49 thoughts on “Raspberry Pi UPS monitor (with Nginx web monitoring)”

1. Logan, this is awesome. And though I know you’re not tech support, I can’t get it to work completely. The test emails work, sudo upsc works and shows me what I need, and I get the wall messages. However, the system won’t email me the alerts.

   Thoughts?

   • What action are you performing to get the wall alerts? Are you unplugging the UPS from the wall power?

     Did you add the EXEC to /etc/nut/upsmon.conf and make sure your script is executable?

     What if you just run the script itself from the command line? You should get a blank template email to prove that the script and your SMTP server work.

2. Hi,

   Great guide, one of the best I have been able to find.

   One issue thought that I can’t seem to get around, is that emails are sent correctly, and are being received, however I cannot get a meaningful from name. All the emails come up as from “derived from envelope by postmaster@***.

   I also note that there is no name in the To field.

   Any ideas?

   Thanks!

   • Are you relaying them through Gmail? Mine come from from my extra account name.

3. Could you explain a little more on exactly how to edit the NGINX config file please? I’m having issues accessing the UPS via a webpage.

   • Sorry, I was assuming a familiarity with Nginx. The configuration files for Nginx are generally located in /etc/nginx/sites-available and you’ll need to use some sort of text editor (e.g., emacs, vi, nano, etc…) to edit the files.

     This is a rough outline, but a basic config with NUT would look something like this.
     server {
listen 80;
server_name domain.com;
root /var/www/html;
index index.php index.html;
location / {
try_files $uri $uri/ =404;
}
location /nut {
alias /usr/share/nut/www/;
try_files $uri $uri/ /index.html;
}
location /cgi-bin/ {
gzip off;
root /usr/lib;
include fastcgi_params;
fastcgi_pass unix:/var/run/fcgiwrap.socket;
fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
}
location ~* \.php$ {
try_files $uri =404;
fastcgi_split_path_info ^(.+\.php)(/.+)$;
fastcgi_pass unix:/var/run/php5-fpm.sock;
fastcgi_param SCRIPT_FILENAME $request_filename;
fastcgi_index index.php;
include fastcgi_params;
}
}


4. Hi Logan,
   Thanks for posting this guide. I just purchased a CyberPower CP1500PFCLCD UPS for my home ESXi server and originally wanted to use use the CyberPower PowerPanel Business Edition for Virtual Machines, but my ESXi server does not support DirectPath IO so I searched for another solution. I came across your guide and was able to set up the NUT suite on an old Raspberry PI B+ and get it working, however is seems that after a few hours of running the pi starts reporting “Broadcast message from nut@rp1 (somewhere) (Sat Nov 18 01:38:54 2017): UPS cyberpower1@localhost is unavailable” and upsc reports “Error: Data stale”. Rebooting the pi solves it for a bit but then it starts failing again. I am thinking that setting this up on a new pi3 might be the answer but wonder if you have any thoughts on this and appreciate any input you may have.

   Thanks,
   Dale

   • Dale, I don’t ever think I received that error, but did you try the tweaks section of my post? It helped to clear up a lot of the issues I was having. Also, did you try a new USB cable?

5. Hi Logan,
   Thanks for the reply. I noticed that sometimes when I get this, the nut-driver is reporting correctly and restarting the nut-server works, yet other times I need to restart both. I will try the tweaks section and a new cable too.

6. HI Logan,
   Your tweaks did the trick and this is now working great for me including the web page and alerting. Thanks again for posting this guide; not only do I have alerting now for my new UPS but I learned a lot in the process!

   • Glad it worked, but I can’t take credit. I searched high and low until I found the fix in this article.

7. Hello Logan, I recently set this up and it is working perfect, but I ran into an issue. When rebooting the pi the UPS reboots as well… This is not optimal as the UPS is running a server and my entire network..

   I have the exact same UPS model as you and was wondering if you had a fix for this.

   Thanks,
   –Cody.

   • That’s really weird. Mine does not do that (I’m not running on a Raspberry Pi anymore, but when I was, it didn’t do that either).

