Hi Chris,
David C. Rankin wrote:
I wonder why systemd doesn't do this by default?
It's not systemd to blame. The timer unit files are shipped by the respective projects, like util-linux, man-db, mlocate, shadow, logrotate, ...
systemd.timer(5) says AccuracySec= ... Within this time window, the expiry time will be placed at a host-specific, randomized, but stable position that is synchronized between all local timer units. This is done in order to optimize power consumption to suppress unnecessary CPU wake-ups. So I think it is systemd's `fault' that they all collide. I think its theory is flawed. The CPU here will happily move from doing nothing to doing slightly less for a unit and back again without getting hot and altering the fan speed. But ig the typical timer units collide then the fan speed moves up through the gears and keeps going at more than idle speed for some time, long after the units have finished and the CPU clocked back down. That must impact power consumption too. -- Cheers, Ralph.