Hello everyone, I didn't think I would be back so soon honestly..

After my small DR adventure I thought I had finally reached some stability and yet here I am.

Today my SSD (Samsung 970 EVO PLUS, 250 GB, M.2, NVMe), purchased 7 months ago a week after I had had the issue with data loss, suddenly disappeared from boot leaving the error message "PXE-E61: Media test failure, check cable".

I leave it to your imagination what my face was..

There were what they seem now to be symptoms: unexpected programs freezes and HDD behaviour. Few but curious for a SSD. Then a couple of days ago the PC went on suddenly by itself.. I know, it's becoming a horror story. It turned on by itself from sleep because it seems somehow mouse and/or keyboard had lost their power management settings?! I had disabled them from BIOS. I checked: they were still disabled.. They were not disabled on Windows though. From that moment on, i think no more than a couple of days, every time I turned it on the SSD was not booting automatically. It was still visible and I had to boot it manually. Weird, until today.

Now. I went in with Linux and it can see and access the SSD. Since my experience with data loss I adopted a tight policy on back up, in fact I immediately made an image after the first few suspicious freezes not to mention all the mass storage data are in a separate HDD, so no data in jeopardy this time. But what now?

After Linux it seemed the SSD got shaken a bit and went visible on boot, but entered recovery. After completing recovery it went missing again.. So annoying.

I purchased a Kingston 16 gb USB flash drive (one of the most sold, supposedly tested quality) where I currently keep Linux. One thing I noticed every time I restarted the PC and went back from Linux to Windows my USB wifi dongle went unresponsive and needed physical removal in order to work again. Why I am saying this is because perhaps the motherboard might be the issue here.

I ordered a SSD to SATA adapter to see if anything changes.

What do you think?

Personally I would:

1. Install another hard disk and take the suspect one out
2. install windows
3. Install drivers and re-flash BIOS if possible
4. Install your "burn in test" software of choice or mainboard vendor diagnostic tools and test system. Also READ YOUR WINDOWS EVENT LOGS. no-one ever does this, but they are a wealth of info.

next steps depend on results. If no issue, get a new Hard Disk. Old one likely warranty, but In my experience, not worth the hassle and postage cost / effort.

you could also put the SSD in another system and use vendor test on it. I rarely will give any HDD or SSD a second chance, I mean how can you really vet it? It already crapped out after a mere 7 months, and if you cant find the exact cause or reason, how can you trust it?

I find spending 2 hours on the above quickest and easiest... and I see a LOT of PCs a week.

Hello @HaQue, thank you for replying. So, basically in your 4 steps you are explaining how to create a new environment to better test the motherboard, right?



I agree. Can't it be even quicker? Can I not flash the BIOS anytime independently and simply use a Linux flash drive to test the motherboard? Any burn-in test for Linux?


What do you mean? What else otherwise. I bought it on Bezoshop, I haven't checked about warranty yet.

Gone after 7 months. What a drag. Any other brands presumably more reliable for the future?

Another thing: if I give it back for warranty I would like to destroy all data if possible. I could access it from Linux last time I tried, what is the best way to erase it compleately?

Lots of questions I know
Thanks in advance for all you help

I was saying how I would do it. I have more experience in this way and comfortable with it, so it is easier to notice what part is failing, if anything.
The general gist is to test. if you have a way to do it that suits then by all means use it.



Ive had hard disks sent back and paid 1/2 the cost of the drive for the privilege. I don't find it worth it, but as above you might want to.



I doubt the vendor would spend too much time trying to recover your data to snoop through it, but I do understand privacy worries.
you could do something like what is described here:
https://superuser.com/questions/1284450/quickest-way-to-wipe-an-ssd-clean-of-all-its-partitions-for-repartitioning-in-li

such as

HaQue is correct in that overwriting the data will erase the user data, but since you have a modern SSD, you can use the sanitize feature set which erases much more quickly and relocated sectors as well:



In fact, I've seen drives where this operation resets the firmware and makes the drive "work again", not that I would trust it though.

Thank you guys for clarifying.



The SSD have 5 years limited warranty. Aren't these sort of issues covered?

I'm fed up already. Do you advice any other SSD brand I can try my luck with this time?

