I install Debian jessie vanilla kernel 4.1.6

At the first loading gives out a mistake: failed to bring up eth0. Ip the address isn't given on dhcp.

But on my router it is specified correctly.

I can't change settings. First start. the system expects reset and connection. That it is possible to make in this case.

Repeated reset and expectation doesn't yield results. The keyboard doesn't work

Power cycle helps, reset doesn't.

I tried to disconnect and include power supply 3 times. But for some reason doesn't help. Thanks, I will try once again

Another issue is a bit weird but you will need to have a working keyboard for that. Some boards hangs if rt clock setting is invalid so you need to set it manually. I guess old kernel 3.4.108 is working ok?

 

BTW: USB works only if you power via power connector.

Sorry, I incorrectly told. The keyboard and usb works perfectly. CTRL+alt+del reboots system. Yes, the old kernel works without problems. The IP address is given and everything works.

At me is Debian on the computer. I tried to connect a memory card to it and to register static ip the address manually. In file/etc/network/interfaces so:

auto lo

iface lo inet loobback

auto eth0

iface eth0 inet static

address 192.168.1.34

netmask 255.255.255.0

But it didn't help. The mistake repeats. Or I was mistaken and it is necessary to specify the address in the file:

 / etc/network/interfaces.default?

If I understand correctly your eth0 is not initialized properly. This is a know issues on some boards ... cut the power off then on (not reboot). Something setting a real time clock is a workaround solution to this.

 

Perhaps u-boot hardware initialization needs some optimization.

 

Setting an fixed IP won't solve this problem.

Whether this problem can be connected with that on an old kernel 3.4.108 I executed a script cd /root ./nand-sata-install to transfer system to ssd. The script was executed. The system asked reset. But there was a set of mistakes and cyclic reset. After that I copied a new image of jessie vanilla kernel 4.1.6 on the card and I try to establish it.

Perhaps therefore there is a mistake?

Whether it is worth formatting my ssd a disk?

Your SSD, partition 1 must be prepared otherwise it doesn't work.

 

Running from SSD give some speed boots.

I understood, I will return then on an old kernel 3.4.108

Hi,

 

I put the latest armbian image on my sd card and boot. The installation is stuck with the ssh key generation for more than 15 minutes. I tried again (downloaded image again, flashed again, ...) same result.

 

From the manual I know this should take about 3 minutes. What is going wrong?

 

Thanks, 

H.

- which hardware and which image exactly

- do you have internet connection

- anything connected to the board?

- which hardware and which image exactly

- do you have internet connection

- anything connected to the board?

 

- Cubietruck, Debian Wheezy Vanilla, Version 4.1.6

- Network is connected, however the device is not visible when I scan the network

- a SATA drive, a keyboard and a screen (obviously =) are connected

 

PS: Screen went black in the meantime without any further notice.

Update: Testing without the harddrive and keyboard. It behaves the same.

I download exact image and burn it to the SD card. Everything is normal.

 

If device fails to init - cut the power and try again.

 

If screen goes off just like that, than power supply can be a problem.

 

P.S.

Some boards have odd problem - you need to set RTC clock otherwise you could end up without eth0.

hmmm, ... odd. 

 

Would it be possible that the speed of the sd card somehow causes this behavior? I only have an old class 2 card...

You assumption about the RTC clock was right. I could see some debug prints on the boot screen that say something like eth0 couldn't be enabled cause of RTC issues.

 

I don't really understand this. The device runs since 1.5 years without any problem. 

This bug emerged within last few months and affect only some boards. I am not aware if this is already fixed, perhaps in recent u-boot.

 

Since my board works well is hard for me to "fix the problem which does not exists"

 

However, we can do this together. I can build few test builds or you or anyone else can try to get an working combination. Most likely we should pick some older u-boot and lost some functions.

 

PS

I am moving later this thread to Allwinner subforum and pin it.

I'd be happy to help. 

I spread a screenshot with a mistake.

Here such situation. Maybe someone will prompt as to correct?

Access to file system is. I can make changes

Now login, type following commands and tell me what do you get:

 

hwclock -r 

 

date -s "2015-01-01 10:10:00"

 

hwclock -w

 

Reboot and see if it works.

I have almost same problem. In the next branch.

How to make set RTC clock otherwise?

Last night I already googled for RTC clock reset (or so). The post said you need a power source of some specific voltage and you connect that to certain pins. I am very sure I would brick my device when trying that as I do not have the things needed but bad luck ^^ Unfortunately, I don't find that post at the moment...

 

 

But honestly I wonder if those two problems, the ssh key generation and the RTC causes the eth0 to be down, are really related. I do not see any connection.

Thanks for idea, but I doubt that matter in energy. My cubieboard 3 steadily works with an old kernel. Such mistakes aren't present there. But it is worth trying also it

Problem in that that I can't enter anything into consoles. the system expects reset.

Consoles aren't present. Teams aren't entered. Reacts only to ctrl+alt+del.

If you tell as to make that teams were entered I will make it

So it doesn't finish the first run procedure?

 

It should, even without a network.

Yes. procedure doesn't come to the end. I waited for hour. After reset everything repeats. Stops on the same place. Thus problems two:

1. Doesn't receive ip the address.

2. Constantly expects reset and there is no console

I see, I need to rework this.

 

I think you have console on Jessie.

I will be glad to hear any help or council that the console started working

So Vadim and me seem to have the same issue, right?

 

But I have an old but running installation that boots from nand... What would happen when I perform the reset procedure Igor described above within the running installation, shut down and insert the sd card? 

 

PS: what does this actually do?

 

hwclock -r --> read time from hardware

date -s "2015-01-01 10:10:00" --> set OS system time

hwclock -w --> set hardware clock to OS system time?

Yes, the issue is the same.

 

Also now I am able to recreate it.

 

When I boot to kernel 3.4.108 and soft reboot to 4.1.6 almost always results in this error.

 

Working on fix.