[Solved] Cloudron down after Ubuntu/Docker update (Docker service fails to start)

[osobo]

Hello all,

Just wanted to share what happened to me and the way to fix it in case you experience the same issue. @cloudron_admin please delete the post if not relevant.

After applying recent Ubuntu security updates and rebooting, my Cloudron instance became unreachable.

cloudron-support --troubleshoot

showed Docker was down, and

systemctl start docker

failed while running dockerd manually worked fine.

I've worked it out with the support of ChatGPT. Root cause : An outdated Cloudron systemd override file:

/etc/systemd/system/docker.service.d/cloudron.conf

This override replaces Docker’s default ExecStart and uses older flags (like cgroupfs) that are no longer compatible with newer Docker versions. As a result, Docker fails to start via systemd.

Fix : Remove the override and reload systemd:

sudo rm /etc/systemd/system/docker.service.d/cloudron.conf
sudo systemctl daemon-reexec
sudo systemctl daemon-reload
sudo systemctl restart docker

Verify Docker is running:

systemctl status docker

Then restart Cloudron:

sudo systemctl restart box

Result

Docker starts normally again, containers come back up, and the Cloudron dashboard becomes accessible.

Notes
Modern Docker versions already handle most of the options that the override was forcing.
Removing the override lets Docker run with its correct defaults (including proper cgroup handling).
This issue may appear after unattended upgrades or manual system updates.

Hope this helps anyone running into the same issue

[jdbaker]

Thanks!
This happened with my Hostinger VPS installation a couple days ago, so I rolled back to a snapshot. I had another Ubuntu security update pending reboot this morning, so after reading your post, I figured I'd give it a whirl, but this time the Cloudron came back up on its own.

[Valexico]

Same issue here, same solution except I had to reboot the server (sudo systemctl restart docker not working)
I had a bit of panic for a sunday evening !
Waiting to hear from the Cloudron team to confirm which are the recommended fix !

For the records, it happened after my scheduled weekly cloudron update + reboot. So I guess it could happen to any user

[osobo]

@jdbaker Glad to hear everything is back on track. Indeed my VPS is hosted by Hostinger too.

[james]

Hello @osobo
Do I understand it correct that you did a manual system update and this was not an unattended security update?

[Valexico]

Hi @james
In my case i Didn't do any manual update : I let cloudron manage updates I just schedule a reboot once a week

[osobo]

Hello @james. I think so. Following Cloudron’s notification to reboot the server I’ve simply clicked on the « reboot » link within the dashboard and lost the access from there.

[james]

Hello @valexico and @osobo
Thanks for that.
Can you two please ssh into your Cloudron server and post the output of following command?

docker --version

[osobo]

@james Docker version 29.1.5, build 0e6fee6

[Valexico]

Same here
Docker version 29.1.5, build 0e6fee6

[james]

Hello @osobo and @valexico
Thanks.
Do you still have the /etc/systemd/system/docker.service.d/cloudron.conf file?
Because on two of my Cloudron systems it is also Docker version 29.1.5, build 0e6fee6 and there was no issue with docker.

[Valexico]

No I deleted the file to fix the issue as described by @osobo

[Valexico]

my claude chat if it helps
https://claude.ai/share/acd62c73-b4e5-4e49-9e84-7e06d4229815

[osobo]

Same here. The file has been deleted with the fix.

[jdbaker]

For what its worth, the issue occurred for me last Thursday night after using the reboot option in the dashboard, as instructed by the typical Cloudron notification. The VPS came back up, but the Cloudron dashboard was unreachable. In the VPS terminal, the troubleshooter showed docker was not starting. Two additional VPS reboots did not resolve it, so I simply rolled back, and then everything appeared to be working.

Later that night, Cloudron updated from 9.1.5 to 9.1.6.

When another reboot notification appeared yesterday, I took another snapshot of the VPS before rebooting, and everything came back up just fine.

However, I did NOT delete my cloudron.conf file, and it still exists.

My docker version is 29.1.5, build 0e6fee6.

[Valexico]

@james
So, should we do something to restore our cloudron to nominal configuration ? Recreate this cloudron.conf file somehow ?

[james]

Hello @valexico

@Valexico said:

So, should we do something to restore our cloudron to nominal configuration ? Recreate this cloudron.conf file somehow ?

That would be best I think.
If this would reproduce the previous issue, that be interesting to get an understanding what is causing this.

The default content from /etc/systemd/system/docker.service.d/cloudron.conf is:

[Service]
ExecStart=
ExecStart=/usr/bin/dockerd -H fd:// --log-driver=journald --exec-opt native.cgroupdriver=cgroupfs --storage-driver=overlay2 --experimental --ip6tables --userland-proxy=false