Comment 35 for bug 1337200

> My new theory is dbus-daemon is spinning out of control because of upower, probably because indicator-power is still subscribed to it.

That, and powerd as well for the record.

Thanks for the logs!

 * udevadm.log doesn't look surprising: We get one event for each of the four power_supply devices every 10 s. While sending events for all those (AC, full battery, etc.) is a bit stupid, this is a frequency which we should be able to keep up with easily. There are no bursts of events at any time.

 * upower monitor is surprising: while its "changed" events for ac, line_power_usb, and line_power_wireless pretty much exactly reflect the uevents, the ones for battery_battery happen much more frequently, about once per second. I don't know why that happens (this would now require starting upowerd in the foreground and watching its debug output). This is the place where upower 0.99's "have properties actually changed" behaviour helps immensely, even though it doesn't explain alone why upower itself wakes up so often. If the "wakes up once per second" still happens with 0.99, we need the debug info. But still, there are no bursts here which seem to correlate to an app waking up.

 * dbus-monitor.log shows roughly the same number of events, but it doesn't contain timestamp information, so it's not possible to see bursts there. To do that, please run dbus-monitor --system | ts -s '%.S' (see comment 23).

At this point, is there a major reason to not update upowerd to 0.99 and pull in the corresponding powerd, system-settings, and indicator-power changes?