This is very helpful and I discovered some beautiful things in KinD, Keep working hard, doing the good things and always remember your Vision, Piyush Thank you very much.
Hello Anil, Yes control plane is also running on VMs and it needs HA. When you are using a managed cloud service and regional cluster , they already take of this by deploying your control plane nodes in 3 availability zones for HA.
Hi Piyush, I have 3 part question for you: 1. Does CKA exam uses kind to setup Cluster/Control-plane? 2. Could we have mapped the first node to the multi-node cluster we created later? 3. How can we get our nodes to scale automatically with the help of CPU/memory metric or if some other metric is also possible like RPS/RT, this seems possible with autoscaling in AWS?
Hello Sanyam, 1) They will have a remote desktop machine based on Ubuntu and they will provide you around 6 clusters in which you will be performing tasks. As there will be tasks around cluster upgrade, which is not possible with Kind, I dont think, they are using Kind. That is the reason we will be using multiple Virtual machines based cluster using kubeadm from video 27 onwards which is similar to the exam setup. You can check the below as well docs.linuxfoundation.org/tc-docs/certification/lf-handbook2/exam-user-interface/examui-performance-based-exams 2) We actually created 2 clusters, 1st cluster only had 1 node which serves as a control plane + worker node and other cluster has 3 nodes 1 master and 2 worker nodes 3) Yes, to scale out and add more nodes you use cluster autoscaling which is a feature provided by managed cloud services such as eks, aks , gke etc
"Great video! I'm learning a lot." I have a doubt about the installation of Kubernetes, How to decide which type of Kubernetes has to install means by a specification of Hardware or instance type or anything else?
Thank you for the feedback. It depends on various factors such as cost, purpose , high availability and performance requirements, the level of control you want and so on. I have discussed that in detail in one of the later videos, stay tuned. For now, we have installed Kind which is ideal for learning, POC and a quick setup of kubernetes locally, this is better than minikube as it is lightweight , doesnt require much resources and supports multi-node setup.
KIND is not part of CKA exam but we have used it so that we can get started quickly without understanding the components and later we have used kubeadm which needed a good understanding of Kubernetes
Make sure you dont just watch the videos. CKA exam is complete hands-on, you have to practise a lot to clear the exam. try to complete all the tasks given in the GitHub repo. Yes, this will be more than enough
The commands are kubectl and its same for any kubernetes cluster however, minikube doesnt support multi node cluster , you will get only a single node also it is heavier than KinD. That is why I'd highly recommend KinD
Minikube uses your local machine to setup a single node cluster however, Kind creates multiple containers on your machine and each container acts as a seperate node, hence it support multi-node cluster.
i tried setting up in my vmware with a rhel8 vm it works fine and fails after that starting control plane section This is the error , failed to create cluster : failed to init node with kubeadm How to fix?
That is why I did not start with kubeadm installation. I started with KIND so that every beginner can spin up the cluster in the beginning and once we have covered enough concepts, we can spin up the kubeadm cluster which is there in day27 video. I am assuming my audience are beginner to Kubernetes and if they face any installation/setup issue , they wont be able to fix that in the beginning, as we have not covered a lot of concepts now. We need to understand logging, different kubectl commands, static pods, TLS/SSL and other important concepts before that. I'd suggest setting up kind cluster now and you can follow along with kubeadm in day27 and let me know if you still face any issue.
@shivishdurkalhan-dev , @TechTutorialswithPiyush , I am facing the exact same issue while trying to setup cluster using kind on RHEL 8.5 VM It fails at the time of starting control-plane. "This error is likely caused by: - The kubelet is not running - The kubelet is unhealthy due to a misconfiguration of the node in some way (required cgroups disabled)" Please let me know how to resolve this.
It's only there in intro and outro. I tolerate a lot of comments for the quality content that I'm providing free of cost, I guess you can tolerate a couple of minutes of annoying music 😄
@@chaudharyu.8453 No hate to anyone, I'm always happy to hear the constructive feedback which I can work on to improve the content quality but I don't see such comments on premium paid courses or even udemy courses.
sir please help me ,i trying to resolve it from past 2 days but not working single node cluster build successfully but problem occur when build multinode cluster and i am using the fedora os ---------------------------------------------------------------------------------------------------------------------- failed to create cluster: failed to join node with kubeadm: command "docker exec --privileged nono-worker2 kubeadm join --config /kind/kubeadm.conf --v=6" failed with error: exit status 1