Unique use for HA on a live music tour and community input question about Networked UPS backups

Configuration
1

I am the Video and Networking Engineer for a live music tour (the band is Pretty Lights for those who are interested, and we stream all of our shows live at Twitch ). One of the challenges that I face is that we have UPS battery backups back stage and at our FOH (mix position in the crowd) and occasionally we encounter power outages (usually when someone accidentally unplugs a cable or if a circuit breaker blows). The UPS is obviously there to keep critical systems and network components powered in these events. The problem we have is that we cannot hear the UPS screaming for help over the music.

I have purchased some LED strobes that I’m trying to set up to notify us that there’s a power failure on the UPS systems. My thought is to use HA to turn on those strobe lights which can be easily done with a smart plug. The challenge is that there is no integration for either the APC or Cyberpower UPS systems that we are using. Each of the 3 UPS systems have network interfaces but Im struggling to get data from them and I’m hoping that someone in the community may have come across these before and might have worked out how to pull data from them.

I’m using Two (2) APC SMT3000RM2UC 2700w Rack Mount Backup & Surge Protectors that have APC “SmartConnect” which will email me about a power failure but sometimes we don’t have the best internet connection on site and again the loud and bass heavy environment sometimes makes noticing phone notifications difficult.

We also have One CyberPower OR500LCDRM1U 300w UPS with the optional RMCARD205 UPS Management Module installed which provides Ethernet connectivity.

I know that there is a NUT integration for Home Assistant but none of these UPSs are connected to computers via USB and to be honest I’d rather poll the UPS directly via it’s Ethernet port if possible.

My fallback solution would be to use an extra smart plug parallel to where each of the UPS systems plug in and poll that smart plug and when the plug becomes unavailable then trigger the smart plug with the strobe to turn on but that doesn’t seem as elegant of a solution as getting data straight from the UPSs themselves.

If anyone has any advice or suggestions I’d really appreciate it, thanks!

2

I think in a case like this I would probably lay a shielded category 5 twisted pair cable from the UPSs to the FOH and then have a laptop placed there with some surveillance running over the UPS status.
A cat5 can be up to 100m between active equipment and you might need a switch near one of the UPS too to connect all your UPSs.
The switch should be secured by a UPS too and the laptop in the FOH would probably have its own battery acting as UPS.
The laptop makes it possible to run different software including HA in as a guestOS in a VM and then have the HA client running in the hostOS.
Cat5 cables should not be that expensive and if it is shielded, then electro-magnetic noise will not be a problem too. WiFi and other wireless connection in a concert setup might be a bit risky. There can be a lot of noise around power cables and much of the wireless gear used by musicians, such as microphones, probably also use the same wireless frequency spectrum.

3

The APC UPS daemon or SNMP won’t connect to them?

4

I don’t believe the APC UPS supports SNMP, I do think that the CyberPower one does though. I’ll do some experimentation with SNMP

I did see this page you linked me to the other night when I first started researching this and after re-reading it just now I saw that the mention of the NIS server attachment. I might be able to get that to work, I was just hoping to have a more “out of the box” solution with a direct integration for the UPSs that utilized just their ethernet connections natively without the need to run another server and daemon combination.

Thanks for your feedback, you may have pushed me in the right direction!

1 Like
5

Connectivity to where the UPS is located isn’t an issue for me, I have a 50gbps fiber network from backstage to FOH, I’m just trying to avoid adding extra computers/hosts in an already very crowded environment. My hope was that someone already had dealt with these devices before and knew how to talk to them directly.

I do also have a robust wifi network being used for devices and sensors used by the show already so adding smart plugs as my sensors is probably the easiest path forward if I can’t get data straight from the UPSs themselves.

1 Like
6

NUT is probably the way forward then.

https://networkupstools.org/stable-hcl.html

User have reported three different APC Smart-UPS 3000 models and its a bit tricky to figure out which one is similar to yours, but maybe try to the program and see.
https://networkupstools.org/ddl/APC/index.html

7

I run nut in a docker container. It was messy running nut on the host OS and docker makes it more portable.

In your case I think it will allow you to have a single host running multiple nut containers(one for each ups) and HA container.

There is no official NUT container so I will edit the post with what i use.

2 Likes
8

I got below from somewhere I modified to my needs.

entrypoint.sh

#! /bin/sh -e

if [ -d /run/secrets ] && [ -s /run/secrets/$SECRET ]; then
  API_PASSWORD=$(cat /run/secrets/$SECRET)
fi

