[SOLVED] possible to configure kernel to completely avoid using a core of multicore processor?

dnp · 12-27-2019, 06:45 AM

i just got the following,

Code:

syslog:Dec 27 07:09:46 carpdiem kernel: [33152702.203117] [Hardware Error]: Corrected error, no action required.
syslog:Dec 27 07:09:46 carpdiem kernel: [33152702.203277] [Hardware Error]: CPU:2 (15:2:0) MC2_STATUS[-|CE|MiscV|-|-|-|-|CECC]: 0x98144000010c0176
syslog:Dec 27 07:09:46 carpdiem kernel: [33152702.203430] [Hardware Error]: MC2 Error: VB Data ECC or parity error.
syslog:Dec 27 07:09:46 carpdiem kernel: [33152702.203792] [Hardware Error]: cache level: L2, tx: DATA, mem-tx: EV

if things become worse or repetitive, is it possible to configure my kernel/system to just not use this one core?

i'm on Slackware 14.2
and uname -a shows

Code:

Linux carpdiem 4.4.14 #2 SMP Fri Jun 24 13:38:27 CDT 2016 x86_64 AMD FX(tm)-8320 Eight-Core Processor AuthenticAMD GNU/Linux

pan64 · 12-27-2019, 11:09 AM

I'm not really sure if it was a core (or something else).

Quote:

Now what does that really mean?
*shrug*, could be firmware/drivers/overheating/poor-CPU-seating/DIMM-seating/faulty-motherboard/faulty-CPU/faulty-DIMM.

https://centosfaq.org/centos/kernelhardware-error/

additionally:
https://www.linuxquestions.org/quest....php?p=5742874

rnturn · 12-27-2019, 11:48 AM

Quote:

Originally Posted by dnp

i just got the following,

Code:

syslog:Dec 27 07:09:46 carpdiem kernel: [33152702.203117] [Hardware Error]: Corrected error, no action required.
syslog:Dec 27 07:09:46 carpdiem kernel: [33152702.203277] [Hardware Error]: CPU:2 (15:2:0) MC2_STATUS[-|CE|MiscV|-|-|-|-|CECC]: 0x98144000010c0176
syslog:Dec 27 07:09:46 carpdiem kernel: [33152702.203430] [Hardware Error]: MC2 Error: VB Data ECC or parity error.
syslog:Dec 27 07:09:46 carpdiem kernel: [33152702.203792] [Hardware Error]: cache level: L2, tx: DATA, mem-tx: EV

if things become worse or repetitive, is it possible to configure my kernel/system to just not use this one core?

Is the same CPU involved in every occurrence of these error messages?

There are tools that let you define what CPUs an individual process runs on but, AFAIK, they don't setup up a means to disable a specific CPU globally. (Probably because the kernel folks don't recommend it.)

You could try using the "isolcpus=2" boot time command line to keep CPU2 from being used by the scheduler. I think I'd use this as a last resort, though. See the file "/usr/src/linux/Documentation/admin-guide/kernel-parameters.txt" for more.

BUT... if this is really infrequent and it's non-fatal -- it looks like the system successfully recovered from the error, after all -- I would just keep an eye on it for now. If it begins happening so frequently that your logs are growing like crazy, then try the "isolcpus" option.

HTH...

dnp · 12-27-2019, 12:08 PM

Quote:

Originally Posted by rnturn

You could try using the "isolcpus=2" boot time command line to keep CPU2 from being used by the scheduler. I think I'd use this as a last resort, though. See the file "/usr/src/linux/Documentation/admin-guide/kernel-parameters.txt" for more.

BUT... if this is really infrequent and it's non-fatal -- it looks like the system successfully recovered from the error, after all -- I would just keep an eye on it for now. If it begins happening so frequently that your logs are growing like crazy, then try the "isolcpus" option.

thank-you, this sounds pretty much what i was hoping existed.
and, yes, i hope it's infrequent or a random sunspot occurance