     Are you running NUT from the packages installed via apt?
     What distro are you running?
     What command are you running with the reboot? Have you tried running just a shutdown instead of a reboot?

     This person is having the same issue, but there’s no replies here yet.

     I can send you a copy of my configs, but they should be identical to what I have on the site.

     In the meantime, I’d try asking on the NUT mailing list.

     • I am using the NUT packages via apt on Raspbian GNU/Linux 9 (stretch). To preform the reboot I have tried shutdown -r 0, reboot and poweroff with a replug afterwards. All result in the same behavior.

       I would ask for your configs, but like you say they should be the same. I followed your site to the point.

       Thanks for the quick reply by the way!
       -Cody.

       • Is the Raspberry Pi power plugged into the UPS (battery side)?

         I’d still suggest asking on ServerFault or the NUT mailing list.

8. Logan,
   Thanks for documenting this setup. This is excellent. I just followed your steps to configure Raspberry as a UPS monitor for my Ubiquiti network.

   • Glad to help!

9. This is a great guide, I have it all working on my pi/ups combo. But one thing I cannot figure out is how to get NUT to start automatically when the pi restarts. I read all kinds of conflicting tips on this. Any advice?

   • NUT runs a systemd service, so you should be able to enable everything using standard systemctl commands.
     sudo systemctl enable nut-driver
sudo systemctl enable nut-server
sudo systemctl enable nut-monitor

10. Awesome guide! Was up in running in minutes. Thanks!

    • Glad to help!

11. Great Instructions!!!!! I do not have Nginx installed, I have gnome, how can I go about setting up the rest of this with it?

    Thank You in advance!

    • Nginx is a web server (needed to serve web pages). Gnome is a desktop environment (gives you your toolbars, start menu, etc…). They are not the same thing.

      If want you web monitoring, you’ll need a web server of some sort (Nginx, Apache, etc…).

      • Thank you for your reply, I have this setup in my Pi with Octopi and Octoprint, I will have to check up on the Nginx server I had seen there might be a conflict with the new ver of Octoprint and Nginx…..

12. Thanks for the article, helped me set it up on a rpi that was already running pi-hole.

    To use lighttpd that’s installed along with pi-hole instead of ngnix; edit /etc/lighttpd/external.conf with:

    server.modules += ( “mod_cgi”, “mod_alias” )
    $HTTP[“url”] =~ “/cgi-bin/” {
    dir-listing = “enable”
    alias.url += ( “/cgi-bin/” => “/usr/lib/cgi-bin/” )
    cgi.assign = ( “” => “” )
    }
    alias.url += ( “/nut” => “/usr/share/nut/www” )

    And restart lighttpd:

    sudo systemctl restart lighttpd

    • I’ve never used lighttpd before, thanks for sharing!

    • Tried hundred times without success… lighttpd will not restart

      • Do the logs show anything?
        sudo journalctl -u lighttpd

    • This config doesn’t work. The error in journalctl is:

      Feb 13 09:29:52 raspberrypi lighttpd[1205]: 2019-02-13 09:29:52: (configfile.c.1094) source: cat external.conf 2>/dev/null line: 1 pos: 18 invalid character in variable name
      Feb 13 09:29:52 raspberrypi lighttpd[1205]: 2019-02-13 09:29:52: (configfile.c.1151) configfile parser failed at: (
      Feb 13 09:29:52 raspberrypi lighttpd[1205]: 2019-02-13 09:29:52: (configfile.c.1154) source: /etc/lighttpd/lighttpd.conf line: 77 pos: 1 parser failed somehow near here: (EOL)
      Feb 13 09:29:52 raspberrypi systemd[1]: lighttpd.service: Control process exited, code=exited status=255

      Line 1 position 18 is a space, so I tried removing that and no go, it says line 17 is a syntax error.

      • Try replacing the fancy citation marks (don’t know the proper English words) with the ones you get when typing them on your keyboard. I did that and it works like magic: http://pi.hole/nut
        Don’t forget to restart the web server. sudo systemctl restart lighttpd

      • I forgot something: When I installed nut-cgi, the cgi-bin directory was installed in /var/www. I had to move it into /var/www/html.