if [ ! -e /etc/nut/.setup ]; then
  if [ -e /etc/nut/local/ups.conf ]; then
    cp /etc/nut/local/ups.conf /etc/nut/ups.conf
  else
    if [ -z "$SERIAL" ] && [ $DRIVER = usbhid-ups ] ; then
      echo "** This container may not work without setting for SERIAL **"
    fi
    cat <<EOF >>/etc/nut/ups.conf
[$NAME]
        driver = $DRIVER
        port = $PORT
        desc = "$DESCRIPTION"
EOF
    if [ ! -z "$SERIAL" ]; then
      echo "        serial = \"$SERIAL\"" >> /etc/nut/ups.conf
    fi
    if [ ! -z "$POLLINTERVAL" ]; then
      echo "        pollinterval = $POLLINTERVAL" >> /etc/nut/ups.conf
    fi
    if [ ! -z "$VENDORID" ]; then
      echo "        vendorid = $VENDORID" >> /etc/nut/ups.conf
    fi
    if [ ! -z "$SDORDER" ]; then
      echo "        sdorder = $SDORDER" >> /etc/nut/ups.conf
    fi
  fi
  if [ -e /etc/nut/local/upsd.conf ]; then
    cp /etc/nut/local/upsd.conf /etc/nut/upsd.conf
  else
    cat <<EOF >>/etc/nut/upsd.conf
LISTEN 0.0.0.0
EOF
  fi
  if [ -e /etc/nut/local/upsd.users ]; then
    cp /etc/nut/local/upsd.users /etc/nut/upsd.users
  else
    cat <<EOF >>/etc/nut/upsd.users
[$API_USER]
        password = $API_PASSWORD
        upsmon $SERVER
EOF
  fi
  if [ -e /etc/nut/local/upsmon.conf ]; then
    cp /etc/nut/local/upsmon.conf /etc/nut/upsmon.conf
  else
    cat <<EOF >>/etc/nut/upsmon.conf
MONITOR $NAME@localhost 1 $API_USER $API_PASSWORD $SERVER
RUN_AS_USER $USER
EOF
  fi
  touch /etc/nut/.setup
fi

chgrp $GROUP /etc/nut/*
chmod 640 /etc/nut/*
mkdir -p -m 2750 /dev/shm/nut
chown $USER.$GROUP /dev/shm/nut
[ -e /var/run/nut ] || ln -s /dev/shm/nut /var/run
# Issue #15 - change pid warning message from "No such file" to "Ignoring"
echo 0 > /var/run/nut/upsd.pid && chown $USER.$GROUP /var/run/nut/upsd.pid
echo 0 > /var/run/upsmon.pid

/usr/sbin/upsdrvctl -u root start
/usr/sbin/upsd -u $USER
exec /usr/sbin/upsmon -D

Dockerfile

FROM alpine:latest
MAINTAINER you
ARG BUILD_DATE \
    VCS_REF

LABEL org.label-schema.build-date=$BUILD_DATE \
    org.label-schema.license=GPL-2.0 \
    org.label-schema.name=nut-upsd \
    org.label-schema.vcs-ref=$VCS_REF \
    org.label-schema.vcs-url=https://github.com/instantlinux/docker-tools

#ARG NUT_VERSION=2.8.0-r4

ENV API_USER=upsmon \
    API_PASSWORD= \
    DESCRIPTION=UPS \
    DRIVER=usbhid-ups \
    GROUP=nut \
    NAME=ups \
    POLLINTERVAL= \
    PORT=auto \
    SDORDER= \
    SECRET=nut-upsd-password \
    SERIAL= \
    SERVER=master \
    USER=nut \
    VENDORID=

HEALTHCHECK CMD upsc $$NAME@localhost:3493 2>&1|grep -q stale && exit 1 || true

RUN apk add nut libcrypto1.1 libssl1.1 libusb musl net-snmp-libs
RUN apk -U upgrade 

EXPOSE 3493
COPY entrypoint.sh /usr/local/bin/
ENTRYPOINT /usr/local/bin/entrypoint.sh

compose.yaml

services:

##########################################
#                 NUT                    #
##########################################
  nut:
    image: tmjpugh/nut_230825:latest
    container_name: nut
    hostname: nut01
    restart: "unless-stopped"
    volumes:
      - "/myserver/nut/entrypoint.sh:/usr/local/bin/entrypoint.sh:ro"
      - "/etc/localtime:/etc/localtime:ro"
    environment:
      API_USER: "upsmon"
      API_PASSWORD: "cyber"
      DESCRIPTION: "ups"
      DRIVER: "usbhid-ups"
      GROUP: "nut"
      NAME: "Cyberpower"
      PORT: "auto"
      SDORDER: "-1"
      SECRET: "nut-upsd-password"
      SERVER: "primary"
      USER: "nut"
      VENDORID: "0764"
      PGID: 1002 #i created this group on my system
      PUID: 1002 #i created this user on my system
    networks:
      nut:
    devices:
      - /dev/bus/usb:/dev/bus/usb
    deploy:
      resources:
        limits:
          cpus: '.2'
          memory: 1000M
        reservations:
          cpus: '.1'
          memory: 256M
          
##########################################
#              NETWORKS                  #
##########################################      
networks:
  nutt:
    driver: bridge

I dont expose port because I connect within docker network to HA but if needed 3493 should be exposed.

9

Came across this whilst looking for something else, but thought I’d share my setup.

I have 3 Cyberpower UPSs in my network, 1 is connected to a windows machine running the Cyberpower Power Panel software, and the other 2 are connected to Raspberry Pi 3s via USB, each running NUT.

HA runs on the Windows machine, with the NUT addon to get sensor data from the 2 RPis, and SNMP sensors to get data from Power Panel.

So the OP could simply attach RPis running NUT to effectively network the UPSs.

Its a much cheaper way than buying networked UPS hardware.

Hope that helps.