On Oct 20, 2009, at 3:20 PM, Arjen de Korte wrote:
Citeren Alan Stern <[email protected]>:
Is there anything else I can do on my end? Or is this with the
nut developers at this point?
Well, there's nothing the kernel developers can do about it. :-)
The nut developers should try to follow up on the anomalies I found
in
the usbmon trace.
I'm afraid there is little the NUT developers can do either. If we
can't retrieve reports that should be available (according to the
report descriptor) and instead get 'broken pipe' from libusb, there
is little more we can do than trying to reconnect to the device.
What might be a problem here too, is that one should make sure that
either one instance of usbhid-ups is running at the same time (if
only one UPS is attached) or that each one uniquely matches a single
device (if multiple UPS are attached). Especially when debugging,
make sure to kill of any remaining instances to prevent them
battling for the same device.
Arjen,
could this be a result of setting pollfreq too low?
http://article.gmane.org/gmane.comp.monitoring.nut.user/5043
If this is not how pollfreq works, then I stand corrected - but if the
device is fully polled once every 30 seconds, and a power outage
occurs in that 30-second window, I would think that NUT would get an
interrupt report.
_______________________________________________
Nut-upsuser mailing list
[email protected]
http://lists.alioth.debian.org/mailman/listinfo/nut-upsuser
