docker is dead long live podman-docker

[kiloutyg]

After more test it seems that basically networks and right management will need to be thoroughly experimented on and modified.

• Commenting all networks config from the compose files manage to allow the containers to launch, but might be an issue when reaching the reverse proxy stage.
• chownind and chmoding to 1000:1000 and 777 allowed to pass most of the command inside the containers but still did not allow the right and owning change command used in the entrypoints.

[kiloutyg]

The containers are not staying alive, always getting a sigwich at some time for some reasons.

[kiloutyg]

The issue about right management is not entirely dealt with, the entrypoint still does not have the rights to chown or chmod db, node_module and vendor (maybe some others i don't remember) files.

The in-between containers networks management seems a bit different than vanilla docker, mainly around ip affectation, podman seems to affect the IP at the building stage and at the mounting ones, and both use the same parameter present in the compose file which create a conflict, the address being already affected at the building stage. I need to either change the strategy to fully build locally or find how to affect the address correctly at the good stage.

Both of the previous issues are probably finding their sources in the rootless nature of podman.

[kiloutyg]

MariaDB docker-entrypoint looks to chown local files which cannot be done. I think i managed to launch the containers, but traefik still show a bad gateway message, which might be coherent since i had to delete the docauposte2's traefik network address, which is the gateway for this specific network.

[kiloutyg]

Ok, taken by despair, i tried to install docker, and it worked ! So we will probably just use it and docker compose instead of all the podman stuff.

Podman still might be interesting to use and redev the containers on it, but it'll take some time.

[kiloutyg]

This is being put in standby for now