@Mathes75

not sure I really fixed it but you can try

cd /opt/mm
git pull
./install/install.sh electron

Let me know if there are still problems after reboot or not.

@Mathes75

probably python missing, see https://khassel.gitlab.io/magicmirror/faq/dependencies/#use-the-fat-image

You can change the image in the .env file.

other questions (not clear from description):

Are you using mmos or are you using Raspberry Pi OS or another OS?

If not mmos, how did you the install? With the install script inside the repo?

Can you provide output of docker compose config?

the dbus errors in the mm logs are only noise, see https://forum.magicmirror.builders/topic/20032/docker-watchtower-container-restarting-and-mm-container-errors-after-upgrade/7#

If the problem occurs:

• can you provide the logs of the labwc container? docker logs labwc
• can yo try to restart the containers without reboot with cd /opt/mm/run && docker compose up -d --force-recreate?

@htilburgs said in MMM-MyTasklist:

Is there a tule for the Ports? Which I can use or not use?

you could add a config param for the port so that everyone who collides with 8123 could choose another one without hacking the source code

I found a good documented PR in another project which explains the dbus errors in detail.

@scottwalsh said in Docker Watchtower container restarting and mm container errors after upgrade:

Did you have any thoughts on these errors that appear in the mm container? Seems to be running fine.

they are only noise, you can ignore them.

I thought I sorted out the dbus errors but they are back again, will update my setup so they disappear in the future.

there is a line

WATCHTOWER_IMAGE="containrrr/watchtower:latest"

in the file /opt/mm/run/.env.

Can you change this to

WATCHTOWER_IMAGE="ghcr.io/nicholas-fedor/watchtower:latest"

and restart with

cd /opt/mm/run
docker compose up -d

Background is that the original watchtower is unmaintained and I suspect the original image is to old now. The new image is from a maintained fork, I use this already in other setups.

@sdetweil

they changed several things in Nov./Dec. 2025

• they now use cloud-init in Raspberry Pi OS
• the new Pi Imager v2 does not allow customizations (user, ssh, wifi, …) when using an image file from disk

Links:

• Cloud-init on Raspberry Pi OS
• How to add your own images to Imager

@HeikoGr

Thanks! Your configuration inspired me to write down and update my own.

I don’t mean to hijack your thread, but since development environments are always a matter of personal preference, I thought I’d share this as an alternative.

this is now fixed with https://github.com/MagicMirrorOrg/MagicMirror/pull/3946 and will be in the next release (Jan 1st)

@jca2112

I opened https://github.com/MagicMirrorOrg/MagicMirror/issues/3944

I did not look into details, the changes made in this file caused this.

@jca2112

this is a bug introduced with last mm version.

Let me know if you are willing to open an issue here, otherwise I will do …

this is unrelated to mm, the default url used in the module is https://www.youtube.com/embed/yf5cEJULZXk?si=Dx852YRN5q6NHj0K and I see the same error 153 in my browser.

Clicking on “Watch video on YouTube” will move you to https://www.youtube.com/watch?v=yf5cEJULZXk with the content “This live stream recording is not available.”

So this NASA stream is at least currently not available …

Maybe of interest for @KristjanESPERANTO

@schlomm

you linked the lite versions, for getting the version with desktop lite must be replaced by full in the urls …

@Nneuland

if See below for more Configuration Options is really in your config remove it …

@sdetweil said in HDMI turns on without movement:

@karsten13 have you had a chance to test universal pir on Trixie yet?

I’m on it …

@Hobbes-0 said in HDMI turns on without movement:

I tested the offCommand directly in the Terminal, the monitor goes off, and again after a minute or so goes back on. So this does not come from the MM.

@sdetweil so this problem is not related to universal pir or mm

I tested universal pir on mmos, which is on trixie but uses the os lite version, mm and labwc/wayland are started as containers.

I cannot reproduce the unexpected activation there.

(I found another permission problem which I will fix later) fixed

Other difference between trixie and bookworm is a breaking version change of gpiomon so you have to use different commands:

bookworm: gpiomon -r -b gpiochip0 23
trixie: gpiomon -e rising -c 0 23

@sdetweil said in Help to install bluelinky:

bluelinky wans node up to 20.9 or 22, and NOT ANY 21

message comes from mm, he is on an outdated version 2.29.0 …

@chrisfr1976

#server_functions is an alias for js/server_functions.js which is defined in package.json:

        "imports": {
                "#module_functions": {
                        "default": "./js/module_functions.js"
                },
                "#server_functions": {
                        "default": "./js/server_functions.js"
                }
        },

Because axios is the only diff your package.json should contain above lines.

I have no idea anymore (maybe corrupt sd card).

@chrisfr1976

what did you change in package.json?

git diff package.json