On Wed, 2010-08-18 at 10:41 -0400, Steve Blackwell wrote: > On Wed, 18 Aug 2010 08:44:16 +0300 > Gilboa Davara <gilboad@xxxxxxxxx> wrote: > > > On Tue, 2010-08-17 at 13:08 -0400, Steve Blackwell wrote: > > > On Wed, 18 Aug 2010 02:12:16 +0930 > > > Tim <ignored_mailbox@xxxxxxxxxxxx> wrote: > > > > > > > On Tue, 2010-08-17 at 12:05 -0400, Steve Blackwell wrote: > > > > > I've been looking at my logs some more. I don't understand these > > > > > messages: > > > > > > > > > > Aug 17 10:30:50 steve kernel: CPU0: Temperature above > > > > > threshold, cpu clock throttled (total events = 455) > > > > > Aug 17 10:30:50 steve kernel: CPU1: Temperature above > > > > > threshold, cpu clock throttled (total events = 455) > > > > > Aug 17 10:30:50 steve kernel: CPU1: Temperature/speed normal > > > > > Aug 17 10:30:50 steve kernel: CPU0: Temperature/speed normal > > > > > > > > And the CPU overheating as well as your hard drive? > > > > > > > > Is the computer in a hot room? Are the fans working? Is the > > > > ventilation blocked? Is the computer wedged in between things > > > > that restrict airflow? Are things full of fluff and dust? > > > > > > > > > > > Well it would seems so but I don't trust the messages. It doesn't > > > seem reasonable that the CPUs go overtemp and then immediately cool > > > down enough to be OK. > > > > Actually it is possible. > > Your CPU has auto-throttle support. Read: When the CPU passes a > > certain temperature threshold, it automatically clocks down (or > > inserts NOPs) in-order to prevent is from burning out. Never the > > less, if your machine's cooling is sufficient you shouldn't see this > > message. > > > > If you CPU's high and low water mark are the same (E.g. 90C), the CPU > > will reach 90C, throttle, and drop to 89C - all in one second. > > I'd suggest you configure lm_sensros and monitor the CPU and board > > temperature. > > $ sensors-detect > > $ /etc/init.d/lm_sensors restart > > $ sensors -s > > $ sensors > > > > - Gilboa > > P.S. can you post your hardware configuration? > > > > Running sensors-detect produced the same /etc/sysconfig/lm_sensors > file that I already had. Running sensors shows this: > > # sensors > atk0110-acpi-0 > Adapter: ACPI interface > Vcore Voltage: +1.42 V (min = +1.45 V, max = +1.75 V) > +3.3 Voltage: +1.68 V (min = +3.00 V, max = +3.60 V) > +5.0 Voltage: +1.62 V (min = +4.50 V, max = +5.50 V) > +12.0 Voltage: +11.98 V (min = +11.20 V, max = +13.20 V) > CPU FAN Speed: 56250 RPM (min = 0 RPM) > CHASSIS FAN Speed: 0 RPM (min = 0 RPM) > POWER FAN Speed: 0 RPM (min = 0 RPM) > CPU Temperature: +62.0°C (high = +90.0°C, crit = +125.0°C) > MB Temperature: +49.0°C (high = +70.0°C, crit = +125.0°C) > Power Temperature: +24.0°C (high = +80.0°C, crit = +125.0°C) > > The first thing that jumps out at me is that I think I need a new PSU! > How is this machine even running if 3 of the 4 voltages are low? Doubt it. Look at the CPU fan speed. Most likely lm_sensors fails to convert the raw sensor readings to real values. I'd suggest to reboot the machine, and check if your machine's BIOS has sensors/monitor tab / menu. If it has, you may be able to view the actual PSU voltage and fan speeds from within your BIOS. > > The second thing is that the temps are just fine. So why do I keep > getting these messages in the logs? Perhaps because the power rails are > low? Doubt it. Temperatures looks OK. > > The chassis and power fans are 2 wire so no data. OK. > > lshw dumps a lot of information. Anything in particular you are looking > for? CPU, motherboard make / model, PSU information, memory configuration, etc. > > I think I have solved my lockup problem. I'll write a separate post > about that. Good to hear. - Gilboa -- users mailing list users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
