How to use .env file with Quadlet .kube?

[poperigby]

Does podman kube play <service>.yaml automatically look for .env files in the directory of the YAML file, or do I have to manually specify them?

[rhatdan]

Is .env a kubernetes concept?
@ygalblum PTAL

[poperigby]

Is there any feature that would allow me to keep sensitive information separate from the configuration of containers? If not, that would be useful.

[rhatdan]

You can use environment variables via systemd if you want, or environment variables set in a quadlet. Or use podman secrets.
Not sure how this is much different?

[poperigby]

Is it possible to then use those in my YAML file?

[rhatdan]

I believe you can use the secrets in the YAML file as environment variables. @ygalblum @umohnani8 Thoughts?

[ygalblum]

As I wrote earlier, in order to use a secret in a K8S Pod, the Podman secret itself must also be in K8S Secret. As noted in the Create secrets section of my blog, since the demo uses both .container (which translated to a podman run command) and .kube (which translated to a podman kube play command) files, the same information (MYSQL_ROOT_PASSWORD) is stored in two secrets.

The reason is that when using a secret with podman run the entire content of the secret is uses, while when using a secret with podman kube play Podman parses the content of the secret as a K8S Secret and only the data field (and the keys within it) is used.

Currently, I don't think there is an automatic way to combine separated configuration and data (in this case secret) and I understand the request.

However, we need to keep in mind that podman kube play tries to stay aligned with the behavior of kubectl. Both support substitution based on the environment variables defined in the Pod spec (see here). So, how will such an automatic process know whether a placeholder should be substituted with Host machine variables or with the ones defined in the Pod Spec?

Not sure how to address this