[Nut-upsuser] UPS Product additions to NUT HCL list for compatible Tripp Lite UPS systems
clepple at gmail.com
Wed Oct 16 01:32:57 UTC 2013
On Oct 15, 2013, at 3:40 PM, Eric Cobb wrote:
> My name is Eric Cobb and I am a Product Specialist with Tripp Lite. We have tested your latest NUT distributions against our HID USB compatible UPS systems and would like to know what is needed and/or who is needed to contact in order to updated the listing of compatible Tripp Lite UPS systems in your HCL list?
From end users of NUT, we usually request the information mentioned under "Report new devices, omissions or errors" at http://www.networkupstools.org/stable-hcl.html#footnotes
However, I don't want to be pedantic about getting the exact output of 'upsc' from the entire Tripp Lite product line - especially if several models share the same USB controller. Getting the 'upsc' output from a representative model should be sufficient.
What would be helpful is to understand which models are similar, and whether there are any attributes we can use to differentiate between various generations of the hardware (for instance, at some point the SMART2200RMXL2U changed from the proprietary serial-over-HID to standard PDC HID).
Also, we occasionally run across cosmetic errors in the output data: http://thread.gmane.org/gmane.comp.monitoring.nut.devel/6482
While this shouldn't affect the shutdown sequence, we do try to correct the returned measurements if we can identify a good algorithm. If all of the voltages are in microvolts (including the config voltages), we can apply a multiplier across the board. But if only some of the exponents are wrong, it is harder to write a general rule that won't break when the HID descriptor in the firmware is fixed. If there is a firmware revision that we can compare against, that can help isolate the bug fixes.
The last piece (not visible on the HCL, but included in the data files for hotplug/udev) is the list of USB vendor/product IDs. If a new Tripp Lite product ID were to be encountered by NUT, the driver would not automatically attach to it unless the "productid" option is specified. It is a simple way to tell users that we did not know whether that device would work when that version of NUT was released, but that they can give it a shot (potentially getting incorrectly scaled voltages). We are open to suggestions on how to improve that heuristic, especially if Tripp Lite has allocated separate ranges of product IDs for UPS and non-UPS hardware. Here's the code in question:
In that file, the product IDs are matched individually, and there is a tripplite_claim() function at the bottom of the file which tries to offer advice for other IDs.
clepple at gmail
More information about the Nut-upsuser