Since the switch to 2025 (link or not, I don’t know), my Zigbee2mqtt integration crashes about every 10 days. It started in January and I’m now on my 4th restart of Gladys to get everything back to normal.
I notice it because my zigbee devices stop responding. When I go to the Zigbee2mqtt configuration, the service is disabled, and it’s impossible to enable it (« an error occurred, check the service configuration »). There is a red cross between the Gladys module and Mqtt and between Mqtt and Zigbee2Mqtt.
I turn off the Beelink, restart it and it’s back up (for about ten days).
I don’t think I’ve seen this issue reported on the forum. Any idea what the bug might be?
As for the logs, I’m having trouble.
When I run your command it tells me that the container does not exist. When I run sudo docker ps, it outputs nothing.
Digging deeper I realize that the files are hidden: doing ls at the root I get nothing, doing « ls -a » I have folders (« .config » for example).
So I tried « sudo docker ps -a » and there I only have one container « gladys ».
I must have missed something
Hello,
Following the system update, I again have containers present when I run sudo docker ps.
But when I try to check the logs, I can only go back as far as yesterday morning.
I’ll wait for the next crash (if it happens again, since maybe the update will have fixed everything).
It crashed again today. Just 15 days after the last crash and 10 days after the last reboot.
I ran sudo docker logs gladys-z2m-zigbee2mqtt: it tells me the container doesn’t exist.
I ran sudo docker ps: nothing. No containers listed.
I rebooted the system: the Zigbee is operational again, but the commands still return no result.
I ran updates again, but no better — I still officially have no containers running, and yet everything works correctly. But I can’t tell what happened.
Yes, a complete reinstallation is possible since I have the Gladys Plus backup. But what I find strange is that it worked for 2 months without problems, and that it has been crashing since 2025.
Then, I’m thinking of two things:
The installation I currently have is already based on a backup: I was on a Raspberry Pi, and I migrated to a Beelink by doing a normal installation, then at startup I immediately restored a backup. Wouldn’t that create like two instances of the installation? Which would mean that when I run sudo docker ps, it returns nothing because on the initial installation I don’t have any docker running; and yet everything works because it’s my second instance doing the job. It may be silly, but well, not knowing how it works, I tell myself that might be an explanation.
About my crashes: it crashed again yesterday. And I notice I’m on version 4.55.2. Whereas on Monday you announced the release of 4.55. So I wonder if my installation crashes with every update. Do you have the dates (even approximate) of the updates made available to devices in 2025?
So, I’d like to try to understand, but if it’s hopeless, I’ll redo an installation
No, restoring a backup does not start a second container, it just restores a database file inside the same container.
Here are all the dates:
I think in your case something is wrong, because if a sudo docker ps shows nothing while Gladys is running, that’s just bizarre. To me there’s an underlying software issue beneath Gladys:
Either an issue with the version of Ubuntu Server you installed. Maybe the version had a bug the day you installed Ubuntu.
Or an issue with the version of Docker you installed
Or a mistake during the installation
If you want us to have a call one day to look at it together, no problem — I’m always available
