[Nut-upsuser] Bug or feature? polling frequency = (pollfreq + pollinterval)

[Charles Lepple]

On Sep 19, 2017, at 1:36 PM, René Berber <rberber at fastmail.fm> wrote:
> My guess is that they are the same version (Synology keeps it updated).

Is there a way to check which version they are using? (Just curious - probably will not affect the problem you described.)

> Read about pollfreq, and pollinterval, but didn't see NUT's code.

The usbhid-ups code is here:

   https://github.com/networkupstools/nut/blob/v2.7.4/drivers/usbhid-ups.c#L851-L871

> The values of those are:
> driver.parameter.pollfreq: 30
> driver.parameter.pollinterval: 5
> 
> I found that power consumption (ups.realpower) only changes every 35
> seconds, which is (pollfreq + pollinterval).
> 
> Does this looks correct?  I would expect changes at pollfreq.

Since we can't control UPS latency, the polling does not have very tight timing. I would expect it to be closer to 30 seconds, though.

If you run the driver with a few "-D" options, you will see a timestamp at the beginning of each line, and you should be able to follow the various Quick and Full update cycles.

You could comment out some of the mapping lines in the mge_hid2nut table if you don't want to waste time reading certain values:

   https://github.com/networkupstools/nut/blob/v2.7.4/drivers/mge-hid.c#L1083

Note that there are some duplicates - your UPS will not have every one of those HID PDC usages (4th column).