Minikube dashboard requires internet

[A-Crawford]

What Happened?

Hello,

I am attempting to run minikube on an airgapped windows 11 system. It must use HyperV, as docker desktop is not available. I am attempting to follow the introuductory, hello-minikube, learning.

I have been able to get Minikube started by running minikube start --download-only on a windows pc and transfering across .minikube and starting minikube on the airgapped pc. However, when running minikube dashboard the terminal hangs on 'checking proxy health...'.

I looked into the log files and saw that Minikube was trying to fetch the dashboard and metric images from the web and that is why I assume it was failing. To try and solve this i;

• Started a Minikube on a network attached PC, got dashboard running etc. and transferred the .minikube files to the airgapped PC. This failed and gave lots of powershell errors
• Did the same as above but did minikube cache add for all images listed under minkube image list as I saw that the cache folder in .minikube did not have all of them
• As cache add is depriciated I did the same with image load
• Started Minikube with download only, and then added all of the images required, then transferred and ran minikube start
• Saved the docker images for dashboard and metric to .tar files from the registry, and on the airgapped pc loaded these .tar into minikube

All of these resulted in the dashboard not starting.

I have found that when running minikube start --vm-driver=hypverv --download-only there I get:

😄 minikube v1.37.0 on Microsoft Windows 11 Pro 10.0.26100.6899 Build 26100.6899
✨ Using the hyperv driver based on user configuration
💿 Downloading VM boot image ...
> minikube-v1.37.0-amd64.iso....: 65 B / 65 B [---------] 100.00% ? p/s 0s
> minikube-v1.37.0-amd64.iso: 370.78 MiB / 370.78 MiB 100.00% 40.96 MiB p
👍 Starting "minikube" primary control-plane node in "minikube" cluster
💾 Downloading Kubernetes v1.34.0 preload ...
> preloaded-images-k8s-v18-v1...: 337.07 MiB / 337.07 MiB 100.00% 39.59 M
> kubectl.exe.sha256: 64 B / 64 B [---------------------] 100.00% ? p/s 0s
> kubectl.exe: 59.26 MiB / 59.26 MiB [-------] 100.00% 82.63 MiB p/s 900ms
✅ Download complete!

I have attempted to remedy not having the dashboard and metrics scraper included by running

minikube image load docker.io/kubernetesui/dashboard:v2.7.0
minikube image tag docker.io/kubernetesui/dashboard:v2.7.0 docker.io/kubernetes/dashboard:v2.7.0
minikube image load docker.io/kubernetesui/metrics-scraper:v1.0.8
minikube image load docker.io/kubernetesui/metrics-scraper:v1.0.8

Minikube starts successfully on the airgapped pc

😄 minikube v1.37.0 on Microsoft Windows 11 Pro 10.0.26100.6899 Build 26100.6899
E1107 19:16:17.493486 16284 start.go:829] api.Load failed for minikube: filestore "minikube": Docker machine "minikube" does not exist. Use "docker-machine ls" to list machines. Use "docker-machine create" to add a new one.
✨ Using the hyperv driver based on existing profile
👍 Starting "minikube" primary control-plane node in "minikube" cluster
🔥 Creating hyperv VM (CPUs=2, Memory=6144MB, Disk=20000MB) ...
❗ Failing to connect to https://registry.k8s.io/ from both inside the minikube VM and host machine
💡 To pull new external images, you may need to configure a proxy: https://minikube.sigs.k8s.io/docs/reference/networking/proxy/
🐳 Preparing Kubernetes v1.34.0 on Docker 28.4.0 ...
🔗 Configuring bridge CNI (Container Networking Interface) ...
🔎 Verifying Kubernetes components...
▪ Using image gcr.io/k8s-minikube/storage-provisioner:v5
🌟 Enabled addons: storage-provisioner, default-storageclass
🏄 Done! kubectl is now configured to use "minikube" cluster and "default" namespace by default

however on running minikube dashboard the following messages are shown:

🔌 Enabling dashboard ...
▪ Using image docker.io/kubernetesui/dashboard:v2.7.0
▪ Using image docker.io/kubernetesui/metrics-scraper:v1.0.8
💡 Some dashboard features require the metrics-server addon. To enable all features please run:

    minikube addons enable metrics-server

🤔 Verifying dashboard health ...
🚀 Launching proxy ...
🤔 Verifying proxy health ...

and nothing happens.

when running Image list there is none of the dashboard images.

registry.k8s.io/pause:3.10.1
registry.k8s.io/kube-scheduler:v1.34.0
registry.k8s.io/kube-proxy:v1.34.0
registry.k8s.io/kube-controller-manager:v1.34.0
registry.k8s.io/kube-apiserver:v1.34.0
registry.k8s.io/etcd:3.6.4-0
registry.k8s.io/coredns/coredns:v1.12.1
gcr.io/k8s-minikube/storage-provisioner:v5

I would add the images now that minikube is running but I would not be able to transfer it to a differnt pc if I understand correctly. I have tried but on running the image I get a plethora of errors.
Is there a way to transfer a functioning, fully equipped minikube configuration (.minikube) to a different PC?

I cannot copy the log files from the airgapped pc but I have replicated the error locally.
Interestingly, once the problem has occured, if I recconnect to the internet I am still unable to start the dashboard.

Attach the log file

log.txt

Operating System

Windows

Driver

Hyper-V

[A-Crawford]

On further investigation I have found that the dashboard will succesfully start if I perform

minikube image pull docker.io/kubernetesui/dashboard:v2.7.0
minikube image pull docker.io/kubernetesui/metrics-scraper:v1.0.8

before going offline and starting the dashboard.

However, if I pull the images and save them to a .tar either with docker or minikube and then load those saved .tar files into a new minikube cluster - the dashboard will not start without internet.

[A-Crawford]

I believe I have found the cause of the issue but not the solution.
In the yaml scrips the dashboard and metric-scraper images are called by :@.

However, when use docker save / load or minikube image save /load, the sha changes for the images and they no longer align with the ones specified in the yaml - which leads to continously attempting to try and pull the image from the internet.

Is there any fix for this, or solutions on how to ensure the SHA does not change when save and loading images?

[A-Crawford]

I have found a temporary solution. Due to ther SHA values changing when save and loading I found that amending the deployment to not include the SHA solved the issue.

I created a new instance of minikube
Ran minikube dashboard so that the namespace was created
then ran the following

kubectl set image deployments/kubernetes-dashboard kubernetes-dashboard=docker.io/kubernetesui/dashboard:v2.7.0 --namespace=kubernetes-dashboard
kubectl set image deployments/dashboard-metrics-scraper dashboard-metrics-scraper=docker.io/kubernetesui/metrics-scraper:v1.0.8 --namespace=kubernetes-dashboard

To update the images