13. great info thx

14. Hi Logan,

    Thank you for the tutorial, I have gotten everything to function properly except the Nginx server. I’m new to Nginx and not sure if I have set it up properly. I am able to start the server and get the Nginx welcome page when I visit the ip address, but I cant get to the ip address/nut I get a 404.

    Here is my .conf for the server does any stick out as improperly written?

    user root;
    worker_processes auto;
    pid /run/nginx.pid;
    include /etc/nginx/modules-enabled/*.conf;

    events {
    worker_connections 1024;
    }

    http {
    tcp_nodelay on;
    keepalive_timeout 65;
    types_hash_max_size 2048;

    server_names_hash_bucket_size 128 ;

    include /etc/nginx/mime.types;
    default_type application/octet-stream;

    ssl_protocols TLSv1 TLSv1.1 TLSv1.2; # Dropping SSLv3, ref: POODLE
    ssl_prefer_server_ciphers on;

    ##
    # Logging Settings
    ##

    access_log /var/log/nginx/access.log;
    error_log /var/log/nginx/error.log;

    ##
    # Gzip Settings
    ##

    gzip on;
    gzip_disable “msie6”;

    include /etc/nginx/conf.d/*.conf;
    include /etc/nginx/sites-enabled/*;

    server {
    listen 80;
    server_name localhost;
    root /var/www/html;
    index index.php eli.html;
    location / {
    try_files $uri $uri/ =404;
    }

    location /nut {
    alias /usr/share/nut/www/;
    try_files $uri $uri/ /index.html;

    }

    location /cgi-bin/ {
    gzip off;
    root /usr/lib;
    include fastcgi_params;
    fastcgi_pass unix:/var/run/fcgiwrap.socket;
    fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
    }

    location ~* \.php$ {
    try_files $uri =404;
    fastcgi_split_path_info ^(.+\.php)(/.+)$;
    fastcgi_pass unix:/var/run/php5-fpm.sock;
    fastcgi_param SCRIPT_FILENAME $request_filename;
    fastcgi_index index.php;
    include fastcgi_params;
    }
    }
    }

    • First thing, the nginx.conf and your website configs should be two separate files, not mixed into one file like here.

      If you run sudo nginx -t, what happens?

      Second, does /var/log/nginx/error.log show anything? What about /var/log/nginx/access.log?

      Also, I would change server_name localhost; to server_name XX.XX.XX.XX; (where XX.XX.XX.XX is the IP of your RPi). Unless you’re visiting the webpage directly from the RPi, localhost won’t work anywhere on your network.

      One more thing, are you using PHP5 or PHP7? This line is different for PHP7, Google it for the syntax.
      fastcgi_pass unix:/var/run/php5-fpm.sock;

      • Thanks for the quick reply and all the help Logan! its working perfectly now.

        -Eli

15. Hello dear friend, creator of the post,

    Could you indicate what type of raspberrry pi you are using to monitor a UPS with the USB HID communication protocol?

    I have a raspberry pi 3b + and previously I had a UPS salicru sps 900 with usb 2.0 and NUT worked perfectly with my pi 3b +. Now I have purchased a UPS Salicru SOHO + 1600 but the USB is not USB 2.0 normal, the specifications indicate that it is a USB HID. My raspberry pi 3b + is not able to detect HID devices since they are connected and disconnected continuously and changing device. Since I can not find a solution, I would like to know what exact model of raspberry pi you use to monitor your UPS with HID usb. I am waiting for your news or any possible solution or HUB converter. I would be very grateful for your help. a cordial greeting

    • At the time, I was using a RPi3 B+, however, I’ve since switched an x86 virtual machine.

