On Sun, 18 Oct 2009, Justin Piszcz wrote:
Thanks Alan,
I have performed all of the steps you recommended, there is also a bug in
the docs, the directory seems to be different, or at least if you have
everything compiled into the kernel and not using modules.
Start nut:
# /etc/init.d/nut start
Starting Network UPS Tools: upsdrvctl upsd upsmon.
Within ~20-30 seconds it already broke:
============================================================================
Broadcast Message from nut@xxxxxxxxxxxxxxxx
(somewhere) at 17:22 ...
Well, the usbmon trace shows that wherever the problem lies, it isn't
in the kernel. There were a few problems in the device itself -- it
was queried for a report and did not provide any data. The nut program
should be smart enough to retry when this happens, but it didn't.
It did try to re-establish communcation with the device, but it appears
that nut got caught in a loop during this attempt. During this loop
nut did nothing but transfer a bunch of string descriptors from the
device, over and over again at 2-second intervals. The loop lasted for
almost a minute, until another device error occurred. Then nut
recovered properly -- until a third error happened and nut got stuck in
its loop again.
Alan Stern