I have now found what is actually sending the messages into every single shell. The reason is that the kernel is using "Emergency" for these messages and rsyslog has following rule in /etc/rsyslog.conf: # Emergency messages to everyone logged on (wall) *.emerg :omusrmsg:* After commenting out the rule, the messages are still logged, but will not spam every single shell any more. I think that should limit the problem down a bit: Some change in or around kernel 4.4.49-16 is now sending emergency messages when we trigger thermal events. The messages directly out of the kernel are misleading and have the level emergency: 2017-03-09T17:12:09.755648+10:00 gerald6 kernel: [123760.500209] mce: [Hardware Error]: Machine check events logged 2017-03-09T17:12:09.755649+10:00 gerald6 kernel: [123760.500224] mce: [Hardware Error]: CPU 3: Machine Check: 0 Bank 128: 0000000088030003 2017-03-09T17:12:09.755655+10:00 gerald6 kernel: [123760.500238] mce: [Hardware Error]: TSC 12d27289fb91e 2017-03-09T17:12:09.755656+10:00 gerald6 kernel: [123760.500249] mce: [Hardware Error]: PROCESSOR 0:306a9 TIME 1489043529 SOCKET 0 APIC 3 microcode 1c The messages are then interpreted by mcelog (this is the latest git) and again logged into syslog and look sensible: 2017-03-09T17:12:09.773831+10:00 gerald6 mcelog[17636]: Hardware event. This is not a software error. 2017-03-09T17:12:09.774797+10:00 gerald6 mcelog[17636]: MCE 0 2017-03-09T17:12:09.774977+10:00 gerald6 mcelog[17636]: CPU 3 THERMAL EVENT TSC 12d27289fb91e 2017-03-09T17:12:09.775133+10:00 gerald6 mcelog[17636]: TIME 1489043529 Thu Mar 9 17:12:09 2017 2017-03-09T17:12:09.775288+10:00 gerald6 mcelog[17636]: Processor 3 heated above trip temperature. Throttling enabled. 2017-03-09T17:12:09.775438+10:00 gerald6 mcelog[17636]: Please check your system cooling. Performance will be impacted 2017-03-09T17:12:09.775582+10:00 gerald6 mcelog[17636]: Running trigger `unknown-error-trigger' 2017-03-09T17:12:09.775726+10:00 gerald6 mcelog[17636]: STATUS 88030003 MCGSTATUS 0 2017-03-09T17:12:09.775868+10:00 gerald6 mcelog[17636]: MCGCAP c09 APICID 3 SOCKETID 0 2017-03-09T17:12:09.776004+10:00 gerald6 mcelog[17636]: CPUID Vendor Intel Family 6 Model 58