Pine64, Ubuntu Xenial, ownCloud

[jean_bruder]

Dear Pine64 users,

My fresh installed Pine64 is running Ubuntu Xenial after some strange crashes when trying to do a system upgrade : I had to made all the update using the serial console ... Now, everything is working fine, except ... ownCloud !

Apache2 is installed with PHP7/MySQL. A simple PHP file displaying the phpinfo() informations works fine.

Now ownCLoud 9 is installed and properly configured, according to the online owncloud installation manual. The setup seems to run fine (using the MyDSL DB), but the system become unresponsive on both the SSH remote console and the serial console when the owncloud pages starts to get displayed. No way to get any debug messages.

How can the system get debugged in order to try to understand what happens ?

Many thanks in advance,
Cheers,
Jean

[longsleep]

Your Pine64 is most likely crashing because of undervoltage.

[jean_bruder]

Dear longsleep,

First of all, I have to thank you for the kernel you wrote for the Pine64 board ! The Pine64 is another nice tiny board for playing with Linux :-)

Can you please explain that "under-voltage issue" : I use a standard 5V, 2Amps power supply. Should I found a specific one (have you some referrals to provide) ?

I found yesterday some interesting things : The web server is hanging while providing to my browser some JS scripts files, after serving many other content before. During this time one core of the CPU reach between 90 and 100% CPU usage for a short time (max. 5 secondes), then stop responding ...

Is there any way to set the CPU governance mode to "On Demand", like on some phones ? Maybe it will help to avoid such an issue ?

Many thanks in advance for you reply,
Best regards,
Jean

[rahlquist]

Can you please explain that "under-voltage issue" : I use a standard 5V, 2Amps power supply. Should I found a specific one (have you some referrals to provide) ?

I found yesterday some interesting things : The web server is hanging while providing to my browser some JS scripts files, after serving many other content before. During this time one core of the CPU reach between 90 and 100% CPU usage for a short time (max. 5 secondes), then stop responding ...

He is expressing the theory that as your cpu works harder to meet the demand you mention, its drawing more current to do the work. When thats happening his theory is your power supply trying to supply the current you need is causing a voltage drop which then causes the cpu to drop below operational levels and stop responding.

You could test if its bad JS or the power supply by finding something else to load the cpu a bit and see if it locks up then too. Just dont burn it up by running it at 100% for hours without a heatsink

[tkaiser]

You could test if its bad JS or the power supply by finding something else to load the cpu a bit and see if it locks up then too. Just dont burn it up by running it at 100% for hours without a heatsink

This is a non issue if the OS image in question has the latest fixes applied since we did some extensive work to improve thermal throttling behaviour. So best advice is to update kernel and u-boot as follows:

Code:

bash <(curl -s https://raw.githubusercontent.com/longsleep/build-pine64-image/master/simpleimage/platform-scripts/pine64_update_uboot.sh)
bash <(curl -s https://raw.githubusercontent.com/longsleep/build-pine64-image/master/simpleimage/platform-scripts/pine64_update_kernel.sh)
bash <(curl -s https://raw.githubusercontent.com/longsleep/build-pine64-image/master/simpleimage/platform-scripts/pine64_fix_whatever.sh)

In case the stress utility can be installed by running 'apt-get install stress' then checking for undervoltage is easy: 'stress -m 2 -c 2'. If the board powers off it's undervoltage.

[jean_bruder]

Thanks to all of you, my ownCloud server is now stable and up'n'running :-)