Closed PierreRustOrange closed 9 months ago
Hello, This is a very weird problem indeed. I don't remember having a problem with this CPU model on Centos 7. I see that this machine uses a Real-Time Kernel. I never tested the sensor on this kind of kernel. I will do some tests with a machine that have a very close (Intel Xeon Gold 6130) model of this CPU and get back to you.
Thanks for your feedback, it's quite weird indeed. I've ruled out the CPU : we have another server with the exact same CPU where the sensor is running fine. It's probably an issue with the OS / kernel thus.
When I look into /sys/devices
I don't even have a msr
sub-directory !
Hello,
The problem is the missing support of the msr
perf_event PMU by this kernel version.
The libpfm library throws this error because it needs to read the /sys/devices/<pmu>/type
and /sys/devices/<pmu>/events/<event>
files to setup some PMUs.
I tested your kernel-rt version (3.10.0-693.2.2.rt56.623.el7
) and the /sys/devices/msr
is also missing and I have the same error as you. I also tried with the next version (3.10.0-693.11.1.rt56.632.el7
) but it wasn't working either.
The closest version of the kernel-rt where the /sys/devices/msr
is present and the sensor works correctly is the 3.10.0-957.1.3.rt56.913.el7
. The latest version of the kernel-rt (3.10.0-1160.42.2.rt56.1182.el7
) works too.
Unfortunately, it seems that you cannot fix this problem without upgrading the kernel of this machine.
Thanks a lot for this analysis ! I'll see if I can upgrade the kernel but that might be complicated ...
I have an issue on a server where the sensor fails with following message :
Could not get encoding for event 'TSC' : code -4
.Issues similar to this one have already been raised, but it is not the same problem than in #1 or #25 as the sensor is built here with the patched version of libpfm4 and I have been using the same sensor's container image successfully on other servers.
I suspect it has something to do with the version of the kernel and / or the generation of cpu used, but I couldn't find anything obvious in the libpfm4 source.
This is how the sensor is started:
And here is the full output of the sensor :
The CPU on this serveur is a Xeon Gold 6142, here is what
lscpu
returns:any idea ?