kubernetes: kubeadm blocks waiting for 'control plane' in Ubuntu Wily

@luxas I was in peace but your statement “We have worked very hard” makes me feel uncomfortable. This statement sounds like that Google hired somebody who can only do leetcode problems but not be able to make real product bug free by “working very hard”. I do not have to work hard but my work is bug free.

@zhang8473 , you have no idea about how to debug. You should learn it first by google. It’t not @lukemarsden @luxas or any other k8s members’ problem. No body have same environment as you. You even did not give any useful information. I had advised you run kubelet directly already, have you done it? You say you got a “connection refused”, but have you seen your kube-apiserver is running and why? If you can’t run your cluster by kubeadm. why not try old way?

@zhang8473 neither @luxas nor I are google employees, we are both volunteers on kubernetes. And we did work hard on kubeadm, and released it in alpha to get early feedback. Please be kind, and we will work together to solve your problem. Thanks!

On Mon, 3 Oct 2016, 22:36 zhang8473, notifications@github.com wrote:

@luxas https://github.com/luxas I was in peace but your statement “We have worked very hard” makes me feel uncomfortable. This statement sounds like that Google hired someones who can only do leetcode problems but not be able to make real product bug free by “working very hard”. I do not have to work hard but my work is bug free.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/kubernetes/kubernetes/issues/33689#issuecomment-251218861, or mute the thread https://github.com/notifications/unsubscribe-auth/AAQJ0mLtQouq2GO4-w2NYlBCJRrHpwsXks5qwWdNgaJpZM4KJSbH .

@zhang8473 , 2333333333, hahhhhh, It’a mistake here. the guide for centos tell us to run kubelet, but the guide for ubuntu doesn’t.

@zhang8473 This is open-source, and as indicated is ‘alpha’ quality. So no, it’s not mature yet. User feedback such as yours is essential to evolve the software towards maturity, but such feedback needs to be supplied in a respectful manner. Personal attacks are not acceptable, as per the Kubernetes Code of Conduct:

https://github.com/kubernetes/kubernetes/blob/master/code-of-conduct.md

@bulletRush I think this applies to you as well. I can understand that you may be frustrated, but you also have an obligation to be respectful.

Well, I hope you calm down and realize that we actively chose to not “support” Wily now in the first release, but we are going to increase our scope as we have resources to test it, and develop further

We have worked very hard on this “product” as you call it, and dealing with every possible configuration the user may have clearly is impossible. You’re welcome to help contributing if you want to of course.

Please make sure kubelet is running before running kubeadm init, if it has been disabled of some reason

It’s not necessary to do systemctl start kubelet on Ubuntu, as the package automatically starts itself. That’s the convention for debian packages.

The convention for rpms is to not start the services automatically, so therefore we’ve left it to the user and documented it

@zhang8473

• first, remember kubelet argments via ps aux | grep kubelet. For me, it’s:

/usr/bin/kubelet --kubeconfig=/etc/kubernetes/kubelet.conf --require-kubeconfig=true --pod-manifest-path=/etc/kubernetes/manifests --allow-privileged=true --network-plugin=cni --cni-conf-dir=/etc/cni/net.d --cni-bin-dir=/opt/cni/bin --cluster-dns=100.64.0.10 --cluster-domain=cluster.local --v=4

• stop kubelet service: systemctl stop kubelet
• run kubelet in one terminal directly use command in first step and see what happend when kubelet startup