[Nut-upsuser] New UPS Support: Eaton 5S 1000

[Charles Lepple]

On May 5, 2016, at 9:25 AM, Ken Marsh <ken.marsh at sparkpost.com> wrote:
> 
> Well, whaddya know, when run with -DDDDD it works! upsc consistently returns data. Maybe it's a timing thing, or a select affected by I/O or something tricky like that. Nonetheless, I ran it, hit upsc eaton multiple times, collected the output and attached it. If this doesn't give you what you need, I can retry with fewer/more D's or whatever you like.

Probably is timing, then. See below.

> I couldn't find any crashes in dmesg, just this over and over:
> 
> [167914.284728] usb 6-1: USB disconnect, device number 21
> [167915.292292] usb 6-1: new low-speed USB device number 22 using uhci_hcd

This part isn't good. While the usbhid-ups can attempt to recover from this, it means that for a second at a time (those times in brackets are seconds since boot), the UPS is logically disconnected from the USB bus. (This is what happens if you have a phone with a worn-out USB connector, and it beeps if you wiggle the plug.) After it reconnects, due to the low bit-rate of low-speed USB, it takes 5-10 seconds for the driver to reconnect.

Some people have reported success with higher-quality USB cables (2.0 rated, even though the UPS will never use the faster modes). Others have moved the cable to a non-3.0 port. A hub might also clean up the signal, but then you need to make sure the hub power supply is on battery as well.

You should only see the "new low-speed USB device" message once for the UPS, when it is first plugged in (or when the system is booted).

-- 
Charles Lepple
clepple at gmail