The short story of a Windows refugee

General history:

I'm a PC nerd / gamer who has been trying to escape from the usual suspects. Recently, driven by economic constraints and the increasing aggressiveness of corporations to force as many people as possible into any subscription models, forced. I came to IT out of personal interest and worked there for many years, for example as a mid-tier corporate sysadmin, system architect, system engineer. What you had to do in a company that says they are like a family ;D

I was mainly active in the enterprise world, i.e. Windows, HP, Cisco – the usual suspects. To be honest, for a long time I thought linux was a nice gimmick that is to be welcomed, because it might be an alternative at some point (apart from later more professional solutions such as mail systems and hypervisors).

Technical history:

For several years I have been running an UNRAID server based on SATA, mainly as a file server, Docker/Plex media server. Very successful, I have never had major problems with unraid that I could not have solved myself. While looking for alternatives to the conventional raidcontroller/enterprise disk/server operating systems that I had been using until then, I came across unraid.

Because it was intended as an interesting test project at the time, I used older, used hardware to build this system - i7 3xxx, ddr3, consumer board, cheap sata controllers not ideal. I was pleasantly surprised how satisfying and trouble-free this system ran for years and quietly developed from a test system to the main server. That was probably a problem, as it turned out now, because the inconspicuousness lulled me into a false, naïve sense of security.

Since I have more time than money in the last few months (; D) I came up with the glorious idea to finally make the server "right". From economically rosier times, I still have a lot of hardware that I wanted to use for this, namely a low end server board with low power i5, some enterprise HDDs and an enterprise PCIe NVMe SSD. I thought to use these "treasures" from better times to create a more professional system and move my data there. Fortunately, I didn't come up with the stupid idea to do a hotswap, the old server continues to run happily – under conditions that I would describe as a harsh endurance test environment.

Creation of the new media server:

So I procured additional materials, small parts and finally the UNRAID license, a financially exhausting process for me that took some time. Assembling the server

- bequiet silentbase 601

- Supermicro Board 8x SATA, 2x M.2, 4x GBit

- i5-9xxx

- 16GB RAM Non ECC

- 7x 10TB HGST

- Power supply 850W (I still had that and you know how it looks with the hardware prices)

went surprisingly smoothly, almost suspiciously ;D I hadn't thought of a Pool Cache SSD at first. In the old server, several sata ssds run as a pool cache so inconspicuously that I did not focus on this aspect at first. I initially installed unraid as a test system, but then used the pre-Christmas offer for full licensing at short notice. The creation of the base array went just as smoothly (2 parity, 5 data).

What happened:

I finally came to the Docker topic, mainly for the media server. As mentioned, I used plex so far, because of their "professionalization" of the license model and out of technical interest I wanted to try jellyfish. So I needed a cache ssd and wanted to use the board-side m.2 slot for it. Since I have already switched several clients to linux in the last time, my stocks are meagre and I don't need to complain about the cost development of nand memory here. A decommissioned m600 of dubious reliability (I have already lost one, a week after the warranty expired) was still there, I thought it was right to experiment for the time being. Just make a backup as soon as possible, I thought.

She lasted 3 days and took all the experiments with jellyfish with her.

Cache drive missing docker unable to start.

Annoying but something like that happens, I thought. XD I wanted to continue with the docker/jellyfish tests but there was no expendable m.2 nvme available and so I came up with

my preciousss

Into the game. A samsung enterprise pcie nvme 6.4tb an artifact from better times, which I always dealt with like raw eggs, not least because it cost an arm and a leg even when it still existed and was considered affordable.

Unfortunately, I had reformatted it from ntfs to ext4 just before, that will be important. XD

So my sweetheart installed, fits the motherboard as if they were made for each other, temperatures are fine. I left the filesystem at ext4, that will be important. So I recreated pool cache, reinstalled docker, rebuilt jellyfish databases and continued to play around with it. It lasted 3 days then the cat came. The silentbase case has the power switch at the top of the case, very exposed, and my cats are heavy enough to trigger a 4 second forced shutoff. This is exactly what she did, during an ongoing parity check and while reading a jellyfish library.

Annoying but usually not a problem, I thought.

Cache drive missing, unable to start docker.

I waited for the parity check that was triggered by the forced shutoff to at least have a clean array parity before I took care of the cache pool further (and secured the power switch against further animal attacks)- 12h later…

After a clean reboot, the SSD was back, theoretically. Practically the filesystem was destroyed and could not be reformatted within unraid. All attempts to get the SSD to work also failed. In linux and windows.

Status at the current state of information:

I lost very valuable hardware from a combination of

- Animal misconduct :)

- Linux ext4 in combination with the filsystem history of the ssd

- Own lack of corresponding skills combined with naïve trust in open source solutions

What did I learn from it, technically:

- Linux is not to be trusted in general, in particular the use of Ext4 in connection with NVMe disks seems to be an almost malicious trap. I am stunned how casually a simple power failure or an unclean shutdown can lead to a permanent defect of hardware, which is actually already secured against such things by itself. I'll put it this way: with ntfs this wouldn't have happened (or so i think at least) ;)

What I learned from it, meta level (not so positive):

- If you have a problem with hardware in connection with Windows, you have options, there are still accommodating people and in the worst case you could (at least theoretically) get paid support. If you have a problem in the linux world you are really alone, even more you expose yourself to the most vicious community imaginable, which tends to use your problems to damage you psychologically, at least that was my strong impression in the course of my research on Linux file systems, possibilities for superblock rescue and such things. "Haha, you would have made a backup" "There's nothing you can do, just throw it away"

- Private computing was yesterday, you have two options today: Throw your data into the throats of faceless corporations that see you as livestock that can be farmed without moral ethical concerns – but in return you keep your data in access longer, it just doesn't belong to you anymore.

