The guide generates some files on one of three ETCD
hosts. It then copies files from host 1 to 2 and 3.
Due to that some file paths differ.
Update step 7 to reflect that and to match step 6.
The command "kubeadm kubeconfig user" is missing
some examples and more details on why it is needed.
- Add a new section "Generating kubeconfig files
for additional users" under the kubeadm-certs page.
- Link to this section from the kubeadm-kubeconfig reference page.
- Link to this section from the create-cluster-kubeadm page.
There were a couple of reported problems with this guide:
- The introductory paragraph talks about single control plane nodes
and does not mention the different options for HA etcd. Clear the
language to reduce the confusion and cross-link to the ha-topology
page.
- The hostname / IP detection in kubeadm can end
up with values not suitable for the certificates that kubeadm
generates for all etcd instances. Ensure that the hostnames / IPs
are pinned by the user in the example script.
Side cleanup related to the dockershim removal:
- Use containerd in the setup example and don't mention docker as
a requirement.
* Fixes grammatical errors
* Removes period as recommended in review
* Update content/en/docs/setup/production-environment/tools/kubespray.md
Co-authored-by: Rey Lejano <rlejano@gmail.com>
Co-authored-by: Rey Lejano <rlejano@gmail.com>
* Fix broken anchor link of #type-nodeport in service.md
* Update anchor link from #nodeport to #type-nodeport
Update anchor link from #nodeport to #type-nodeport in overview.md
* Update anchor link from #nodeport to #type-nodeport
Update anchor link from #nodeport to #type-nodeport in ingress.md
* Update anchor link from #nodeport to #type-nodeport
Update anchor link from #nodeport to #type-nodeport in troubleshooting-kubeadm.md
* Update anchor link from #nodeport to #type-nodeport
Update anchor link from #nodeport to #type-nodeport in connecting-frontend-backend.md
* Update anchor link from #nodeport to #type-nodeport
Update anchor link from #nodeport to #type-nodeport in source-ip.md
The "kubelet-finalize" step does not work on worker nodes,
because commonly they do not have the cluster CA key and all
"init" phases have a pre-step to check for the existence of
the CA key, designating the cluster as one that uses external CA
or not.
Changing this behavior is complicated in kubeadm, thus manually
instruct the user how to do edit the file and to restart the kubelet.
This is already what we do in:
https://kubernetes.io/docs/tasks/administer-cluster/kubeadm/kubeadm-certs/#check-certificate-expiration
- Include more details about custom image repositories, such
as making sure that paths in custom repositories comply with
kubeadm execution.
- Move the section "without internet connection" above
the section about custom images.
- Link the "create-cluster-kubeadm.md" guide these new section
as a pre-req step for nodes.
Priyanshu Ahlawat