如何在k8s内部部署karmada

[SZT0728]

看了karmada的启动脚本 都是对给定的一个k8s集群从外部进行部署karmada，有没什么方式可以在k8s集群内部部署karmada的？？:

What do you think about this question?: 有官方开发者支持解答一下疑问吗

Environment:

• Karmada version:
• Kubernetes version:
• Others:

[RainbowMango]

Have you checked Installing Karmada document?

[SZT0728]

when i use kubectl-karmada install karmada，the output failed message is follow:

I0111 13:59:09.146318 477525 deploy.go:217] Update APIService 'v1alpha1.cluster.karmada.io'
F0111 13:59:09.149806 477525 deploy.go:95] apiservices.apiregistration.k8s.io "v1alpha1.cluster.karmada.io" not found

how can fix this？？

[SZT0728]

Environment：kubectl-karmada is build from source code which tag is v1.0.0
k8s server version: version.Info{Major:"1", Minor:"20", GitVersion:"v1.20.5", GitCommit:"6b1d87acf3c8253c123756b9e61dac642678305f", GitTreeState:"clean", BuildDate:"2021-03-18T01:02:01Z", GoVersion:"go1.15.8", Compiler:"gc", Platform:"linux/amd64"}

[RainbowMango]

Oh, this is a bug that has been fixed by #1228.

I think you can re-build the kubectl-karmada with the latest code. Or, create the APIService below manually before run kubectl-karmada init:

karmada/artifacts/deploy/apiservice.yaml

Lines 1 to 16 in eb4b3e2

	apiVersion: apiregistration.k8s.io/v1
	kind: APIService
	metadata:
	name: v1alpha1.cluster.karmada.io
	labels:
	app: karmada-aggregated-apiserver
	apiserver: "true"
	spec:
	insecureSkipTLSVerify: true
	group: cluster.karmada.io
	groupPriorityMinimum: 2000
	service:
	name: karmada-aggregated-apiserver
	namespace: karmada-system
	version: v1alpha1
	versionPriority: 10

[RainbowMango]

cc @prodanlabs

[prodanlabs]

Yes, unfortunately, a karmadactl needs to be compiled manually.

[prodanlabs]

hi @RainbowMango
Is it possible to add a karmadactl to the release page, what do you think?

[orangedeng]

I went into the same issue and I can't continue to install karmada after init failed.

[RainbowMango]

hi @RainbowMango Is it possible to add a karmadactl to the release page, what do you think?

We can't do that even it will work. But we can cut a new release v1.0.1 for this.

[prodanlabs]

ll karmada after init failed.

you can compile the latest karmadactl from source，you need a golang environment.

git clone https://github.com/karmada-io/karmada.git
cd karmada
make kubectl-karmada  

[prodanlabs]

What is the current-context of kubeconfig?
cat /root/.kube/k8s-cluster1.config | grep current-context | sed 's/: /\n/g' | sed '1d'

[RainbowMango]

That is probably because karmada-controller-manager can't access member cluster, I don't know where you installed Karmada, but please check the endpoint of the member cluster, by following command:

kubectl get clusters k8s-cluster1 -o jsonpath='{.spec.apiEndpoint}'

Or, you can describe the clusters to see the conditions/events:

kubectl describe clusters member1

We usually use hack/create-cluster.sh to create a kind cluster for testing, we'll update the endpoint to deal with the network issues.

[SZT0728]

What is the current-context of kubeconfig? cat /root/.kube/k8s-cluster1.config | grep current-context | sed 's/: /\n/g' | sed '1d'

the karmada-scheduler's log show that
[core] grpc: addrConn.createTransport failed to connect to {karmada-scheduler-estimator-kind-k8s-cluster1:10352 karmada-scheduler-estimator-kind-k8s-cluster1:10352 0 }. Err: connection error: desc = "transport: Error while dialing dial tcp: lookup karmada-scheduler-estimator-kind-k8s-cluster1 on 10.1.0.10:53: no such host". Reconnecting...
and i guess that the k8s-cluster1 is create by kind and run in docker，and the karmada running in k8s,maybe because they are not the same net spaces, so the karmada failed to connect k8s-cluster1，is it？

[prodanlabs]

What is the current-context of kubeconfig? cat /root/.kube/k8s-cluster1.config | grep current-context | sed 's/: /\n/g' | sed '1d'

the karmada-scheduler's log show that [core] grpc: addrConn.createTransport failed to connect to {karmada-scheduler-estimator-kind-k8s-cluster1:10352 karmada-scheduler-estimator-kind-k8s-cluster1:10352 0 }. Err: connection error: desc = "transport: Error while dialing dial tcp: lookup karmada-scheduler-estimator-kind-k8s-cluster1 on 10.1.0.10:53: no such host". Reconnecting... and i guess that the k8s-cluster1 is create by kind and run in docker，and the karmada running in k8s,maybe because they are not the same net spaces, so the karmada failed to connect k8s-cluster1，is it？

karmada-scheduler-estimator is a non-essential component.You can look at the karmada-controller-manager log

 kubectl -n karmada-system logs -f karmada-controller-manager-b954bdf4c-q77zp

[orangedeng]

I noticed that a new host cluster is created inside my current k8s cluster. I think we are looking forward to let the specific cluster to become the host cluster. Is there any way to do so?

[pigletfly]

You can try installing with helm

[RainbowMango]

/close
Hi @SZT0728 Thanks for reporting, If this issue still exist, please feel free to let us know.

[karmada-bot]

@RainbowMango: Closing this issue.

In response to this:

/close
Hi @SZT0728 Thanks for reporting, If this issue still exist, please feel free to let us know.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.