Yes, it’s good 
For my part everything became stable again after reinstalling Node-RED following Pierre-Gilles’ tutorial and by adding -u node-red:dialout \.
Given the number of steps I carried out beforehand, I don’t know what finally fixed things.
Thanks everyone
\nHello,\nI’m going to see what @tiboys did to check if that works in my case. Because my setup is very heavy and, besides, this morning I feel like I’ve frozen ZigbbeeMQTT2.
Is that generally the case or is it only for your situation since you use serial-port?
I checked the Node-RED website and the « default » command doesn’t seem to have changed (see: https://
Hello @pierre-gilles
Based on the previous messages, it seems I’m not the only one who lost Node RD after the V3 version change. But @tiboys seems to have solved the problem more easily; I’m waiting to hear back from him.
Actually it was a big makeshift hack without fully understanding it, as usual! 
I copy-paste commands without knowing what they trigger.
Basically, I followed your guide @Psoy but it errored on « device=/dev/ttyUSB0 \ », so sudo chown -R 1000:1000 /var/lib/node-red told me it couldn’t find the path.
So I started over to remove the nodered docker and then I followed the original guide adding the -u node-red:dialout option in the docker run -d command group.
I tried to install version 0.15 of serial port but it messed up and I ended up with 1.0.1 and everything works.
In my tinkering, I created a mynodered container but it’s not « started » so I don’t think it has any impact.
Sorry I can’t give more details ..
I can run any command that could give you more info.
Hello
I was away for two days and when I came back I put everything back to square one. Apparently there is a much simpler method
« This morning I feel like I froze ZigbbeeMQTT2. » I took the opportunity to add a USB hub to ease the RPI. To be safe I removed it and restarted
I haven’t had time yet to check whether Node-RED is still running on my machine.
As for a USB hub, I tried using one but without success.
Otherwise, you just need to add a label to block Node-RED updates with Watchtower.
In my case I prefer to lock a container that works well and only update it in certain cases, carefully controlling what I’m doing!
1 Like
Good evening,
I have, despite myself, just reinstalled Node-RED following the update from version 0.15.0 to 1.0.1, which happened by itself. So I took the opportunity to redo the Node-RED installation procedure. I followed @tiboys’ message.
I therefore followed Pierre-Gilles’ tutorial, adding the line -u node-red:dialout \\. I had already done that but it didn’t work, because I think I hadn’t written that line in the right place,
here is the command that worked:
docker run -d \\
--log-opt max-size=10m \\
--restart=always \\
--privileged \\
-u root \\
--network=host \\
--name node_red \\
-u node-red:dialout \\
-v /var/lib/node-red:/data \\
nodered/node-red
Then I restarted Node-RED. At that moment Node-RED rebooted by itself. So I added VonOx’s line:
sudo chown -R 1000:1000 /var/lib/node-red
and with that Node-RED started again, I can open it. After installing the Serialport palette 1.0.1: problem « connection lost » and when you close the Node-RED page, it doesn’t reappear when you try to reopen it. So for me the only option was to use the Serialport palette 0.15.0. Once this version of the palette is installed everything goes back to normal.
If someone has managed to get the Node-RED palette v1.0.1 running, I’d appreciate them explaining what they did. Because I think I’ll have to redo all these steps at every update of the Serialport palette.
@spenceur, I’m interested in blocking Node-RED updates, but I looked at your message and I didn’t understand it; could you detail the steps to block Node-RED updates? Thanks in advance
I will modify the tutorial Rflink, Node-RED and Gladys.
Have a good evening
Hello,
If I understood correctly, to start your container you do this:
You just need to add this option
--label com.centurylinklabs.watchtower.enable=false
Or
-l com.centurylinklabs.watchtower.enable=false
Which gives something like:
docker run -d \
--log-opt max-size=10m \
--restart=always \
--privileged \
-u root \
--network=host \
--name node_red \
-u node-red:dialout \
-v /var/lib/node-red:/data \
--label com.centurylinklabs.watchtower.enable=false
nodered/node-red
By the way, is it normal that your command has -u twice?
Be careful: by doing this, watchtower will no longer automatically update your container.
You’ll have to monitor the evolution of each locked container independently to perform the update and do it correctly 
Hi @Psoy, my victory was short-lived 
After a few days of running without problems, I lost the node serial port in Node-RED. It disappeared and it’s impossible to reinstall it via the menu (it keeps spinning but always returns to « install »).
I try the procedure via the console with the command for 0.15 then a Node-RED restart, and when I try to connect, the Node
Hello @tiboys
this happened to me last night, I was on version 0.15.0 and Node-RED moved to version 1.0.1, so Node-RED stopped and I was forced to reinstall everything
After doing the steps from @bjm when you select the Serialport palette you only see version 1.0.1, you install it and when you check that your palette is properly installed, it’s version 0.15.0 that appears. I don’t know why, but that’s how it is 
Thanks @spenceur
I don’t know what -u means, so having two of them in that command doesn’t surprise me unfortunately 
I copy-paste and if it works, so much the better.
I’m open to explanations, or to getting info on how to set that up/program it? It’s in Docker, it’s Linux, I can’t find my way around it.
Thanks for your answer, I’ll reprogram it, at the next crash
Hello, I put the link to the command’s documentation in my message, here is an excerpt :
@Einstein8854 was talking about it above :
That’s normal, it’s already more technical, it’s the complete opposite of what @pierre-gilles wants.
If you get your hands dirty without understanding what you’re doing, it’s not simple.
I strongly advise you to lock your Node-RED version.
Otherwise, @VonOx, @pierre-gilles, I wonder if you shouldn’t configure Watchtower in the images so that it only checks Gladys and not the other containers.
Because otherwise you’ll end up with lots of topics that don’t directly concern Gladys, but systems on which Gladys is plugged in or at least plugged into Gladys’s « service » MQTT.
2 Likes
I know it’s Pierre-Gilles’s fight, but it’s very tempting to be able to install software that lets me use RFLink, the Xiaomi Gateway and Tuya over Wi‑Fi. Since I installed Node-RED, I’ve learned Linux commands, JavaScript… but the road is still long and when you don’t have the basics it’s much harder to know where to look for information, that’s why I’m calling on you to help me, and I thank you again. 
That makes sense indeed, we provide support for other projects that
Hello @VonOx,
you’re more familiar with it than I am and despite my searches on the web I haven’t found anything.
Wouldn’t there be a UI for Watchtower, similar to the one we have for Zigbee on port 8080, that would on the one hand allow viewing Watchtower’s configuration and settings and on the other hand let you modify the settings on demand? Because Watchtower is really useful for keeping things up to date. The idea would be for Watchtower to alert about available updates for the containers in our Docker and for the user to be able to trigger the update manually or not?
No, that doesn’t exist, and there’s no point anyway because there is no configuration file — watchtower is a CLI.
If you want something stable, don’t use Docker images with the « latest » tag. As @spenceur says there are tags for each version (we do that for gladys, by the way)
1 Like
I deleted my flows, uninstalled the serial-port palette 1.0.1, reinstalled 0.15 and it’s working again. I hope it won’t update itself on its own…
Wait and See ..