Repeated corruption of SDcards

[nochkin]

Why vfat? Was it crashing before you started doing it? I assume so, but wanted to make sure.
Anyway, vfat is not considered as a reliable fs. It's just being used due to wide compatibility only, not for reliability.

If you mount the whole /var/log on your flash drive, then you can at least check it by reading your drive on a PC.

It would be also interested to see what's on the HDMI output, but I guess the Pi is running headless on your side.

[pz1]

Why vfat?

Because then I can read the logs on my PC.

[pz1]

Your initial message indicated that you checked sda1 and sda2, and both are clean.

Correct, and yet it refused to boot. So what is the value of a file system check?
Next time it crashes I will do the check once more, and also try to reboot with a HDMI screen connected to see how far it gets on this "clean" disk.

[nochkin]

Your initial message indicated that you checked sda1 and sda2, and both are clean.

Correct, and yet it refused to boot. So what is the value of a file system check?

The value is to check file system corruptions and inconsistencies.
If it was checked clean, that means your file system is probably fine and the issue could be somewhere else.

I know this is a step which you may not be able to do, but I'll try anyway:
When it crashes next time, please try to connect it to a monitor over HDMI to see what's on the screen.
If you reboot, you may see some errors too.
The source of the crash could be somewhere at the higher level (Z-Way bug?).

[pz1]

I know this is a step which you may not be able to do, but I'll try anyway:
When it crashes next time, please try to connect it to a monitor over HDMI to see what's on the screen.
If you reboot, you may see some errors too.

I think I said that I intended to do that in one of my earlier replies. As a matter of fact it is my top prio now.

[nochkin]

Oh, I probably missed that. Sorry for that, we went through too many messages :-)
Let me know what you'll find.

[pz1]

Unfortunately after 13 days I had to reboot the production machine. So a break in the durability test.

The good news is that the rig rebooted without any problems. In the past year or so that would not have happened!

The major difference is that as reported before I now do write the log files to an USB disk.
With this reboot, I took the opportunity to also make the first (boot) partition of the SDCard read_only. I still do use swapfile in memory on this production Pi. I am a little bit more hopeful now

[nochkin]

Interesting. We'll see how it goes. Thank you for update.

[pz1]

In the meantime I have made the usage of a USB-stick a bit more robust by binding it via its unique identifier. A little later we had a power outage of a quarter of an hour. The system came up again without a problem. Progress!

Looks like the USB disk addition has solved the corruption problem, but I remain uneasy not knowing the real cause. As a scientist I should revert the configuration and see what happens, but I lost inspiration to go through that again

[nochkin]

I feel you. Sometimes the stability is more preferred over curiosity