.

Talk to a tech at a store, get someone that knows what they are doing to do this for you if you are not a computer tech.
stuff like this is easy for some of us, and others see blue. some reach for a pipewrench.....
like I mentioned once before, these are very volatile, super super easy to damage. taking out of package is often enough to wipe them out.
an RMA is a pain, but is the fastest way to deal with crap like this.

The story I told was for a reason, 512gb m2, $540 each and I needed 2 to work in RAID stripe.
you think it is hard to pick 2 off the shelf and put them in the pc?, ofc not, super easy.
yet every time, they sent at least 1 dead 1.
months of this till I got mad, sent the last 2 back and demanded my money back.
They do not care about you or your money, you are a drop in the ocean of customers.
you are a note, an email, a pest. unless you bring it to the owner or the managers attention.

I still deal with this company and I still give them a lot, they do NOT care.

good luck

 

[Fnord Velkor]

Still awaiting delivery of new stick but ended up doing a fresh W10 install on a different drive and lo and behold the NVMe is there after all ...

... However it seems to have been split into a "System Reserved" 500MB section and the rest as a healthy primary partition (which I named "WD BLACK" just to make sure I was looking at the right gubbins). See pic below.

So maybe there is a way of resurrecting this drive by stitching the bits together?

Or maybe I have to delete the partitions and format the drive to get it back useable?

(BTW Disk 2 in the pic is a SSD with UBUNTU on it.)

 

[Andy Doggett]

Still awaiting delivery of new stick but ended up doing a fresh W10 install on a different drive and lo and behold the NVMe is there after all ...

... However it seems to have been split into a "System Reserved" 500MB section and the rest as a healthy primary partition (which I named "WD BLACK" just to make sure I was looking at the right gubbins). See pic below.

So maybe there is a way of resurrecting this drive by stitching the bits together?

Or maybe I have to delete the partitions and format the drive to get it back useable?

(BTW Disk 2 in the pic is a SSD with UBUNTU on it.)

View attachment 320572

System reserved is usually W10 boot manager and stuff, created when one installs W10 on a fresh drive. Re: your BSOD, the drive could be failing but there could also be a bad or damaged connection between it and the motherboard, either on the drive itself or the motherboard, most likely the motherboard since you've had several failed drives. At least that's what the interwebs tell me. Hopefully it's just the drive .

 

[Morbad]

Still awaiting delivery of new stick but ended up doing a fresh W10 install on a different drive and lo and behold the NVMe is there after all ...

... However it seems to have been split into a "System Reserved" 500MB section and the rest as a healthy primary partition (which I named "WD BLACK" just to make sure I was looking at the right gubbins). See pic below.

So maybe there is a way of resurrecting this drive by stitching the bits together?

Or maybe I have to delete the partitions and format the drive to get it back useable?

(BTW Disk 2 in the pic is a SSD with UBUNTU on it.)

View attachment 320572

If Linux saw it, Windows should see it. That is not surprising. There is probably nothing at all wrong with the main partition on the drive, likely never has been.

That System Reserved partition is the EFI System Partition from the old install, which looks like it's been repurposed by the new install as your new OS drive only has the primary and recovery partitions.

You might be able to delete and rebuild the EFI System Partition and recovery partition on your original drive.

Skip step 1:

Redirecting

If that is not possible, you should clean both drives in diskpart so WIndows stops getting confused. There should be no extraneous bit of old Windows installs when installing Windows fresh, or you get stuff in the wrong places. These drives should be completely blank.

System reserved is usually W10 boot manager and stuff, created when one installs W10 on a fresh drive. Re: your BSOD, the drive could be failing but there could also be a bad or damaged connection between it and the motherboard, either on the drive itself or the motherboard, most likely the motherboard since you've had several failed drives. At least that's what the interwebs tell me. Hopefully it's just the drive .

