PFsense machine getting very hot with CPU around 100%



  • Hi Guys,

    I've installed PFsense 2.2.5 64 bit on a  HP thin client GT7720 with 2.3GHz AMD Turion X2 ZM-84, 3 Gb Ram and extra dual gigabit Broadcom NIC, but I've seen that CPU is always around 100% and the machine starts to get very hot, Has anyone have similar experience or could advise me what to do to solve this?
    I'm a noob on PFsense, so no much experience on BSD, so please guide me to solve this issue.

    Regards,
    Crisman.



  • Look at Diagnostics: System Activity to see what's using your CPU.

    Also turn on PowerD and select the AMD thermal sensor in System: Advanced: Miscellaneous.



  • A few minutes ago I turned on my PFsense machine and after a few minutes with the CPU at 100% the values went to normal around 10-20%, but the temperature was going up and stayed around 85-90 ºC, I've turned on PowerD and the AMD thermal sensor but the temperatures won't go below that last values, very strange??

    Before I moved to PFsense I was running IPCop and IPFire and the temperature don't go more than 55 ºC.

    So, what else could be done?

    Thanks.



  • 10-20% CPU is still way too much for an idle pfSense. My AMD G-T40E with less than half the performance of your CPU is using 3-5% CPU during light load.

    Again, look at Diagnostics: System Activity to see what's using your CPU.

    Also, your thin client does seem to have poor cooling. Lots of dust in it?


  • Banned

    It doesn't matter if 10-20% CPU is too much. If you CPU is 85-90ºC, then replace the failed fan, remove those five tons of accumulated dust, remove dead spiders, mice etc. from the case, re-apply thermal grease on CPU, etc.



  • The problem is not on the fan, the Thin Client is almost new (3 months), without any dust and the fan runs cool, if before installed PFsense I had installed other OS and even the Windows version for thin client and the machine won't get so hot, the problem should be in some drivers maybe from the FreeBSD, I had no problems with linux and windows, so whats causing this on PFsense?

    Thanks.


  • Banned

    Dude, it's irrelevant. If your CPU being used heats it up to 90ºC, your problem is shitty cooling.



  • Sorry but I don't agree with you, last week I was testing IPfire (linux) and the CPU never goes more than 55 ºC even with teh CPU around 70%. It must be something else, later at home I will install Windows and run some stress tools to see if the machine gets so hot, but I still say that it must be something on the kernel drivers for FreeBSD relating that hardware that it not working.

    Thanks.


  • Banned

    And it still doesn't matter, your CPU/case cooling is stil the same shit. Look: it's just shit:



  • It's not shit but it's laptop cooling and regulation. Look for fan speed settings in BIOS. It's not uncommon that laptops get heat problems under load with other OS:es than Windows.



  • I have this problem with one of my main systems. For mine its just reporting. It has IPMI so I can check the temperatures a different way. CPU temp reported by the GUI is 66 C and the temp reported by IMPI is 32 C. When I was on 2.1.5, it reported the temperature correctly, or at least a lot closer.



  • @pLu:

    It's not shit but it's laptop cooling and regulation. Look for fan speed settings in BIOS. It's not uncommon that laptops get heat problems under load with other OS:es than Windows.

    The problem is that when the computer boots the CPU has heavy load (90-95%) then after 3-4 minutes goes down until 10-20% load and the temperature still goes up, I will replace the Fan with a more powerful and see what it happens.

    Thanks.



  • 20% load.. WHAT is causing that? You still haven't answered that question.



  • @Harvy66:

    20% load.. WHAT is causing that? You still haven't answered that question.

    Hi,

    That's because I was using Squid Proxy but when I uninstalled it the CPU was almost all time idle but in fact that hasn't changed the high temperatures.

    BTW: I've installed just for testing this weekend the OPNSense, I couldn't see the temperatures on the GUI but the machine was WARM no so hot that it almost burns my fingers when I've installed PFSense, so this could not be a FAN problem.

    Thanks.
    Crisman


  • Banned

    This is still extremely amusing. Of course it IS shitty cooling design problem. What happens when your CPU is used? The machine will melt the plastic, or what? Horrible fan, horrible heatsink, horrible airflow in the case. Why people would buy crap like this goes beyond me. RMA the shit. The whole thing is like a poorly designed laptop without LCD.



  • @crisman:

    BTW: I've installed just for testing this weekend the OPNSense, I couldn't see the temperatures on the GUI but the machine was WARM no so hot that it almost burns my fingers when I've installed PFSense, so this could not be a FAN problem.

    sysctl -a | grep temp
    

    How can it not be a fan problem?



  • Hi
    I don't think it's a fan problem, but we need evidence of what processes are running and what your CPUs are busy doing.
    Here is mine for example… (Diagnostics menu / system activity)

    last pid: 85388;  load averages:  0.00,  0.01,  0.00  up 171+01:28:55    20:04:35
    151 processes: 5 running, 120 sleeping, 26 waiting

    Mem: 35M Active, 292M Inact, 258M Wired, 52K Cache, 279M Buf, 3357M Free
    Swap: 8192M Total, 8192M Free

    PID USERNAME PRI NICE  SIZE    RES STATE  C  TIME    WCPU COMMAND
      11 root    155 ki31    0K    64K CPU3    3  25.4H 100.00% [idle{idle: cpu3}]
      11 root    155 ki31    0K    64K CPU2    2  25.4H 100.00% [idle{idle: cpu2}]
      11 root    155 ki31    0K    64K RUN    1  25.4H 100.00% [idle{idle: cpu1}]
      11 root    155 ki31    0K    64K CPU0    0  25.3H 100.00% [idle{idle: cpu0}]
    81695 root      22    0  223M 31856K piperd  0  0:00  0.68% php-fpm: pool lighty (php-fpm)
        0 root    -16    0    0K  192K swapin  0  2:09  0.00% [kernel{swapper}]
      12 root    -92    -    0K  416K WAIT    0  1:17  0.00% [intr{irq24: bge0}]
        6 root    -16    -    0K    16K pftm    0  1:15  0.00% [pf purge]
    9196 proxy    20    0  220M  105M kqread  1  0:53  0.00% (squid-1) -f /usr/pbi/squid-amd64/local/et
      12 root    -92    -    0K  416K WAIT    2  0:47  0.00% [intr{irq25: bge1}]
      12 root    -60    -    0K  416K WAIT    3  0:30  0.00% [intr{swi4: clock}]
      23 root      16    -    0K    16K syncer  0  0:27  0.00% [syncer]
      12 root    -88    -    0K  416K WAIT    0  0:21  0.00% [intr{irq16: uhci0 uhc}]
    46258 root      52  20 17136K  2348K wait    0  0:20  0.00% /bin/sh /var/db/rrd/updaterrd.sh
    24844 root      20    0 12456K  2128K select  3  0:13  0.00% /usr/local/sbin/apinger -c /var/etc/apinge
    20836 root      20    0 16804K  2304K bpf    2  0:11  0.00% /usr/local/sbin/filterlog -i pflog0 -p /va
    59435 root      20    0 14656K  2336K select  1  0:09  0.00% /usr/sbin/syslogd -s -c -c -l /var/dhcpd/v
      15 root    -16    -    0K    16K -      0  0:09  0.00% [rand_harvestq]

    If you can do the same, we can see where the CPU cycles are being used, which will cause the CPU to warm up.


Log in to reply