      If the UPS is showing up as a HID device, that sounds like a keyboard/mouse. I don’t believe a new RPi would fix your issue, since it seems to be something with the USB ports on the UPS itself. This post makes it sound like the USB ports are only for charging devices, not for sending data…
      https://cofme.com/salicru-renews-sps-soho-series/

      In addition, I don’t even see a supported driver on this page…
      https://networkupstools.org/stable-hcl.html

16. Thank you very much for the post, in the end I managed to get almost everything to work.
    In case some person is interested my ups were disconnected every 20 seconds through the usb. Instead of the usb driver blazer I put the usbhid-ups. As I read if NUT communicates with my UPS within the first 20 seconds the raspberry disconnects the USB device and this is due to the port of my UPS. It is a 1.0 HID and has this problem. Before I had another UPS with USB 2.0 port and never disconnected from my raspberry even if I had no installed nut.

    Would you be so kind as to be able to answer 2 questions please?

    1) how can I access from outside the network to the ip of my raspberry to be able to view the web created with nginx? I have a host created but I can not access the information of the ups. I try it in the following way:

    10.180.12.47/nut “space” my host name

    but I get a message “not found ngninx”

    2) Is there any way to create an alert through nut or through nginx to be able to receive an alert when the battery is charged 100% or when it reaches 25%?

    Be kind by your work and by answering all the questions. Have a good night

    • Glad you got it working, thanks for sharing!

      1) Did you open a port in your firewall (either 80 or preferably 443)? You’ll also need to forward requests on that port to the IP of your Nginx server.

      2) I’m not sure if NUT supports those types of alerts out of the box, but you could easily write a script to check for that data.

17. Thanks for a great tutorial. I got my RPI 3 B+ (Stretch) working with NUT. My only issue is that if I reboot the Pi, which I admit isn’t very often, NUT doesn’t auto-start because the nut-driver complains it can’t find the UPS, even though it is listed in the USB ports.

    Typically, unplugging and plugging back in the UPS to a different USB port will allow the driver to start and everything works again.

    Same configuration as the post, same driver, just a Tripp-Lite instead of a CyberPower.

    Anyone have any tips as to what might be happening?

    • That’s weird. When you restart the Pi, can you run lsusb and see the UPS?
      Do the logs say anything about why it’s not connecting? Check /var/log/syslog as well as dmesg.

      • Thanks! I’ll check that out in the next week or so when I have time.

        lsusb detects it no problem every time. I’m almost wondering if I should try introducing a startup delay in the startup scripts.

18. Hi Logan,

    Thank you for the amazing guide!

    This is my first Pi project so please excuse my ignorance.

    I have gotten everything to work until the email step.

    “Next, we’re going to edit the configuration file at /etc/msmtprc to include our information. Substitute your new Gmail username and password below.”

    I install msmtp with no issues via terminal using the install and update command you give.

    However, when I go to the /etc/ directory, there is no msmtprc file for me to find the configuration file.

    Where am I going wrong?

    Thank you

    • I’m 97% positive that you need to create the file, I seem to remember running into the same issue. Use a text editor like nano, vim, or emacs.

      • Hi Logan,

        I have done so and mimicked the systemconf example in the help docs of msmtp, but now get a “msmtp: /etc/msmtprc: input error” whenever I try and use your test printf command above (or any other for that matter).

        Any thoughts?

        • It sounds like a syntax error in your file. Did you check your /etc/msmtprc file against mine (below)? I’m assuming you’re using Gmail as your SMTP relay?
          # Accounts will inherit settings from this section
defaults
auth on
tls on
tls_certcheck on
tls_trust_file /etc/ssl/certs/ca-certificates.crt

          account default
host smtp.gmail.com
protocol smtp
port 587
from your_email@gmail.com
user your_email@gmail.com
password your_app_password

          • Hi Logan,

            Yes I have done so.

            Same error unfortunately.

            Could I email you directly with my config file and you can have a look?

            Thank you

            • Yep that’s fine (my email is here). You can omit your username/password and I’ll use mine.

              What are the permissions on the file? Mine are 644.

              I think this is the source that’s generating the error, but I don’t know C. It seems like this code parses your config file, so I’m still thinking it’s something syntax-related…

              • Hi Logan,

                Have emailed you.

                Thank you

This site uses Akismet to reduce spam. Learn how your comment data is processed.