I trust Samsung EVO's but honestly I don't really care that much what I use as long as it isn't Kingston. They are all fairly similar, the nuances about speed etc don't bother me as much as other people. Maybe it is because I don't have a need to get into the weeds as much.

Really, just read some reviews on some current available drives in your area/online and choose one

I agree. The difference in speed between brands and models is not crucial for everyday use. I chose Samsung because I though I could rest easy for a few years..



Sorry if I'm stressing this again but is it normal practice to charge even for these kind of issues? I had a look at the policy:



I haven't read anywhere stating any fee for the service, at least for these type of issues, namely failure cause by circumstances allegedly beyond the user's control - I mean the SSD is physically perfect, no dirt, scratches or missing labels. Am I missing some important part known by only who belong to this line of job?
Of course if SAMSUNG, in its sole discretion, determines somehow the issue is imputable to me than that is another story. Will this be the case?

Thanks

Hi,
No you are not missing anything. I have not read the policies closely, but all I am saying is that some people will "cut their losses" and just buy a new drive. This may mean they lose out some money or don't receive the support or new drive for free.

I would recommend looking at your policy, and making the decision to either ignore it and get back up and running -OR- going to the Vendor and trying to make them honour their policy. The choice is yours.

Personally I have access to hundreds of disks, so they are quite disposable.

I am fairly certain the WD hard disk I sent back under warranty that never worked right from purchase, I had to pay the postage cost to WD. They did replace it free.

Hi @HaQue, thanks again for taking the time to answer.

I have read the policy and AFAICT the policy doesn't seem to mention any extra costs or any weird condition under which they may be charging a fee (aside for...in its sole discretion...). I haven't read about the shipment, but I am pretty sure that will be on me at least the way forward. Therefore I am going to take my chances and ship it back once I have tried to erase it and understood what is the right procedure to claim the warranty. Hope there won't be surprises..

There may be some information in the shipment instructions. I think for the WD I needed to first get authorisation, and a code to add on the label. This is what RMA means, Return Merchandise Authorisation. This may be where they say that shipping cost to the vendor is either your or their responsibility.

BTW make sure you follow the packing instructions closely - they don't have to do anything if you don't abide by the return rules.

Good Luck!

Whichever you ship it back, keep the tracing number - it comes in handy from time to time. I mention this because a couple of times, I had to trace a return to make sure it arrived safely.

_________________
"Take care of thy backups and thy restores shall take care of thee." Ben Franklin revisited

Thank you! I'll follow your advice

Do you want to know how the story ends?

Amazon reimbursed me which I really appreciate as I don't have to go through the whole warranty procedure.

What it bugged me was the fact I couldn't access the drive to erase the files before shipping it back, until today. This morning I opened windows disk management utility and there it was "offline" showing the message: the disk is offline because it has a signature collision with another disk that is online. I then put it online and voilà everything was there. I seized the opportunity, formatted it and secure erased it.

Given that this is the first time in my computing career I hear of signature collision. How come it developed a signature collision all of a sudden? I didn't installed any new drive that could have possibly interfere..

I am afraid for the end we have to wait a bit..



Wrong!! As soon as the SSD became inaccessible I loaded the last backup image (made from the SSD) on an old spare drive which is the one I am currently using. Therefore the collision was caused by having loaded the same image on two drives both online.
The question is when exactly the drive went back to life?

I bought a new Samsung SSD and transferred the system from the old drive to the SSD with Samsung data migration tool.

Again I had a signature collision, I didn't know yet what exactly a signature collision entailed, so I set once again online the disk.
The moment you put online an offline drive with signature collision, windows forces a new random signature in the MBR causing the BCD entries to become orphaned.

What I can do now is re-migrate my system on the SSD at the expense of its lifetime or recover the original signature.

I am trying to go for the second. Here is a how to with disk part.

My only doubt: being the syntax of a disk signature {<DWORD> | <GUID>} (e.g. {A2911A08-36FC-11EA-9C97-1C1B0D2B4F9B} ). Which part of the syntax do I have to assign to make the signature change successful?

I read the signature is a is four-byte identifier, in the above example should be it A2911A08? The disk part command to change signature is "uniqueid disk ID=[NEW SIGNATURE]". Would the new signature be A2911A08 only or A2911A08-36FC-11EA-9C97-1C1B0D2B4F9B or any other combination?

Any help is greatly appreciated
Thank you