On Jan 12, 2020, at 12:58 PM, Gene Heskett wrote: > > But this doesn't pass the smell test, 3 other machines that are trained > to reboot when power returns and have no ups fitted, were NOT rebooted > as evidenced by their reported uptimes, one of which shows a 144 day > uptime. So why should this have been logged via a -wall bc to every > login it is servicing? > > There was no power failure IOW.
The thing is that many UPS models will report "on battery" when the voltage fluctuates out of range, but a switching power supply might still be able to coast through it-- especially if the supply is oversized for the load at the time of the abnormality. Since both "outages" (really, the on-battery states) were exactly 10 seconds long means that the UPS was probably trying to "debounce" the voltage fluctuation by not switching back to line power immediately after the voltage went back to normal. For APCs and MGEs, I think it is typically 5 seconds. Several people have rigged up longer timers with upssched to not send the wall message until it has been on battery for, say, a full minute. _______________________________________________ Nut-upsuser mailing list [email protected] https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/nut-upsuser
