Multiple "Mirrors" from a single device

[DarrenO 0]

@sdetweil
Thanks. That has sort of help me out and pointed me in the right direction.

Is it just as simple as copying the existing:

• mm.sh file -->> mm2.sh
• pm2_MagicMirror_new.json -->> pm2_MagicMirror2_new.json

and editing the pm2_MagicMirror2_new.json file to reference:

• the app name of MagicMirror2
• the script of mm2.sh (is there anything to change in this?)
• watching the config2.js file for changes

and then executing pm2 start config2.js --name MagicMirror2

[sdetweil]

@DarrenO-0 the json file my installer script creates is just to give the app name in pm2 a meaningful mame,

if you use
pm2 start mmfoo.sh
the the app name will be mmfoo

the watch function is broken, pm2 now only supports watching a folder for changes, but file updates don’t count as a folder change

can you change the name of the script, yes, anything you like

once you update the json file , just
pm2 start xxxx.json
and it will take the info from there

[BKeyport]

As I rewrote that section in the docs - I’ll give a (semi) live example (My second mirror is down due to lack of space, I’m working on a solution) - this is what works for me at the moment. I do realize I’m doubling up on some things.

In my Pi’s config, I also reverted to X11 desktop rather than wayland for reasons unrelated. I have not tested this on Wayland.

---

mm.sh

cd /home/bkey1970/MagicMirror
DISPLAY=:0 npm start

mm2.sh

cd /home/bkey1970/MagicMirror
export MM_CONFIG_FILE=config/config2.js
export MM_PORT=8081
DISPLAY=:0 npm start

the start of config.js:

var config = {
	address: "0.0.0.0",
	port: 8080,
	ipWhitelist: [],
	language: "en",
	timeFormat: 12,
	units: "imperial",
	//logLevel: ["INFO", "LOG", "WARN", "ERROR", "DEBUG"],
	modules: [

The start of config2.js:

var config = {
	address: "0.0.0.0",
	port: 8081, 
	electronOptions: { 
		x: 1920
	},
	ipWhitelist: [], 
	language: "en",
	timeFormat: 12,
	units: "imperial",
	//logLevel: ["INFO", "LOG", "WARN", "ERROR", "DEBUG"],
	customCss: "css/custom2.css",
	modules: [

PM2 commands to tie it all together:

pm2 start /home/<username>/<path>/mm.sh --name "MM Main"
pm2 start /home/<username>/<path>/mm2.sh --name "MM 2nd"
pm2 save

[DarrenO 0]

@sdetweil
OK, so no requirement to update the “name” and “watch” values in the .json file, but it doesn’t hurt to do so from an fyi perspective, right???

When I run “pm2 start config2.js”, I can then access the second instance via browser on another device with no issue, but it also takes over the display directly connected to the MM host device.
I can view both MM instances via a browser on different devices (great, as that was one of my original questions), but what if i want to view instance #1 via the display directly connected to the MM host device and instance #2 via a remote browser (or vice versa?

My .sh files to start each MM instance are as follows:

mm.sh
#!/bin/bash
cd ~/MagicMirror
DISPLAY=:0 npm start

mm2.sh
#!/bin/bash
cd ~/MagicMirror
export MM_CONFIG_FILE=config/config2.js
export MM_PORT=8081
npm start
#DISPLAY=:0 npm start

Should I change anything, or what would I need to change, to have the desired outcome described above?

[sdetweil]

@DarrenO-0 npm start run the ui locally and the server for remotes

npm run server starts the server for remote and browsers locally

both cases remotes need the server config to allow outside connection

pm2 start xxxx
where xxxx is the script name ( end in . sh)
OR
the json file where it can get the info

the different config is set thru an environment variable
MM_CONFIG_FILE=

in linux by default env variables are NOT passed to children processes
so you have to use export to
make it passed on

on Windows env variables ARE passed to children processes by default

the port number is used on the remote system to access the appropriate instance

http://mm-server-address:mm-instance-port-number

[DarrenO 0]

@sdetweil
so mm2.sh should/would be:
#!/bin/bash
cd ~/MagicMirror
export MM_CONFIG_FILE=config/config2.js
export MM_PORT=8081
npm run server
#DISPLAY=:0 npm start

while mm.sh remains untouched

When I have this in place:

• instance #1 correctly displays on the screen attached to the MM host… tick
• i can also view instance #1 on a browser (on port 8080)… tick
• can view instance #2 in a browser (on port 8081)… tick

However, when I run mm2.sh while I am SSH’d to the MM, I am not returned to the command line so I can save the pm2 setup.
I just have a prompt saying “Ready to go! Please point your browser to http://ipaddress:8081”

I have to Ctrl-C to cancel out, which then stops the server function and the #2 instance is no longer displayed

