Igor

Network manager is taking care of the network by default. Not ifupdown. That's why interfaces is empty.

There were some issues with CPU speed change and governor. At least on similar Allwinner H3 chips. Try to edit /etc/default/cpufrequtils and change governor, set min and max to the same value. Remember to restart cpufrequtils service or reboot.

There is always room for improvement -> https://www.armbian.com/get-involved I will certainly not touch this anytime soon.

1. Your logs: armbianmonitor -u ? 2. Your boot loader environment: printenv ?

Camera problems are off topic, outside technical support - "on board hardware only" and outside my focus. Use search 1st.

Check here: https://github.com/armbian/config https://github.com/armbian/config/blob/master/debian-config-jobs#L423-L430 Our defaults are Network manager and hostapd for AP. I am not very familiar with NM way of setting AP but I think hostapd has more features and is more resilient. We also use our own patched version. We don't use stock Debian/Ubuntu virtually since day one. For your use case it is difficult to say. Hostapd is harder to work with, but you have more control. I doubt you need that in case driver is done by the book. First we need to get to the bottom of the problem why it doesn't work. Is it the driver and why, is it Network manager? Trying complementary tools before jumping into the code is usually better ...

Upstream Broadcomm drivers could also not be in a perfect condition. You have to ask their maintainers. It could be some known regression/problem. Unrelated to Armbian: https://duckduckgo.com/?q=brcmfmac+ap+mode Edit: you can also try setting AP with hostapd to rule out Network manager problems. For hostapd way you need to tell Network manager not to use this particular wireless device ...

Read my post again. Everything from /boot Wrote on mobile

Ahaa. There was a problem with a kernel upgrade. You must have removed /boot mount ... copy content from /boot to SDcards /boot and reboot. Keep SDcards /boot mounted under /boot ...

This means nothing, in supplied logs everything looks fine. Overlayfs support is enabled as module (and all other dependencies) and you will need to debug Docker in details why it doesn't work. Is overlayfs module loaded? Make a new image from sources and see if everything works there. If not, file an issue that Docker doesn't work. I doubt we can check it in less than one month.

With reboot you have much better chances. Unloading/loading on this historical kernel 3.4.y ... not sure. This driver is buggy. Did you check if the module name is dhd at all? I don't recall from my head now, but we had: ap6210, ap6211, ap6212 and dhd for different variants of this AP62xx chip. Old kernel, which you are trying to use, is messy in any case. Modern kernel has brcmfmac and no special modes. It should work STA/AP out of the box.

https://forum.armbian.com/search/?q=op_mode=2

Older kernel wireless driver (for chip on Opi zero +2) needs to be loaded with op_mode=2 otherwise AP doesn't work.

One great options is to help make it. I am pretty sure that all current maintainers have a big pile of hardware and very little time. More load is almost impossible to add ... I talked with @balbes150 about this problem long time ago. He is having boxes, I am having boards. Have it or have a free access to them. This is still the same. We as a community, not just he and I, wants to see project develop further. We (those few people who stands out with their work) can't produce more and more. It doesn't go that way ... it's more, more, more ... and once nothing more That is bad for all. The one wants to work but he can't and community lost one of the prime contributors. If you or someone else wants to see this and that board/hardware supported, become its maintainer. It's a long term micro job and we can arrange samples or other random boards from our (over)stock and donations cuts for compensation/motivation. Making board officially supported means answering to its problems over longer period of time. Someone has to take that prime responsibility and someone has to serve as a backup. IMO current maintainers could jump on backup roles at best.

Why? I got confirmation on Twitter: That's fine for "No official support" images.

Probably this: https://www.cyberciti.biz/faq/howto-regenerate-openssh-host-keys/

Have some troubles which need to be addressed. https://github.com/armbian/config/issues/33

OMV works, NC cloud as well. Both can't work unless you use some kind of virtualization. Problem is unrelated to Armbian. Case closed.

Network manager, which works perfectly fine on my Cubox / router combo, both kernels: http://ix.io/1qbb

We tried to freeze it once but failed I hope next time there will be more luck. I just hate this Rockchip BSP mess P.S. I remake most of the images yesterday because desktop upgrade was broken. I can't do it for Rockchip without losing some of the functions ... At least not the way we build things now. We plan to enable images rebuilding from kernel which are already stored in repository but there is some build script RFC needed. ASAP.

No, you have to pick one of. Both suits the same purpose as @Tido pointed out and use the same low level services.

Huh, than we might have some regression here. I can check this in 1-2 weeks fastest.

I can understand your reasoning but am not sure about that. We just removed something what wasn't working properly. Software boutique is broken under arm or arm64. I don't recall where exactly. And slow on most of small boards. If you want to know, https://www.armbian.com/search Removing half broken app is always better than leaving it inside. User can always try to install it on his own and if it doesn't work he doesn't blame us but himself. If it is there and don't work they usually put a pressure on us. I would certainly recall if there were more complaints than before. It was my idea to have some kind of software install system on the desktop, but we would need to fix them, use our own version ... we can't afford that.

What could this tell? That we are ignorant? Lazy? Bored? Our project resources are extremely underpowered and staff is under tremendous pressure. Reaching breaking points ... We are just a few, while user count is getting into insane numbers. We need to scale it down ... or we will soon need to tell people to fuck off on a daily base This is a non-commercial project and support costs are already sky high. We can't pay for them, you don't want to. Once again, from my perspective. If you want this to be solved (for you and few others) before 2019/2020 you have to do it (i did explain you how) or hire someone to do it instead of you. We simply don't jump around and wait on people's wishes. Sometimes yes, but mostly not. Project is too large. Many people, including me, are from that era. Debian was chosen to cover most people, but we can't and don't want to please them all.

OMV installation works. '+ owncload is almost certainly not possible without sever hacking.