Sunday, June 7, 2015

On my MacBook Pro Retina 13" with Fedora 22 I received notifications about system errors.

This is the error which triggered the alert:

mce: [Hardware Error]: Machine check events logged

The common advice on the interwebs was to run mcelog, but that didn't show anything.
It turned out that the reason for that is that in Fedora runs it as a daemon and sends the output to syslog.

With journalctl /usr/sbin/mcelog I could see all the errors.

An example event:
apr 05 21:41:32 localhost.localdomain mcelog[891]: MCE 1
apr 05 21:41:32 localhost.localdomain mcelog[891]: CPU 2 THERMAL EVENT TSC 21059c66c67
apr 05 21:41:32 localhost.localdomain mcelog[891]: TIME 1428262844 Sun Apr  5 21:40:44 2015
apr 05 21:41:32 localhost.localdomain mcelog[891]: Processor 2 heated above trip temperature. Throttling enabled.
apr 05 21:41:32 localhost.localdomain mcelog[891]: Please check your system cooling. Performance will be impacted
apr 05 21:41:32 localhost.localdomain mcelog[891]: STATUS 88000bc3 MCGSTATUS 0
apr 05 21:41:32 localhost.localdomain mcelog[891]: MCGCAP c07 APICID 1 SOCKETID 0
apr 05 21:41:32 localhost.localdomain mcelog[891]: CPUID Vendor Intel Family 6 Model 69
apr 05 21:41:32 localhost.localdomain mcelog[891]: Hardware event. This is not a software error.


This happened when all CPU cores were on 100%.