Or rely on open source, that can work, but it doesn't have to work – better not rely on it and don't think that anyone will help you.

Ps

Please excuse the shortcomings of an automatic translation and the sarcasm that may occur in places. The loss hurts me very much - and of course I have backups ;)

Edited December 14, 2025Dec 14 by 2STRonin

8 hours ago, 2STRonin said:

Or rely on open source, that can work, but it doesn't have to work – better not rely on it and don't think that anyone will help you.

You’re mistaken here. Commercial just offers you someone to will ignore you or miscommunicate for weeks on end and still get no actual help.

Software companies like Unraid manage to split the difference, with actual help coming from both sides, but you still will have to handle the work yourself.

Shit happens they say.

I agree, its hard to build a cat-safe computer, we also have a bunch of cats, they examine everything :-)

Anyway, you are somewhat wrong, "Linux" is NOT an operating system, its just a Kernel. Everything around it comes from other people/companies and yes, its a mess.

These "distributions" differ alot, so what you read from Dist-X may not be true/valid/good for Dist-Y.

UNRAID is its own distribution (based on others), which makes it unique.

But I see hope for your insane SSD.

Put it back into a windows system, open diskpart, select the drive and do "clean all" (this may take a long time, dont become nervous). Or use tools like "wipedisk" or something.

At the end, the SSD will be free of all ext4 remains and can be freshly formatted and used in UNRAID. Use zfs or something the next time. Ext4 is not well supported in UNRAID yet (ok, I guess you have learned this part already now)

You’re mistaken here. Commercial just offers you someone to will ignore you or miscommunicate for weeks on end and still get no actual help.

Software companies like Unraid manage to split the difference, with actual help coming from both sides, but you still will have to handle the work yourself.

Just to clarify, I didn't want to blame anyone here. And yes paying consultants or experts to be misunderstood and ignored sounds familiar to me XD

Put it back into a windows system, open diskpart, select the drive and do "clean all" (this may take a long time, dont become nervous). Or use tools like "wipedisk" or something.

At the end, the SSD will be free of all ext4 remains and can be freshly formatted and used in UNRAID. Use zfs or something the next time. Ext4 is not well supported in UNRAID yet (ok, I guess you have learned this part already now)

Yes, I came, saw, and failed. ;) Windows is very unhappy with the SSD. DiskManager takes a long time to open the devices, never a good sign. On the hardware side, everything seems fine, the device reports everything green via status LEDs, also reports correct values but Smart cannot be read. DiskPart Clean works but doesn't help, Clean All reports a write protection that cannot be removed. Initialization in Disk Manager acknowledges with "Device not ready" MK2FS cannot write to sector 0 and therefore cannot create a homogeneous FS. Something is worse but my skills are not yet enough to say what exactly.

1 minute ago, 2STRonin said:

Initialization in Disk Manager acknowledges with "Device not ready" MK2FS cannot write to sector 0 and therefore cannot create a homogeneous FS. Something is worse but my skills are not yet enough to say what exactly.

Sounds like you have "bricked" it. This happened quite often in the early days of SSDs. The Firmwares were not really stable and could be killed "beyond recovery".

Ask the vendor ("samsung"?) if there is a de-brick tool for this particular device. You man need to send it in (which obviously may give you additional costs, a new disk may be cheaper then).

Maybe a firmware update is available and you can flash it yourself.

I think it's bricked too but I insist :not by me ;D

A repair from the manufacturer is unlikely (EOL) and certainly does not make economic sense. A comparable new device is far far beyond my current possibilities. At the moment I see it as a very expensive lesson to keep everything of value away from "Linux", especially NVMe Tech. I don't know yet whether a firmware flash from me is possible at all, in general it is very difficult to get any information about it from Samsung.

6 minutes ago, 2STRonin said:

in general it is very difficult to get any information about it from Samsung

Yeah I know, but I'm afraid this is the only chance you have currently.

6 minutes ago, 2STRonin said:

but I insist :not by me ;D

Killing cats wont solve your problem 😁

Edited December 15, 2025Dec 15 by MAM59

No, no, I can't be angry with the cat (well, let's say at least not for long and not seriously). As described, it is a chain of unfavorable circumstances and I admit at least a partial blame because I simply naively brought such a valuable device into play without knowing about the apparent shortcomings of Ext4 in connection with NVMe controllers (Trim):) NTFS seems to spoil you too much, something like that hasn't happened for a long time.

edit: To be honest - my guilt is bigger than the cat's XD I'm actually old enough to know that you don't put things in uncertain danger that you can't replace or where the loss has a high impact. ;)

Edited December 15, 2025Dec 15 by 2STRonin