Hi all

So i received a message today that my Docker image disk is getting full (70%). I logged in and stopped Docker, changed the value to a bigger size and restarted the service, just to see that it doesn't come up again. I then tried different things, resulting in completely deleting the image and adding the docker containers new through user templates.

Strangely some containers work, some not. Some don't work at all (start up normally, but aren't reachable). And some do work for a couple of minutes, and then don't work again, and then run again. I now also stopped all containers and started them again, just to see that even more don't work now. I also tried to restore the files from a backup, which didn't change anything. In the logs, they are just looking fine, as if they are running perfectly.

I'm not really sure where to look next as it seems to be behaving very randomly in my eyes. I attached the diagnostics from my server.

unraid-diagnostics-20251002-1618.zip

There appears to be a contender constantly restarting; see if you can find which one by looking at their uptimes.

Unfortunately not. All are now running for more than 2h. Also no portconflicts. And i also didn't change any permissions.

Constant log activity like this is not normal:

Oct  2 16:18:06 UNRAID kernel: block nvme0n1: the capability attribute has been deprecated.
Oct  2 16:18:06 UNRAID kernel: docker0: port 48(vethb651948) entered blocking state
Oct  2 16:18:06 UNRAID kernel: docker0: port 48(vethb651948) entered disabled state
Oct  2 16:18:06 UNRAID kernel: vethb651948: entered allmulticast mode
Oct  2 16:18:06 UNRAID kernel: vethb651948: entered promiscuous mode
Oct  2 16:18:07 UNRAID kernel: eth0: renamed from veth31d752d
Oct  2 16:18:07 UNRAID kernel: docker0: port 48(vethb651948) entered blocking state
Oct  2 16:18:07 UNRAID kernel: docker0: port 48(vethb651948) entered forwarding state
Oct  2 16:18:08 UNRAID kernel: docker0: port 48(vethb651948) entered disabled state
Oct  2 16:18:08 UNRAID kernel: veth31d752d: renamed from eth0
Oct  2 16:18:08 UNRAID kernel: docker0: port 48(vethb651948) entered disabled state
Oct  2 16:18:08 UNRAID kernel: vethb651948 (unregistering): left allmulticast mode
Oct  2 16:18:08 UNRAID kernel: vethb651948 (unregistering): left promiscuous mode
Oct  2 16:18:08 UNRAID kernel: docker0: port 48(vethb651948) entered disabled state
Oct  2 16:18:10 UNRAID kernel: docker0: port 35(vethada7201) entered disabled state
Oct  2 16:18:10 UNRAID kernel: veth0738f63: renamed from eth0
Oct  2 16:18:10 UNRAID kernel: docker0: port 35(vethada7201) entered disabled state
Oct  2 16:18:10 UNRAID kernel: vethada7201 (unregistering): left allmulticast mode
Oct  2 16:18:10 UNRAID kernel: vethada7201 (unregistering): left promiscuous mode
Oct  2 16:18:10 UNRAID kernel: docker0: port 35(vethada7201) entered disabled state
Oct  2 16:18:10 UNRAID kernel: docker0: port 35(vethfb3b993) entered blocking state
Oct  2 16:18:10 UNRAID kernel: docker0: port 35(vethfb3b993) entered disabled state
Oct  2 16:18:10 UNRAID kernel: vethfb3b993: entered allmulticast mode
Oct  2 16:18:10 UNRAID kernel: vethfb3b993: entered promiscuous mode
Oct  2 16:18:11 UNRAID kernel: eth0: renamed from veth242aa02
Oct  2 16:18:11 UNRAID kernel: docker0: port 35(vethfb3b993) entered blocking state
Oct  2 16:18:11 UNRAID kernel: docker0: port 35(vethfb3b993) entered forwarding state

Try stopping all the containers, then start just half of them, wait 5 minutes and check the logs, if the activity stopped, start the other half and keep drilling down.

So what i did yesterday, was to stop all containers and reboot the system again. Strangely it had issues with unmounting the volumes (it wrote something like, 'Retry unmounting volumes'). Then it rebooted and then the docker containers were just gone (and it started a paritycheck). Except from the Docker compose containers, these were still here. So i got at it again, recreated the containers from the templates and now it seems to work (they are running fine since yesterday evening).. Very strange behaviour. It honestly doesn't fill me with much confidence that the docker containers are just gone after a reboot (and that the docker image gets corrupted so easily..). Is there an exact procedure, how to restart an Unraidsystem with Docker? Do i have to stop the dockerservice completely before rebooting?

unraid-diagnostics-20251003-1231.zip

Typically you just do a reboot shutdown, and the Docker service should be terminated. If it's not there, there is something preventing the array or pools from unmounting. I also see that you have appdata spread all over the array disks; that's not necessarily a problem, but typically it resides on a pool.

How would i change that?

I got the log from yesterday before the reboot, when it couldn't unmount the disks. I attached it. But i don't see any particular failure.

syslog-192.168.66.94.log

1 hour ago, zotac99 said:

How would i change that?

You can use the mover.

Does the array stop cleanly if you try now?