[Nut-upsuser] Tripp Lite 2010 results

[Charles Lepple]

On Aug 21, 2011, at 9:11 PM, Matt Behrens wrote:

> On 08/21/2011 08:56 PM, Charles Lepple wrote:
>> Thanks for the report. (In both NUT 2.6.0 and 2.6.1, this product  
>> ID is detected automatically.)
>>
>> Just to confirm, though, do your voltage readings look right when  
>> you run upsc? That ID was flagged as returning battery voltages  
>> that are low by a factor of 10.
> No, there are a few problems... battery voltage here actually looks  
> like it might be 10x high.

My mistake - I misread that part of the code. In 2.6.0 and later, we  
apply a scale factor of 0.1 to correct for that bug in the UPS firmware.

>  Also, this is my first time with nut, but I'm pretty sure ups.alarm  
> isn't supposed to say "No battery installed!" ;-)

It's not supposed to, but I suspect another firmware bug.

"ups.alarm" is not meant to be parsed by any of the monitoring clients  
such as upsmon, so if you know the battery is in fact installed (and  
the voltage seems to confirm this), you can safely ignore that.

If it still bothers you, the proper thing for us to do is figure out  
if your model truly can't tell if the battery is installed, or if it  
is inverting the sense of that signal.

We would need two debug logs to compare - one with the battery  
installed, and one without. Check your UPS documentation to be sure  
that it is okay to pull the battery connector when things are powered  
on. (This is pretty standard on high-end rack-mount UPSes, but might  
not be the case for smaller ones.)

Stop the existing NUT setup, being sure that the usbhid-ups process  
has died. Then run "/lib/nut/usbhid-ups -a ups -DD 2>&1 | tee /tmp/log- 
battery-installed.txt". I think that's the path for your version of  
Ubuntu, but in any case, no more than two -D flags should be  
necessary. Let the driver run for a minute or two, then kill it via  
Ctrl-C. Repeat with the battery removed.

Please gzip both logs before emailing them to the list. Thanks!