[Nut-upsdev] NUT I-D: Minimum command and variable set for conformance

[Manuel Wolfshant]

On 1/3/22 13:47, Roger Price wrote:
> I received a comment calling for a minimum command and variable set to 
> be supported for conformance.
>
> The proposal assumes that in the future there will be UPS units which 
> natively include an Attachment Daemon and which support direct 
> internet access from a Management Daemon.  Something like figure 3 in 
> section 3 of the I-D: 
> https://www.ietf.org/archive/id/draft-rprice-ups-management-protocol-05.html#name-ups-and-attachment-daemon-i 
> Such a UPS could claim to be RFC compliant, a marketing advantage, and 
> the comment is that such a claim should require that at least a useful 
> subset of the commands be implemented.
>
> The comment proposes as the minimum set: « Battery state 
> (charging/charged), UPS on main power/UPS on inverter, battery 
> percentage of charge left, and killpower (commanding the UPS to turn 
> itself off) at the least. »
>
> The idea of a minimum command and variable set seems to me to be an 
> interesting idea, but we need to look closely as what is involved.
>
> 1. An Informational RFC cannot lay down conformance standards, but can 
> discuss a "Recommended minimum support for commands and variables"
>
> 2. The minimum set is for a consumer grade appliance. Professional use 
> requires much more to manage for example unattended servers.
>
> To encourage discussion, I propose the following new subsection 9.4 
> Recommended Minimum Support:
>
>  9.4 Recommended Minimum Support
>
>  The features provided by current UPS units vary widely.  However 
> experience
>  shows that a minimum feature set is needed for satisfactory use of 
> the NUT
>  Project software.  For a full list of variables see the Recording 
> Document
>  (7.1).
>
>  9.4.1 Desktop PC Variables
>
>  The following variables form a minimum set suitable for Desktop PC.  
> It is
>  expected that on public power supply failure, the PC will be halted.  
> It will
>  not restart automatically when power returns.
>
>  battery.charge
>  battery.charge.low
>  device.mfr
>  device.model
>  ups.status with statuses OL OB LB FSD
>
>  9.4.2 Unattended servers, additional variables
>
>  The following additional variables are needed in a minimum set 
> suitable for an
>  unattended server.  It is expected that on public power supply 
> failure, the
>  server will be halted.  It will restart automatically when power 
> returns.
>
>  battery.date
>  device.serial
>  ups.delay.shutdown
>  ups.delay.start
>
>  9.4.3 All systems
>
>  Satisfactory use of the NUT Project software requires support for all 
> commands.
>
> All comments welcome.  If the list is agreed, I will add the new 
> subsection 9.4 "Recommended Minimum Support" to the I-D
>
> Roger

A few comments here

- I have not seen any low- and mid-end UPS that knows about killpower so 
not including it as a requirement is a good idea

- The 3KVA Ablerex that I have handy and which powers since ... a long 
time ago ( I've replaced 3-4 times the batteries, at least ) a small 
size DC does not know about battery.date and since it is connected via a 
Serial-to-USB adapter, I am not sure that it knows about device.serial. 
It probably does, but in lsusb -v I can only see the USB adapter

wolfy