I wanted to log stuff to a USB dongle, so I tried a cheap 8 GB stick and after a while it got ruined, won't work anymore even in other computers.
Then I tried another cheap 8 GB stick of another brand, same story.
Then I tried my expensive 128 GB transcend memory (https://www.transcend-info.com/product/ ... tflash-920) which is more SSD-like and has better electronics. It worked for a while but after a day it got ruined, it became read-only and I can't write anything to it anymore, no matter what computer I plug it into.
While the transcend stick became read-only, the other ones can't even be read.
For example I'd put it into a windows computer, open diskpart in the terminal, select the stick, issue "clean" command which should remove all partitions, create new partitions, copy and paste some files into it using explorer and wait a few minutes, then disconnect it using safe removal. I plug the stick back in and it's as if nothing happened, all the data and partitions before the clean command was issued is still there.

Generally speaking some (good) USB sticks have in the controller some provisions to switch to "read only" in order to protect contents when a malfunction of some kind is detected.
Still generally speaking cheaper sticks tend to simply fail and become inaccessible when there is a problem.

So the behaviour you report is "normal", what is not normal at all is that the device had a short life (I don't think that you could exceed the number of writes in a short period of using it with the MIkrotik router).

For some (not all) sticks there are (leaked/unofficial) so-called "manufacturer tools", i.e. the programs that the manufacturer uses in production to initialize the devices, by talking "directly" to the controller and its settings.

Even if you can find the right program, usually there are no documentation/instructions coming with them, so even if in theory they can recover the functionality of a stick, in practice it can be very difficult if not impossible.

I have not seen other reports of similar issues with USB port of the Mikrotik Ax3, maybe you were just (very) unlucky?
Or you suspect that your particular device has a defect?

I have not seen other reports of similar issues with USB port of the Mikrotik Ax3, maybe you were just (very) unlucky?
Or you suspect that your particular device has a defect?

How to tell? It's still under warranty but I need some solid proof and testing is expensive.

I have a USB stick in AX3 for a container solution (Samsung 32Gb).
It's in there for some months now. It never failed once.

I doubt AX3 as a device is ruining that USB stick. This place would be swamped with reports if that was the case.
Also, IF (big if on purpose) that AX3 would be defect, nothing would work with that USB stick. A defect on USB level does not present itself as working for a while and then stop it. It works or it doesn't.

So that leaves config or usage.

Maybe you can tell us a bit more on what you do with it ?
How do you use it ?
How is it configured ?

I wanted to log stuff to a USB dongle

Really ??

Stuff = undefined garbage.
So we need to get some more specifics on what and how.

You could try to use a USB-tester to check if the ax3 is at least supplying a normal usb voltage (5V). If for whatever reason the ax3 delivers a higher voltage then that could probably be a reason for your USB dongles getting ruined. Although it seems unlikely, it's not impossible...

Did you also plug the read-only usb disks into a Linux pc? If so, what does the dmesg log mention about the usb sticks?

I wanted to log stuff to a USB dongle...

You "logging stuff" to a USB drive is probably kiling the flash write cycles, if you really want to do that, you should use a USB hard drive, not a Flash based memory.

Highly unlikely.
I have 2 RB5009's with USB stick I use for writing all log entries to with rotational log files.
Neither of them have failed yet (both in use for well over a year).
I assume USB driver on AX3 is the same as on RB5009 ? Same ROS, same ARM64, so ...