[sdetweil]

@DarrenO-0 in Windows when you want an app to run when started from the command prompt you do

detach command…
in linux you add and & as the last character

comand… &

now… STOPPING those detached/background things are tough…
here is one value for pm2… it EXPECTS the script to block

and so you can kill it thru pm2 stop xxxx
BUT
pm2’s JOB is to start and app AND KEEP IT RUNNING…
so you kill it (ctrl-q) or it fails, pm2 will restart it

SO, if you want both to run and want to do it manually, open a second ssh session
until you are happy and then let pm2 run them both

[DarrenO 0]

@sdetweil
OK, so I launched the second instance of MM via mm2.sh with the “npm run server” command.
I then launched a second SSH session to the MM device and issued a “pm2 save” command.
Rebooted the MM device from the second SSH session but the second MM instance didn’t come up.

[DarrenO 0]

@BKeyport
isn’t this to run two MM’s on a single display and have the second MM displayed from the 1920th pixel on the x axis?

I’m wanting to run two MM’s, each instance on isolated displays - one on the display connected to the host in the traditional method and the other on a remote display like another computer, or TV with an in-built browser.

[sdetweil]

@DarrenO-0 and what does the status say?

pm2 status
pm2 logs 1 (or whatever the number is of the line for the second ) --lines=xxxx
where xxxx is the number of the most recent lines in the log for that app

I have lots of apps on my desktop pm2
pm2 status output

| id │ name               │ mode     │ ↺    │ status    │ cpu      │ memory   │
├────┼────────────────────┼──────────┼──────┼───────────┼──────────┼──────────┤
│ 1  │ MagicMirror        │ fork     │ 27   │ stopped   │ 0%       │ 0b       │
│ 7  │ MagicMirror1       │ fork     │ 279… │ stopped   │ 0%       │ 0b       │
│ 0  │ Smart Mirror       │ fork     │ 31   │ stopped   │ 0%       │ 0b       │
│ 3  │ checkarec          │ fork     │ 0    │ online    │ 0%       │ 3.5mb    │
│ 6  │ smart-mirror       │ fork     │ 9    │ online    │ 0%       │ 3.5mb    │
│ 5  │ smart-mirror ttt   │ fork     │ 10   │ stopped   │ 0%       │ 0b       │
│ 4  │ start_alexa_bk     │ fork     │ 0    │ online    │ 0%       │ 3.5mb    │
│ 2  │ startspot          │ fork     │ 328  │ stopped   │ 0%       │ 0b       │
└────┴────────────────────┴──────────┴──────┴───────────┴──────────┴────────

the count under the circle coloum is restarts of that app

[sdetweil]

@DarrenO-0 the server side still provides access regardless of where the local display output is configured

AND if you run server mode, then the config for the local electron isn’t used

[sdetweil]

@DarrenO-0 said in Multiple "Mirrors" from a single device:

or TV with an in-built browser.

just heads up…

we use the latest Javascript language options
most/many/all TV’s do NOT have this browser version available and cannot be upgraded.
so the MM screen will be black (browser crashed)

when I got here in 2017, I could run MM on my Samsung TV browser
not anymore

[DarrenO 0]

@sdetweil
after running the mm2.sh script (runs as “npm run server”) and then I initiate a new SSH to the MM, I still only see a single instance (the original instance from mm.sh) of MM running when I issue a pm2 status command.
Although I can physically see the second MM displayed as desired in my browser, I’m not seeing the second instance via the pm2 status command.

[DarrenO 0]

@sdetweil
Fair enough.
So i’ll have to have a physical computer like a Pi, NUC or MiniPC attached to the TV via HDMI to display the MM via that computer’s browser.
Not a big deal as the TV would be used for viewing TV channels and online streaming content anyway.

[sdetweil]

@DarrenO-0 how did you add it to pm2 . just doing mm2.sh doesn’t

pm2 start xxxx
xxx is the mm2.sh
or the MagicMirror2.json
it should run and you should see with status

then pm2 save so that it will be saved to the bootup start list

[DarrenO 0]

@sdetweil
yeah, I started it via the mm2.sh script that had the “npm run server” command
I’ve now started the second instance as “pm2 start pm2_MagicMirror2_new.json” and “pm2 status” does show both instances running.
Ran “pm2 save”, rebooted the host and both instances are showing as expected - MM#1 on the display connected to the host device, with MM#1 and MM#2 via a browser on a remote device/devices.

[sdetweil]

@DarrenO-0 awesome!!..

now select which post is the one that you think ‘solved’ this
hit the 3 dots (bottom right of the post) and click on Mark this post as the correct answer

[BKeyport]

@DarrenO-0 Ahh, gotcha. Yes, this is for two on the same system, which linux treats as one huge monitor.