Day 27/40 - Setup a Multi Node Kubernetes Cluster Using Kubeadm

[@ajonie.a6136]

I cannot thank you enough for making these tutorials free. I must say your k8 series is a literal gold mine. Its commendable that you are not editing out the mistakes but rather walking us through your approach to troubleshooting. Thank youuuuuuuu

[@TechTutorialswithPiyush]

Thank you so much buddy for the kind words! Yes, hiding mistakes would create unnecessary confusion and limit the learning process, if we are doing hands-on, mistakes are bound to happen, we have to find the issue and fix it, that's all that matters

[@ashlok_chaudhary]

It's truly the best learning experience to configure so many things, connect them seamlessly, and troubleshoot any errors that arise. Thank you so much for such a detailed video and this wonderful free series!

[@saadahmed3581]

Thank you very very very much,, I wasted 28 hours continuously to setup on ARM64 system, and you just saved me. Take love

[@TechTutorialswithPiyush]

Glad it helped :)

[@harshgupta5113]

Damn, this gave me so much insights. For me the most important part was the troubleshooting of the worker node not able to run the kubectl command. I remember facing this issue in my company but my colleague solved it for me. I asked him on how he did it but I couldn't get it in my head. Your explanation was so simple and effective and now I completely understood the concept. Thank you very much bro.

[@TechTutorialswithPiyush]

Awesome bro! Now next time you show him how it's done!

[@kyawzawaung3712]

Thank you very much for giving this quality content free.

[@TechTutorialswithPiyush]

It's my pleasure

[@VijayPolsani]

Appreciate all your hard work and thinking out loud while configuring

[@TechTutorialswithPiyush]

Thank you

[@bhanubisht8]

Completed the video....!!!!!!

[@abc-edm]

thanks Piyush, nice video with rich info with live troubleshooting ...

[@TechTutorialswithPiyush]

Glad it was helpful!

[@the_hustling_wanderer]

Thanks Piyush for the wonderful video

[@TechTutorialswithPiyush]

You're welcome

[@eapenmani]

Great Learning .Thank you!

[@TechTutorialswithPiyush]

welcome 🙂

[@chiranjeevisundarasetty2240]

Learned allot from this video thank bro

[@TechTutorialswithPiyush]

You're welcome, Keep practicing!

[@varunjoshi4410]

Thank you, this was helpful.

[@TechTutorialswithPiyush]

kind.sigs.k8s.io/docs/user/ingress/

[@GrigTech]

Like and comment, thank you for work you have done, great learning tutorial.

[@TechTutorialswithPiyush]

Thank you! Cheers!

[@suhasdhoke1735]

Thank you ❤

[@itlearntv]

I learn allot of this video of your. thanks

[@TechTutorialswithPiyush]

Glad to hear that you're finding the series helpful.

[@ethagbemichael]

Great job 👏 👍

[@TechTutorialswithPiyush]

Thanks

[@linuxgenius4667]

thats exactly i was waiting for, on how to configure bare matel

[@TechTutorialswithPiyush]

I think you can follow the same steps for bare metal, also, check Kubernetes the hard way

[@raviveer11]

Super thanks for this video thanks a lot

[@TechTutorialswithPiyush]

Welcome 😊

[@TheZikori]

Good job sir👍

[@TechTutorialswithPiyush]

Thanks and welcome

[@itsupportbyharshit6285]

Thanks

[@AbdulMateen-bm3kv]

Thank you very much

[@ajitsahoo4875]

Hy Piyush, why don’t you use GCP for this practice? Is there any video where you explain this concept using GCP?

[@TechTutorialswithPiyush]

replied already on the other comment

[@AyoubKoussy-j6i]

Hi piyush thank you for the wonderful series.I have one question ,I want to create a cluster of hadoop in the top of k8s ,any idea how to start

[@TechTutorialswithPiyush]

Sorry buddy I never worked with Hadoop but you can follow the official documentation and should be good

[@ayushjaiswal167]

Hi Piyush! Thanks for such a nice tutorial. However, I have a question that is it a best and secure practice to place admin.conf file in worker nodes. How such things are handled in a production level setup as per industry standards.

[@TechTutorialswithPiyush]

Thank you! If you are using on premises cluster such as Vmware or bare metal then you have an Admin workstation which has the kubeconfig and master node already have admin.conf which is generated by default. In case of a managed kubernetes service such as GKE, you dont get access to the master node , kubeconfig is getting generated through the cloud console or cloud shell using the user's authentication and authorisation

[@GopiVivekManne]

Excellent demo Piyush!! I love the debugging part too 💯👍 Although the master node requires a minimum of 2 GB RAM and 2 vcpu, I attempted to configure it using a t2.micro instance with kubeadm to see if it would work or not. It failed to initialize, I had to upgrade the node to a t3.small to complete the setup.(kubeadm reset command is helpful)

[@TechTutorialswithPiyush]

You're welcome buddy! Yes, it needed atleast 2 GB RAM, t2.micro would not work

[@bhanubisht8]

so can we do that in free tier account...???

