bug in containerd impacting dozzle and potentially other containers

This thread is info only but might impact omv users running compose plugin.

I have an issue with containerd spamming my syslog with these errors - literally hundreds of logs per minute.

On my systems, I tracked the issue to dozzle and have reported it to the dozzle dev. He can reproduce but says it is impacting other containers and he thinks this is a bug with containerd. No further update at this time.

Hopefully will get resolved with an update to containerd. Issue being tracked here in dozzle github here... https://github.com/amir20/dozzle/issues/4267

containerd[227310]: time="2025-12-02T07:41:39.863011207+11:00" level=error msg="unable to parse \"max 0\" as a uint from Cgroup file \"/sys/fs/cgroup/system.slice/docker-548fe802b6a67a22f4c6f31f3411dac9a0274969dcb4ded5c69aff2dbdaaa126.scope/hugetlb.1GB.events\""

containerd[227310]: time="2025-12-02T07:41:40.864289809+11:00" level=error msg="unable to parse \"max 0\" as a uint from Cgroup file \"/sys/fs/cgroup/system.slice/docker-548fe802b6a67a22f4c6f31f3411dac9a0274969dcb4ded5c69aff2dbdaaa126.scope/hugetlb.2MB.events\""

This issue is not specific to omv8. I have it on omv7 too.

It is being worked on my containerd devs so should be resolved soon

Quote from topo01

How do you track down which container is causing this?

It is tricky. I used trial and error method but there is probably a better way. Down all containers and start them one at a time...

I think the following from the error might be related to the containerID but haven't confirmed docker-0680deeede67693cbd042aa2d785c29076c8a8b85526f374c09953b77b0c7e88.scope

Note that you need to use the id from your error log not the one i posted above as an example

Fix might be out soon with an update to containerd. Here is the latest info I have on the issue.

the containerd version that is being used is v2.2.0 which has the problem. The fix has been merged into release/2.2 branch and will be available in the next release 2.2.1

If you want a work-a-round, you can install the previous version of containerd.

first get a list of versions available via apt

sudo apt list --installed -a containerd.io

root@minipc:~# sudo apt list --installed -a containerd.io
containerd.io/trixie,now 2.2.0-2~debian.13~trixie amd64 [installed,automatic]
containerd.io/trixie 2.1.5-1~debian.13~trixie amd64
containerd.io/trixie 1.7.29-1~debian.13~trixie amd64
containerd.io/trixie 1.7.28-2~debian.13~trixie amd64
containerd.io/trixie 1.7.28-1~debian.13~trixie amd64
containerd.io/trixie 1.7.28-0~debian.13~trixie amd64
containerd.io/trixie 1.7.27-1 amd64

then install the previous version

sudo apt install containerd.io=2.1.5-1~debian.13~trixie

I have tested the above in a omv8 VM and seems to work...

But you need hold back from it being reinstalled as the latest version will be available to install...

work-a-round tested on my prod omv8 environment and all is good so far - no more log spamming by dozzle container.

No worries and I agree the easiest solution is to not run the container.

Me too. Looks like issue is resolved.