On 02/02/2016 22:11, Constantine A. Murenin wrote: > On 2016-02-01 1:07, Roy Marples wrote: >> On 30/01/2016 19:39, Constantine A. Murenin wrote: >>> In general, I personally don't think it ever makes sense to shutdown >>> by default when the temperature is exceeded, since most of these >>> sensors aren't really all that reliable (especially if you're getting >>> them over i2c, with potential bus locking issues and race conditions >>> with BIOS / IPMI; getting a bit sidelined, at the very least, the >>> sensor values should be dampened, which is what's done in OpenBSD's >>> sensorsd, not sure if anything similar is done here). >> >> I disagree with this. >> On my IBM ThinkPad z60m, the fan is so knackered it has to be on it's >> side when compiling. If the BIOS thinks it's too hot then emit's some >> warning bleeps. If it gets hotter then it powers itself down. >> I would much rather the system shutdown gracefully compared to having >> the power rudely yanked. > > Wouldn't the correct solution then be to kill the process-intensive > jobs, instead of shutting down the whole system?
How would you handle a non process intensive master spawning process itensive workers to replace the ones you just killed? Roy
