Nut upsuser - Apr 2017 - usbhid-ups: Failed to open device, skipping. (Permission denied)

Charles Lepple writes:
> On Sun, Apr 2, 2017 at 2:59 AM, Sam Varshavchik <mrsam at
courier-mta.com>
> wrote:
> > Having trouble configuring nut 2.7.4 on Fedora 25, with a new Tripp
Lite
> > UPS.
> >
> > I'm running XFCE, and XFCE's Power Manager sees the UPS just
fine. lsusb
> > gives:
> >
> > Bus 006 Device 002: ID 09ae:3016 Tripp Lite
>
> I'm still catching up on mail, so apologies if this was covered on a
> more recent GitHub issue, but Tripp Lite PID 3016 is extra-sensitive
> to certain combinations of USB cables and newer motherboards. If you
> take NUT and XFCE power manager out of the equation, but the UPS still
> generates a lot of disconnect/timeout/reconnect traffic in the kernel
> logs, you might be up against this issue.
>
> Full gory details here: https://github.com/networkupstools/nut/pull/122
This particular new unit came with its own USB cable, and this unit is  
plugged into a relatively old server. The motherboard is about ten years old.

I have a lot of noise in my syslog. grepping for 'usb' or '3016'
doesn't
find anything useful.

I've getting better results with the last patch from issue 414, which fixes
the bad error handling that was causing the driver to bail out. The driver  
still yells about EAGAIN/EIO, once every four minutes or so:


ioctl(10, USBDEVFS_REAPURBNDELAY, 0x7ffd4d210fe0) = -1 EAGAIN (Resource  
temporarily unavailable)
sendto(3, "<31>Apr 1 12:46:10 usbhid-ups[4983]: libusb_get_report:  
Input/output error", 75, MSG_NOSIGNAL, NULL, 0) = 75

but with the patch it recovers and keeps running. With the last patch from  
414, the driver has now been running for over 24 hours, just quietly barking  
about these timeouts, and reconnects, to anyone who'd listen.

I am open to suggestions for grep food, but I can't find anything in my  
syslog that looks like a kernel complaint about actual USB connectivity.

My weekend project is to build the patched libusb-1.0 branch, and see what  
happens there.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL:
<http://lists.alioth.debian.org/pipermail/nut-upsuser/attachments/20170407/1aacc140/attachment.sig>

On 04/07/2017 09:15 PM, Sam Varshavchik wrote:
> Charles Lepple writes:
>
>>
>> Full gory details here: https://github.com/networkupstools/nut/pull/122
>
> This particular new unit came with its own USB cable, and this unit is
> plugged into a relatively old server. The motherboard is about ten
> years old.
>
I also had this problem with this unit.  My hacky solution solution is
here: https://github.com/sdgathman/trippfix

I'm convinced that the USB interface on these units is braindead -
despite their other excellent qualities (excellent handling of
outages).  Many other people have the same problem with this unit,
including Windows users with the supplied proprietary monitoring
software, so I think it is hardware.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: OpenPGP digital signature
URL:
<http://lists.alioth.debian.org/pipermail/nut-upsuser/attachments/20170407/5fb86b7b/attachment.sig>

Stuart Gathman writes:
> On 04/07/2017 09:15 PM, Sam Varshavchik wrote:
> > Charles Lepple writes:
> >
> >>
> >> Full gory details here:
https://github.com/networkupstools/nut/pull/122
> >
> > This particular new unit came with its own USB cable, and this unit is
> > plugged into a relatively old server. The motherboard is about ten
> > years old.
> >
> I also had this problem with this unit.  My hacky solution solution is
> here: https://github.com/sdgathman/trippfix
>
> I'm convinced that the USB interface on these units is braindead -
> despite their other excellent qualities (excellent handling of
> outages).  Many other people have the same problem with this unit,
> including Windows users with the supplied proprietary monitoring
> software, so I think it is hardware.
What would be the symptoms of this that I would be seeing? I don't see  
anything in syslog, only usbhid-ups getting a periodic timeout, but each  
time it successfully reconnects via USB.

It is not entirely unprecedented for manufacturers to quitely fix defects in  
their hardware, then simply continue to ship it as the same model. It's been
about a week since this unit has been plugged in, with the only issue being  
these usbhid-ups timeouts, which it now apparently recovers from it.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL:
<http://lists.alioth.debian.org/pipermail/nut-upsuser/attachments/20170408/3e372a74/attachment.sig>