[@sjeyendran]

Detail method sharing kubernets installation

[@TechTutorialswithPiyush]

Thank you, I have added all the steps in the Github repo as well, check it out

[@harshgupta5113]

Bro waiting for the K8s the hard way setup video.

[@TechTutorialswithPiyush]

Will create in some time, for sure

[@ajitsahoo4875]

Another question Why don’t you use EKs directly?

[@TechTutorialswithPiyush]

Because it's a managed service, many things including the control plane is managed by the cloud provider hence, limited learning for you. The right way is: Learn Docker --> Kind cluster --> Kubeadm and then EKS/AKS/GKE

[@ganeshpawar231]

Q. I did this set up on EC2 as you did and further deployed ingress controller with LoadBalancer type. Ingress pod is running fine. CCM should have given an Elastic IP but I am not able get External IP for ingress service. It is in pending state. Why? Does it require any other configurations?

[@bhanubisht8]

Comment for target...!!!!!

[@alekhya0510]

Hi piyush , I am following CKA series it is wonderful,I am very thankful.While installing through kubeadm way at the time of initialization kubeinit why do we provide public ip of vm , as public ip keep on changing while restarting.will it cause any error??

[@TechTutorialswithPiyush]

Hello buddy, even if the IP changes, your cluster still gets registered

[@Ramm165]

Hi piyush thank you for the wonderful series.I have one question.I have 2 desktops with 64gb of ram and can i create vms on them and follow this tutorial to setup kubernetes cluster on them ?

[@TechTutorialswithPiyush]

That would be fantastic, go for it and let me know if you face any issues.

[@Ramm165]

@TechTutorialswithPiyush thank you sure

[@Ramm165]

@@TechTutorialswithPiyush installed locally and it worked thank you

[@TechTutorialswithPiyush]

@@Ramm165 great job buddy

[@AlitafSd]

thanks for such a clear explanations. i have one question As we know to build image we required the docker. And we do CI part using the jenkins. So my question is that , in a company does this docker daemon and jenkins installed on same ec2 instance or different?

[@TechTutorialswithPiyush]

Thank you brother 🙏 Usually Jenkins works in a master slave architecture wherein Jenkins master provision multiple build servers to run the builds and these servers uses base image with docker installed. These could be on vms such as ec2 of containers or Kubernetes pods

[@anishkaje250]

Hi Im trying to do this handson but unable to do it. Initially the Kubectl works and I recieved the get pod, get node so on.. After that it asks "specified host is right or not?" How to fix this issue? I followed your document but unable to solve this kubelet is in active state

[@TechTutorialswithPiyush]

Hello Anish, If your kubectl commands are not working , there could be issue with your apiserver, or your kubeconfig is not correct. Can you check the apiserver logs and kubelet logs , also run crictl ps commands to ensure apiserver is running.

[@anishkaje250]

@@TechTutorialswithPiyush okay sure thanks for the response

[@dr.hemantchauhan2613]

😊

[@mubinbhojani2779]

sir please give all steps to install k8s cluster using kubeadm

[@TechTutorialswithPiyush]

I have already shared those, check out the GitHub repo for the same.

[@ganeshpawar231]

Hi Piyush sir, Nice tutorial...! Thanks for sharing each detail of installation from this video. By the way, crictl ps command, containerd.sock error might be due to it has owner root and group root. You are running it as ubuntu user. We can simply change owner to ubuntu of containerd.sock. Hope it will work. What do you think?

[@TechTutorialswithPiyush]

Thanks for the tip, I appreciate the feedback!

[@99nanaji]

Hi Piyush, Please upload the scripts used for multi-node Kubernetes cluster installation to Git. Thank you

[@TechTutorialswithPiyush]

Hello and thank you for the reminder, updated the repo now.

[@abhimanyuhowshe7873]

Why we are explicitly installing runc as contained is using it implicitly?

[@TechTutorialswithPiyush]

even though runc is used by containerd, it's a seperate component. Docker engine communicates with containerd and containerd spins up runc or other oci compliant runtime to run containers

[@abhimanyuhowshe7873]

@TechTutorialswithPiyush but as we installed containerd runs was also installed with it?

[@AnkitKothawade]

Hello brother, just want to ask a question is this course is enough to prove an experience of 2 years . i hope you understood what i am saying. hoping for your response. Thank you!!!