At this point, I'm doubtful there is even a hardware issue with the drives. I don't know what's corrupting the EFI System Partition/bootloader on the OPs drives, but that being the only failure on different drives makes a hardware problem extremely unlikely.

The original black screen crashes are an issue, and the system needs to be comprehensively tested for stability.

 

[cmdr milesred]

System reserved is usually W10 boot manager and stuff, created when one installs W10 on a fresh drive. Re: your BSOD, the drive could be failing but there could also be a bad or damaged connection between it and the motherboard, either on the drive itself or the motherboard, most likely the motherboard since you've had several failed drives. At least that's what the interwebs tell me. Hopefully it's just the drive .

on a mb that's a few years old, yes very possible.
I would clean the connection as best I could.
easiest for this type of connector is use the old drive, clean its contacts, put some contact cleaner on them(like with a Q-tip) then insert and remove a few times.
When inserting a new card into any used slot, I always insert and remove a few times for this very reason too. the tiny contacts in the slot and the card doing this action a few times is usually enough to make sure a good connection is made
doing it a few times, usually you can also feel a better insert.

thanks to just recently having to replace my video card, I ended up cleaning all the heatsinks and ram and all the other cards too, might as well while in there..
good for 2 more years.

 

[Fnord Velkor]

Yeah, what I don't understand is how the stick stopped being bootable, was not detected by windows installer (or repair or whatever), only by making a fresh W10 install on a separate SATA SSD (what is now disk 1 in the pic) allows me to see the stick. What is even more confusing to me is that the "System Reserved" fragment of the stick (now called Disk 3) gets a Drive letter (G) assigned by default, as well as another drive letter (H) for the rest of the stick - this seems to me to be the reason for windows losing the "bootability" of the stick and I wonder what would have caused the fracturing of the file structure - doing a windows defender offline scan was the immediately preceding activity when the blue screen boot error appeared.

If I could resurrect the stick, it would be useful to save me reinstalling stuff but I don't think I want to use it as a boot device again in any case. In fact I am not too keen on having my new stick (Samsung 980 PRO with heatsink) as the OS bootable drive as this reinstalling stuff gets a bit old - plus Ableton might not let me have yet another authorisation (they gave me a further computer authorisation after the last stick failure) and I don't know how much a further licence for their DAW would cost (and I don't want to learn a different prog).

While I was mucking about, I did dig out that previously failing stick from May (an evo970 1TB) and that just consistently produced the "hard-drive failure imminent" boot failure (as before) and no cycling etc made any difference so that seems to have been a genuine SMART condition detection.

Once again - thanks for the posts, very kind of you all to help out a septuagenarian.

 

[Old Duck]

Your remark made me go into BIOS and mess with secure boot - made no difference. At the moment (running UBUNTU) and by default it sets to "other OS" - changing boot order and selecting secure boot to "Windows UEFI" - makes no difference because I don't have secure boot key etc. So in either case I still get blue screen on failure to load W10 with the "critical process died" message when the boot order has the NVMe stick at the top.

So basically Windows is just not seeing the NVMe stick, it thinks a different SSD is drive C and sice that has no OS on it, windows fails and can't repair. I have tried the system restore option in the startup repair sequence and that just has drive C greyed-out with the "you need to enable system restore on this drive ... ". I even tried the repair windows (save your files) option and that won't proceed since "drive C" - has no OS.

I am not up to speed on PC stuff any more so I assume that the BIOS only showing the NVME in the boot order and not listing it under the "PCH Storage Configuration" is normal (since it is not in SATA mode).

I have a new stick arriving tomorrow so this weekend will be strip the PC and play with the physical aspect.

Thanks again to you and everyone that has chipped-in, it is appreciated.

Are you able to boot Windows via Safe Mode?

I ask because I had trouble running my swap file on SSD (NVMe) due to repeated corruption, which sometimes caused serious problems like crashing and then not being able to reboot. Once I disabled virtual memory, everything has been rock-solid since (except Star Citizen, LOL).

