MaxT

Until someone dares sending PR for Current …

If the issue is so important, why no one dares sending relevant PR as suggested in that thread?

2.1.1 seems to be not the latest version existing at your post time, there is v2.1.2 in the repo.

Probably the easiest way to assess chances is to submit PR.

Developing solutions and sharing by sending PR and/or using customize image script might be the way address those.

I guess you need something similar to udev rule for Logitech Fn key: http:// https://github.com/sta-c0000/logitech-keyboard-udev-rules. IMHO it is more reliable and less resource hungry vs software package, but I have not seen alike for a Win key. So it might be necessary to spy on software doing this and develop udev rule for win key. May be someone on the forum having enough experience could help.

It would be great, if you could post results of all your findings here, so that someone else could save a few hours on this problem and rather spend them on another one;)

May be worth looking into armbian-install script? Just a thought.

I would open the box make photos of the board with silk prints visible and search similar box in a relevant section of tv boxes on this forum. Name on the plastic box means nothing, silk prints on the board and SoC itself might help.

Probably marking on the board is the only what really matters.

Thanks, very helpful. Maybe someone will be able to bring some light on how second partition could appear (uboot update script transition bug?), while being not a part of Armbian's functionality.

Context of the TV boxes section might be helpful, especially in the sticky thread re their status in Armbian.

If tester is not available, I'd try powering failing ports with their respective cables but feeding them from pins powering other ports (eg with wires from some molex cable, just need to be careful with polarity and shortcuts) and then powering working ports from pins on the board feeding failing ports - to exclude failure of the power rail feeding these ports on the board. If cables are faulty, I'd check them for electrical connections and if they are ok, would replace capacitors.

Disabling automount (drive letter assignment) in windows before flashing might help - probably it is mount time when windows is checking and changing partition table.

I wouldn't be so categorical - I have rock pi 4b, which never saw android or any os other than armbian, which has same thing- two boot devices on emmc. Since it doesn't bother me, I didn't investigate further.

If I remember correctly, recently there was a similar thread on the forum re GPT corruption under windows. Not sure, but there might be some useful info. Alternatively, flashing under Linux might be an option.

WSL might be easier if you're on windows

Either of those, I personally used Ubuntu running in ESXi VM

Are you joking? Open readme.md in GitHub repo, read it and explore links therein

As per documentation, run on a compatible host: git clone https://github.com/armbian/build cd build ./compile.sh Follow screen menus. Build system will download necessary toolchain, sources, patches, etc and build an image. Whether or not it will work depends on whether sources/patches for a particular board deviated from the hardware due to time or bugs. If someone is regularly testing then image likely will be ok, if not then image might work or not. If the board has community maintained status, then it is up to its users to test/develop/send patches in GitHub. If necessary, it is possible to play with sources, configs, develop and place extra patches, etc. At times it looks not that straightforward, so there will be a learning curve.

I guess they are free to spend that time to develop own samba implementation

It would be interesting to compare corrupt and non corrupt cards, including partition tables, etc. Windows might be playing bad with partition tables - if one only inserts ESXi boot disk into windows machine, such disk becomes non bootable. AFAIR this is because ESXi relies on records order in the gpt table of the disk, while windows strips away blank records so that if eg second and third records are blank then windows will move the fourth record to the second place, while ESXi relies on the fourth record in a table. I personally met this issue a few years ago and it took a while to find the solution of reordering gpt table records back to restore ESXi boot.

If I remember correctly, it is possible to mount such partition flashed to usb or sd within WSL (just ensure that proper support is installed into WSL) and make any changes. It is possible even to script changes to be run under windows, though I can't remember how exactly this is done. Probably AI can hint.

yeah, frustration is quite clear, but probably it is manufacturer who should be most/only interested in that. that also cost quite a lot, so it is probably better to demand where money were paid to, i guess.

AFAIR someone on this forum also experienced booting issue with orange pi board though 4 not 5 with quite similar symptoms. May be worth contacting that person for his results so far - might give some hints on what's going on and whether it's a manufacturing fault