[@TechTutorialswithPiyush]

Hello Ankit, from Kubernetes perspective its good enough but you also would have to learn a managed Kuebrnetes service on Cloud, a few more concepts such as GitOps, helm, ArgoCD, operators, CRDs etc. All the best. For Gitops and ArgoCD , I have a detailed video in #10weeksofcloudops series, check it out

[@pradipakshar]

45:10 🤣🤣🤣🤣

[@TechTutorialswithPiyush]

Happens 🤣 Usually during the long recordings

[@dr.hemantchauhan2613]

❤

[@yashjha2193]

I have practiced, pods, RC, RS, Deploy, ManualScheduling, Node selector, Node affinity, Resource quota and rest.. But Ingress, RBAC can this be practiced using Kubeadm thing? Please tell me

[@TechTutorialswithPiyush]

That is correct!

[@guntreddirajesh]

😍

[@manivannanm9907]

Hi piyush, Thank you for the wonderful series. I Learn a lot of this video. I have one issue after join the worker node, worker node is not ready and it shows (describe worker node) kubelet is not ready -container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:Network plugin returns error: cni plugin not initialized. how we resolve this issue. thank you

[@TechTutorialswithPiyush]

Hello, You need to install the cni plugin such as calico, you can check the GitHub repo day27 folder, it has the troubleshooting steps as well

[@theinteger]

👉👉Can we follow these steps on local VM?👀👀👀

[@TechTutorialswithPiyush]

You can use a virtualized environment as mentioned in the video, either use virtual box or multipaas and then you can spin multiple VMs inside your local VM.

[@theinteger]

@@TechTutorialswithPiyush Great, i will setup one VM for master and other for worker | (ubuntu)

[@roopeshsaravanan8609]

But setting up a multi node k8s cluster in a VM doesn't make sense correct?

[@TechTutorialswithPiyush]

It totally makes sense. We use this kind of setup when we need full control over the control plane, administrative and operational tasks, access to all the underline nodes etc plus it is cheaper when compared to a managed cloud service such as GKE/AKS/EKS etc. A manage cloud service takes care of your administrative and operational tasks, control plane HA and scalability, comes with some additional features such as Node auto provisioning, Cluster Autoscaler etc and that is why it is costlier than self managed kubernetes. You choose the setup based on your needs and requirements.

[@Usama-Tanoli]

Hi Piyush, I set up a Kubernetes cluster with kubeadm on my laptop using VMs in VirtualBox last month. However, when I change my Wi-Fi or internet connection, the cluster stops working due to the IP change. Can you suggest a solution to fix this issue?

[@TechTutorialswithPiyush]

Hello, this shouldn't be the case, what steps have you performed? can you compare it with the steps mentioned in this video?

[@Usama-Tanoli]

​@@TechTutorialswithPiyush i will perform this lab tomorrow.

[@VishalMK-i4f]

calico-node is showing 0/1 running, and found this issue " Readiness probe failed: 2024-11-14 14:42:48.582 [INFO][237] confd/health.go 202: Number of node(s) with BGP peering established = 0"

[@TechTutorialswithPiyush]

In the GitHub repo of day27 folder, I have added a section for this, can you please follow the steps

[@kaarthickpk]

🤩🤍💥💫

[@ganeshpawar231]

Hi sir, I did this also on t2.micro Ubuntu 22.04 LTS. It's working. But thing is it was throwing error when I ran Kubeadm init --pod-network-cidr Error---> Error NumCPU : the number of available CPU is 1 is less than the required 2 Error Mem: the system Ram 957 MB is less than the minimum 1700 MB. I used same command with flag --ignore-preflight-errors=NumCPU,Mem And it got successful. And again joining the cluster was also throwing error I used --ignore-preflight-errors=IsPrivilegedUser with kubeadm join command And it got successful. But thing to remember It is showing 200MB available on master node and 500MB available on worker node . Thanks🎉

[@TechTutorialswithPiyush]

That's sort of a minimum requirement, you can bypass the preflight checks but it could create issues if you run multiple workloads as your control plane won't be able to handle the load but for now you should be good

[@ganeshpawar231]

@@TechTutorialswithPiyushyes, I wanted to check if it is working as you requested in video😄

[@kalyanprogrammer6134]

why there is mismatch between ports range in k8s document to your video totally confusing not even medium learners friendly sorry to say this very time consuming boring content for me time Waste tq

[@TechTutorialswithPiyush]

Thank you for the feedback, have a good day

[@TechTutorialswithPiyush]

Bdw, to answer your question, if you start a series from video 27, you will find it overwhelming. The k8s documentation for ports are generic, In my video, it is ec2 specific hence, extra ports but I guess you will find a better video than this, good luck

[@ajitsahoo4875]

Hy Piyush, why don’t you use GCP for this practice? Is there any video where you explain this concept using GCP?

[@TechTutorialswithPiyush]

Because it's a managed service, many things including the control plane is managed by the cloud provider hence, limited learning for you. The right way is: Learn Docker --> Kind cluster --> Kubeadm and then EKS/AKS/GKE Plus this series was for anyone looking for learn Kubernetes, not cloud specific. You can learn Kubernetes and then move to any cloud for further learning