[Nut-upsuser] USBDEVFS_CONTROL failed cmd usbhid-ups

[Arjen de Korte]

Citeren Arjen de Korte <nut+users op de-korte.org>:

> This is clear, the old UPS doesn't seem to use the input reports (at  
> least not very frequently), so the chances of collisions with  
> feature reports are less.

But looking at the data again, I also noticed that the new UPS doesn't  
use them very often either. So the question is, if this is caused by  
the UPS or not? Both devices are connected to the same (host) system,  
so this probably rules out hardware problems with the USB (host)  
interface or the libusb library we use. So really the only thing that  
is different, is the UPS devices themselves *and* the loads that are  
connected.

What I saw from the logs, is that the load is very different:

MGE UPS SYSTEMS ELLIPSE (old):
    UPS.PowerSummary.RunTimeToEmpty = 2637
    UPS.PowerSummary.PercentLoad = 10 %

EATON ELLIPSE (new):
    UPS.PowerSummary.RunTimeToEmpty = 1200 or 1500 seconds
    UPS.PowerSummary.PercentLoad = 24 or 25 %

The higher number of input reports for the EATON ELLIPSE is probably  
caused by small variations in the load, between 24% and 25% (where it  
reports 1500 or 1200 seconds of runtime left respectively).

In the short interval for the logs for the MGE UPS SYSTEMS ELLIPSE,  
there was no change at all. And if there would have been, chances are  
that with such a tiny load there won't be any change in the runtime,  
so the UPS wouldn't bother to send an interrupt report for that. This  
might be the reason why you see such a different behavior between the  
two.

If possible, could you swap the loads attached to both devices and  
check if this also swaps the problems? It would help to run them in  
debug mode for a couple of hours to see what happens.

Best regards, Arjen

PS  I expect to receive an EATON Ellipse UPS in the coming days, so  
that (hopefully) I'll be able to reproduce the problem you're seeing.
-- 
Please keep list traffic on the list