`connecting to engine` issues while following quickstart on F41 with Podman

[gritty nacelle]

Hello, I am toying with Dagger and started with the Quickstart turorial but I am not able to get it to run on Fedora 41 with Podman, following (sorta) the instructions on https://docs.dagger.io/integrations/podman/. I have the podman-docker RPM installed which supplies the docker symlink, and I have set up the podman machine for rootful execution:

❯ podman machine set --rootful=true
❯ podman machine start
Starting machine "podman-machine-default"
Mounting volume... /home/afri:/home/afri
API forwarding listening on: /run/user/1000/podman/podman-machine-default-api.sock
You can connect Docker API clients by setting DOCKER_HOST using the
following command in your terminal session:

        export DOCKER_HOST='unix:///run/user/1000/podman/podman-machine-default-api.sock'

Machine "podman-machine-default" started successfully

But ultimately my dagger init command keeps getting stuck on the connecting to engine step like the following:

❯ dagger init --sdk=go --source=./dagger
✘ connect 10.7s
! start engine: new client: interrupted
│ ✔ starting engine 0.4s
│ ✘ connecting to engine 10.3s
│ ! new client: interrupted
│ │ ✘ moby.buildkit.v1.Control/Info 0.3s
│ │ ! connection error: desc = "error reading server preface: command [docker exec -i dagger-engine-v0.15.1 buildctl dial-stdio] has exited with exit status 1, make sure the URL is valid, and Docker 18.09 or later is installed on the remote host: stderr=Emulate Docker CLI using podman. Create /etc/containers/nodocker to quiet msg.\nError: dial unix /run/buildkit/buildkitd.sock: connect: no such file or directory\nUsage:\n  dial-stdio [flags]\n\nFlags:\n      --addr string   The address serving the grpc api (default \"unix:///run/buildkit/buildkitd.sock\")\n  -h, --help          help for dial-stdio\n      --timeout int   The timeout in seconds for connecting to the grpc api (default 5)\n\n"
...

I have tried with DOCKER_HOST set up as advised by machine start and without it, fails both ways.

[wind root]

@gritty nacelle have you checked https://docs.dagger.io/integrations/podman/ ?

It's required to configure podman in rootful mode in order for Dagger to work

[gritty nacelle]

yeah, I even say that in my initial message

After a bit more digging it seems I am running into https://stackoverflow.com/questions/76979352/podman-cli-always-defaults-to-local-instead-of-the-default-remote-connection and even with the podman machine setup and the VM running, the plain podman (so docker too) does not run containers in the VM but locally.

Seems like I can proceed with the tutorial if I edit the compat docker script to include --remote

[wind root]

👍 good to know!