This message was deleted.

Where did you struggle?

I have internal repo proxying in place, and I did do:

docker pull internal_repo/docker_io_proxy/rancher/k3s

,

docker pull internal_repo/docker_io_proxy/rancher/k3s-proxy

, and

docker pull internal_repo/docker_io_proxy/rancher/k3s-tools

, and I could see these 3 images showing up via

docker images

. Then I copied the tar file to

/var/lib/rancher/k3s/agent/images/

. Then I tried this command:

k3d cluster create k3d-local --network host --no-lb --k3s-arg "--disable=traefik,servicelb" --k3s-arg "--kube-apiserver-arg=feature-gates=MixedProtocolLBService=true" --host-pid-mode --image internal_repo/docker_io_proxy/rancher/k3s:v1.27.4-k3s1

Here are the messages with error:

INFO[0000] [SimpleConfig] Hostnetwork selected - disabling injection of docker host into the cluster, server load balancer and setting the api port to the k3s default
WARN[0000] No node filter specified
WARN[0000] No node filter specified
INFO[0000] [ClusterConfig] Hostnetwork selected - disabling injection of docker host into the cluster, server load balancer and setting the api port to the k3s default
INFO[0000] Prep: Network
INFO[0000] Re-using existing network 'host' (...)
INFO[0000] Created image volume k3d-k3d-local-images
INFO[0000] Starting new tools node...
INFO[0001] Creating node 'k3d-k3d-local-server-0'
INFO[0001] Using the k3d-tools node to gather environment information
ERRO[0015] Failed to run tools container for cluster 'k3d-local'
INFO[0015] Starting new tools node...
ERRO[0030] Failed to run tools container for cluster 'k3d-local'
ERRO[0030] failed to gather environment information used for cluster creation: failed to run k3d-tools node for cluster 'k3d-local': failed to create node 'k3d-k3d-local-tools': runtime failed to create node 'k3d-k3d-local-tools': failed to create container for node 'k3d-k3d-local-tools': docker failed to pull image '<http://ghcr.io/k3d-io/k3d-tools:5.5.1|ghcr.io/k3d-io/k3d-tools:5.5.1>': docker failed to pull the image '<http://ghcr.io/k3d-io/k3d-tools:5.5.1|ghcr.io/k3d-io/k3d-tools:5.5.1>': Error response from daemon: Get "<https://ghcr.io/v2/>": net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
ERRO[0030] Failed to create cluster >>> Rolling Back
INFO[0030] Deleting cluster 'k3d-local'
INFO[0030] Deleting 1 attached volumes...
FATA[0030] Cluster creation FAILED, all changes have been rolled back!

Glad you made it work. What's your use case for this? Usually air-gapped environments are production, no? Since k3d is a development tool, there was (so far) never the intention to make air-gapped setup a focus point. FWIW, you can set the environment variables

K3D_IMAGE_TOOLS

and

K3D_IMAGE_LOADBALANCER

so you don't have to retag locally. And the K3s image you can specify via `--image`or via config file when creating the cluster. And according to the K3s docs, you could put the images.tar into any directory and volume mount it to

/var/lib/rancher/k3s/agent/images/

so you wouldn't need the import. All of this can be put in a single line command actually. In case you didn't check them beforey here's the K3s docs: https://docs.k3s.io/installation/airgap Feel free to open a k3d issue to get full docs there as well 👍

Thanks. k3d is used in the quickstart guide for Tinkerbell tink stack deployment. But to use tinkerbell on-prem for hosts deployment, it very likely be an airgap environment in that case. I did use the

--image

parameter and passed in the k3d cluster create command as shown before:

--image internal_repo/docker_io_proxy/rancher/k3s:v1.27.4-k3s1

that did not help. I already went through the k3s airgap doc before, and that worked for me. I did the following steps already, but that did not help:

sudo mkdir -p /var/lib/rancher/k3s/agent/images/
sudo cp ./k3s-airgap-images-$ARCH.tar /var/lib/rancher/k3s/agent/images/

not sure about the volume mount one, I just did

docker inspect k3d-k3s-default-server-0

, and I don’t see any volume mount for this docker container ..