Thank you Piyush for this amazing series. And I joined in this course few days back. But your teaching is amazing and this course worth to many students and employess like me , to upgrade.
You have explained kubernetes architecture in simple way. Its really good. And one more thing can be cover multi master node archtecture also in the future videos.
Thank you Salman. Admission controller is more of a CKAD concept and not part of the CKA that is why I have not included it but I will cover all these important topics after we are done with CKA
Thank you Balraj! Actually it helps with keeping the session interactive. I would saved 2X time, had I recorded the videos with the cam on, 2X time saving in recording and 2X time saving in editing but that wouldnt be a good session and people will start losing interest. I had to keep the cam to a minimum size to keep that interaction but I truly understand your concern. It shouldnt hinder the visisbility. I will work on that definitely, thank you once again. This series has been recorded already but I will make some changes for the future videos.
Thank you Piyush, very clear explanation, just want to comment on the end to end flow, I think you missed CRI at the end. I am very enjoying your video
Hi Piyush, Excellent explanation and easy to grasp without getting overwhelmed. Also, can you explain role played by controller manager here 19:33. It was not covered in kubectl create pod process explanation. Thank you.
controller manager controls all the controllers in the cluster such as node controller which ensures that your node is healthy all the time, deployment controller that ensure that your pods are restarted when crashed and so on. It monitor the cluster and make sure that your desired cluster state is matching with the actual cluster state. It talks to apiserver and apiserver communicates the action to be performed to Kubelet
Hi Piyush, can your this series help me prepare for my Azure DevOps interviews that involve Kubernetes? I am struggling because I have little hands-on experience with Kubernetes in my career.
Hello Yogesh, Yes, it will definitely help. Along with this series, check out the Azure DevOps series as well: ru-vid.com/group/PLl4APkPHzsUXseJO1a03CtfRDzr2hivbD Here's the video about Azure devops interview questions as well: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-u5uSDMM9ydc.html Good luck
I thought APIserver will create the pod then update back to ETCD which is the moment ETCD create a record in the database itself. Seems like I am wrong, the record will be created first then the object (pod, deployment) will be created later. Can you help to confirm?
It's totally your call. Azure DevOps is already completed and CKA is still ongoing. If you look it from the perspective of demand in the market then Kubernetes have more weightage than Azure DevOps but it's totally up to you.
kubeconfig has your context , authentication details using which you login to the cluster. kubectl uses kubeconfig to authenticate the client to apiserver. etcd is a key value datastore that stores all the configuration data and your cluster state.
Controller manager is a collection of several controllers that ensure that your desired infra state is always reached. For example, if a pod crashes, it instruct the kubelet to restart the container.
Awesome explanation Piyush, I really loved the video and understood almost everything.. I have few doubts though. 1. So when the pod is being created, will there be 2 entries in ETCD by apiserver? First entry about the fact that a pod has to be created and the second entry after the pod has been created. Right? 2. Does a node have any limit to the number of pods it can create? 3. Does ETCD have any space constraints or is it elastic?
Thank you for the feedback. Let me try to answer: 1) One entry that pod will be created and then it will be updated with the additional details such as nodename on which it is scheduled , last transitioned time, created time, last updates and other metadata. 2) Number of pods in a node is equivalent to number of IPs available, Node is eventually a VM with a dedicated IP assigned to it, also, the capacity of the node CPU, memory, disk etc. This concept will be more clear in day16 video in which I have shared the details about resource request and limits. 3) ETCD is a database and like every DB it will have the space constraints. In one of the later videos, I have discussed how you can setup etcd in multi availability mode, you can also check defragmentation process in ETCD. Hope this helps.
@@TechTutorialswithPiyush Thank you so much for taking some time out and solving each and every doubt. I appreciate. You are the best! I was just watching day 7 video now.