Home / ubuntu / Questions / 328142
In Process
zulu34sx
Asked: 2013-08-04 00:25:23 +0800 CST

Machine stops running randomly

  • 772

I recently installed Linux MINT 15 64bit as a guest in a UBUNTU 12.04 LTS host. Everything is working just fine, except that the complete machine turns off completely (the host, physical machine) and the screen turns black. Has anyone an idea how to fix (or even find) the problem? Is this failing hardware or has it to do with the virtualBox-thing? Any help appreciated.

If I pause the guest when CPU gets over 65°C, the machine doesn't freeze anymore. My CPU is not overclocked. Is there any way to allow only a specific percentage of CPU-usage (or only a specific freq) to prevent overheating? Is there any chance do anything in Hardware without changing the CPU?

Note: the built-in CPU-limiter in vBox doesn't work.

virtualbox

2 Answers

  1. I COULD be wrong, but it could be a hardware issue. You need to provide more details. Full system specs, CPU, GPU, PSU, HDD, and RAM. As much information about these as you can. So don't do anything drastic like buy a new PSU or CPU based on this advice!

    It's possible a power fluctuation is causing the trouble. If your processor is overclocked, it could easily be causing this. It's common for overclocked processors to run fine under normal conditions and then cause your computer to randomly crash. This is caused by power fluctuations, and THOSE are caused by electrical noise of a variety of sorts. Your PSU doesn't provide a perfect stream of power, and under certain conditions, it fluctuates more, sometimes too much, and the CPU shut down when an incorrect voltage is received.

    Similar problems can arise with any component in your computer, the two most common I find, being the RAM and the CPU. My RAM has to be manually revolted to 1.65V, because that is it's specified range, but my Mobo automatically sets it to 1.5V for whatever reason. Cleaning up glitches like that will stabilize your system.

    If you have extra components, or another PC, try swapping parts (PSU, RAM, GPU) as well and see if any of that stops your random crashing. It could be a voltage glitch, or your PSU may need to be replaced for all we know.

    • 0

  2. This could be at least two problems:

    1. Many BIOS have a setting to halt the CPU (to prevent damage to the CPU) once it reaches a certain temp.

    2. There's issue which plagued me for months with an HP P7-1240 (AMD A10-5700) where the default 'ondemand' CPU governor can result in a hang. I switched to use 'conservative' instead and the problem has not reoccurred.

    You need to either build a kernel which does this by default or change it ASAP once you login.

    One problem I am seeing is that even if I build the kernel that at some point soon after I login it will be changed back to 'ondemand'. I don't know what is changing it. Changing it back to 'conservative' seems to 'stick'. [1]

    From the documentation for CONFIG_CPU_FREQ_GOV_CONSERVATIVE (emphasis added):

    'conservative' - this driver is rather similar to the 'ondemand' governor both in its source code and its purpose, the difference is its optimisation for better suitability in a battery powered environment. The frequency is gracefully increased and decreased rather than jumping to 100% when speed is required.

    If you have a desktop machine then you should really be considering the 'ondemand' governor instead, however if you are using a laptop, PDA or even an AMD64 based computer (due to the unacceptable step-by-step latency issues between the minimum and maximum frequency transitions in the CPU) you will probably want to use this governor.

    EDIT: [1] Found it. This is /etc/init.d/ondemand:

    root@care:/tmp# sysv-rc-conf --list ondemand
ondemand     2:on       3:on    4:on    5:on

    disable it:

    root@care:/tmp# sysv-rc-conf ondemand off
root@care:/tmp# sysv-rc-conf --list ondemand
ondemand     2:off      3:off   4:off   5:off
    • 0