 

[cmdr milesred]

I would look into the hirens disks.
I would recommend norton ghost. the version 2003 dos is the best version. you can find a version that is for usb or cd/dvd .
(Ultimate Boot Disk Ghost (1394 Usb1.1 Support) Drive Image Partition Magic Ntfsdos)
if it can make a backup, then you can restore files as you need/want them or an entire disk restore.
if it cannot make a backup due to disk errors, it will tell you.
if that happens, then you can try partitionguru, it will make a backup as long as it can read the drive.
between these 2 you should be able to backup your data.
you can get free versions of these easily, especially ghost, as old as it is, it is still the best. and the one I refer too is customized to boot into just about everything.

as a tech, I always make a backup before any attempt at a repair. and these 2 programs have never once failed me.
I can tell you almost every version of ghost except this one, also sucks and fail is their middle names..

fwiw, I use ghost every time I build a computer and usually shortly after any new software install has proven ok.
it may take 45 minutes to make a backup and store it on preferably a external powered usb drive..whatever..
but then it usually takes 5 to 10 minutes to restore if it gets messed up.

worth thinking about when you have software that has your creative effort saved.

all the stuff I create gets put on usb flash drives and usb powered drives and dvd's regularly. it sucks to lose stuff that took time and effort to create.


thanks to the flakey ssd's I previously mentioned, I had to restore a ghost image every 2 weeks to that pc, reset the bios, restore, then restore the bios.
all because of seriously flakey ssd's.
thank you NORTON. 10 minutes is beautiful.

 

[STRONTIUM DOG]

Yeah, what I don't understand is how the stick stopped being bootable, was not detected by windows installer (or repair or whatever), only by making a fresh W10 install

try to imagine the ssd is playing a game of tetris @ the speed of light blocks being bits
if there is a power loss mid fall (during a write) the blocks are frozen mid fall any attempt to manually restack them at that point is futile (ubuntu)
the drive is now not neatly stacked so win cant read weather its full or empty resulting in drive blindness
a power cycle will restack the blocks neatly were it should have just rebooted...……….
if you made invasive changes "mid fall" I fear that just corrupted the process (resulting in reformat or new os requirement)
well that's my take on things
WOULD YOU LIKE TO KNOW MORE!

 

[Fnord Velkor]

Again - thanks for the posts.

I think I will just reformat the stick and start from scratch again with it. I won't put the OS on it though, I'll leave the OS on the SATA SSD that I have used just now. (I intend building a new PC once the next gen intel gets released - the new 980PRO can be the OS disk for that.)

All of my documents and created stuff are a) never stored on the OS drive and b) backed up onto NAS and USB sticks - videos and music also on DVDR. The only embuqqerance is reinstalling the "involved" progs (FSX takes a day) and re-authorising applications.

I do have an image saved (well, several really) but it wouldn't write / restore / whatever so I think that I shall investigate that boot ghost thingie.

Thanks again.

 

[cmdr milesred]

if you cannot find anything, let me know.
everyone should have ghost.
it and partition guru are able to do what they do, because they do it the right way.
most backup software is crap and uses variations of zip or rar to read and store, this is a joke. this is how you create problems.

 

[yelboc]

If my memory serves me right then on some Linux distros 'fixntfs' can repair MBR records.. You might need to install it - ntfs-3g comes to my ageing braincells..

 

[Bravo17]

What is even more confusing to me is that the "System Reserved" fragment of the stick (now called Disk 3) gets a Drive letter (G) assigned by default, as well as another drive letter (H) for the rest of the stick- this seems to me to be the reason for windows losing the "bootability" of the stick and I wonder what would have caused the fracturing of the file structure

This is totally normal - the new install sees both partitions on the SSD as normal partitions since it hasn't been told that the System Reserved one is "special" (it isn't for the new install) and shouldn't get a drive letter, so it gives it the next free one. And the old C partition gets the next one after that. The partition structure (I assume that is what you meant by file structure) is not "fractured", it looks normal.