Updating existing containers?

[mcdent]

Hi
I have a series of rootless containers running from systemd files. I'd like the ability to have some of them periodically update.
As I understand from https://docs.podman.io/en/latest/markdown/podman-auto-update.1.html I can add
--label io.containers.autoupdate=registry to the ExecStart line of the unit file?

Then I need to have sudo systemctl enable podman-auto-update.timer running to periodically update them via podman-auto-update.service

My main question, assuming the above is correct; can I retrospectively add the --label io.containers.autoupdate=registry to each of the unit files I want updating? Or does this need setting on the command line when I first create the container?

Thanks!

[vrothberg]

Thanks for reaching out, @mcdent!

Adding the generated unit files is dangerous and not something we can support. I assume you used podman generate systemd. Have you heard of Quadlet? That's a new way of running Podman containers under systemd and much easier to use.

[mcdent]

Thanks for your reply. I'm not sure I follow. Are you saying using podman generate systemd to create a systemd unit file is dangerous?
Or do you mean editing these startup files, once created is dangerous?
I kind of thought this was the correct way to run my rootless containers?

And no, I have not looked at Quadlet but I shall do so.
Thanks again!

[vrothberg]

Apologies for being vague.

What I meant is that editing the output of podman generate systemd is generally not something we can support. Editing the command-line of podman as you suggested poses no threat though.

I kind of thought this was the correct way to run my rootless containers?

It is absolutely correct to run your containers with podman generate systemd. But as you might have experienced yourself, the user experience of podman generate systemd is not ideal. Changing something requires to recreate the container, and then do everything again. Quadlet solves this problem in a very elegant way, such that adding the auto-update label is easy and straight-forward.

Feel free to reach out if you have questions on Quadlet!

[mcdent]

Thank you that now makes complete sense 👍 I shall go and check Quadlet out, rather than spending time figuring out the creation of all my containers again :)

[vrothberg]

Excellent! Thanks again for reaching out and happy hacking 👍

[mcdent]

Oooh, one of my next projects was to look at converting some of my containers over to k8s. It looks like starting with Quadlet might also help with this process too, if I am reading correctly :)

[vrothberg]

Yes, absolute! Quadlet also supports running K8s workloads (see https://docs.podman.io/en/latest/markdown/podman-systemd.unit.5.html#kube-units-kube). That's also more elegant the pre-existing podman-kube@ systemd template.

And the Quadlet files are easy to share, just like Compose or K8s YAML.

[rhatdan]

#18891

[mcdent]

Wait, I was just searching for a way to specify the autoupdate feature in a Quadlet file :)
For the time do I need to specify it as part of the ExecStart=podman run --label io.containers.autoupdate=registry ?

Does this mean in the next release it will be a native part of the quadlet unit file ?

Forgive my question if it is somewhat daft, I am still learning!

[vrothberg]

At the present you can set it as a Label in Quadlet but #18891 will make it more convenient.

Such conversations are a great source of inspiration.

[mcdent]

Thanks @vrothberg . I ran in to an issue where I had not specified the correct url/path to the image, so now my Quadlet file looks like:-

[Unit]
Description=Uptime Kuma
After=local-fs.target

[Container]
Image=registry.hub.docker.com/library/louislam/uptime-kuma:1
Volume=/home/mike/docker/uptime-kuma/run/data:/app/data
PublishPort=3001:3001
ContainerName=uptime-kuma
Label="io.containers.autoupdate=registry"
Exec=podman run

[Install]
# Start by default on boot
WantedBy=multi-user.target default.target

Hopefully that is too too far from being correct. The error I am seeing is:-

❯ systemctl --user status uptime-kuma.service
× uptime-kuma.service - Uptime Kuma
     Loaded: loaded (/home/mike/.config/containers/systemd/uptime-kuma.container; generated)
     Active: failed (Result: exit-code) since Wed 2023-06-14 18:49:08 BST; 8s ago
    Process: 329668 ExecStart=/usr/bin/podman run --name=uptime-kuma --cidfile=/run/user/1000/uptime-kuma.cid --replace --rm --log-driver passthrough --runtime /usr/bin/crun --cgroups=spli>
    Process: 329677 ExecStopPost=/usr/bin/podman rm -f -i --cidfile=/run/user/1000/uptime-kuma.cid (code=exited, status=0/SUCCESS)
    Process: 329686 ExecStopPost=rm -f /run/user/1000/uptime-kuma.cid (code=exited, status=0/SUCCESS)
   Main PID: 329668 (code=exited, status=125)
        CPU: 214ms

Jun 14 18:49:07 docker1 systemd[1408]: Starting Uptime Kuma...
Jun 14 18:49:07 docker1 uptime-kuma[329668]: Trying to pull registry.hub.docker.com/library/louislam/uptime-kuma:1...
Jun 14 18:49:08 docker1 uptime-kuma[329668]: Error: initializing source docker://registry.hub.docker.com/library/louislam/uptime-kuma:1: Requesting bearer token: invalid status code from r>
Jun 14 18:49:08 docker1 podman[329668]: 2023-06-14 18:49:07.753027474 +0100 BST m=+0.024107819 image pull  registry.hub.docker.com/library/louislam/uptime-kuma:1
Jun 14 18:49:08 docker1 systemd[1408]: uptime-kuma.service: Main process exited, code=exited, status=125/n/a
Jun 14 18:49:08 docker1 systemd[1408]: uptime-kuma.service: Failed with result 'exit-code'.
Jun 14 18:49:08 docker1 systemd[1408]: Failed to start Uptime Kuma.

Am I missing something obvious? Thanks!