1 00:00:00,000 --> 00:00:10,266 Hi let us understand about various components of kubernetes. Within kubernetes 2 00:00:10,266 --> 00:00:18,099 the working hours is a pod. So pod is going to be the working hour where the pod can have 3 00:00:18,100 --> 00:00:20,766 n number of containers, as we discussed 4 00:00:20,766 --> 00:00:23,566 earlier container is nothing but a entity 5 00:00:23,566 --> 00:00:26,432 which is going to encapsulate an image from 6 00:00:26,433 --> 00:00:29,599 any repository. Let's assume I do have a 7 00:00:29,600 --> 00:00:32,732 repository and it may have n number of 8 00:00:32,732 --> 00:00:35,599 images within it, let's assume I do have an 9 00:00:35,600 --> 00:00:40,233 image http, a web server I can use this 10 00:00:40,233 --> 00:00:42,966 particular image as a part of this particular 11 00:00:42,966 --> 00:00:47,199 pod and pod can have n number of other 12 00:00:47,200 --> 00:00:50,066 containers as well along with this particular 13 00:00:50,066 --> 00:00:54,399 container. Say if I have http I may need some 14 00:00:54,400 --> 00:00:57,200 other containers to manage this particular 15 00:00:57,200 --> 00:01:00,600 container mostly we will be using a mesh 16 00:01:00,600 --> 00:01:03,966 mostly we will be using some sort of a proxy 17 00:01:03,966 --> 00:01:06,599 For now assume there may or there may not 18 00:01:06,600 --> 00:01:10,000 be another container and a pod can have 19 00:01:10,000 --> 00:01:13,266 n number of containers. Kubernetes is all 20 00:01:13,266 --> 00:01:15,732 about managing or orchestrating this 21 00:01:15,733 --> 00:01:19,233 particular pod. I may want to start n number 22 00:01:19,233 --> 00:01:22,799 of pods and manage them within the 23 00:01:22,800 --> 00:01:26,033 infrastructure. So we call that as an 24 00:01:26,033 --> 00:01:28,733 orchestrator for that I am going to use a 25 00:01:28,733 --> 00:01:32,833 client code kubectl. How this particular pod 26 00:01:32,833 --> 00:01:35,599 should get created, so declaratively I will be 27 00:01:35,600 --> 00:01:38,866 defining the structure of the pod that will be 28 00:01:38,866 --> 00:01:43,199 using a yaml file. So this will have the details 29 00:01:43,200 --> 00:01:46,100 on what should be the image and what port 30 00:01:46,100 --> 00:01:48,866 of this particular container should get exposed 31 00:01:48,866 --> 00:01:51,732 and what type of pod should get created all 32 00:01:51,733 --> 00:01:54,099 this informations will be available as a part of 33 00:01:54,100 --> 00:01:58,600 this particular xyz.yaml file. Now I can use this 34 00:01:58,600 --> 00:02:02,633 particular kubectl and use this yaml file and 35 00:02:02,633 --> 00:02:05,599 create the pod. So the pod will get created 36 00:02:05,600 --> 00:02:08,633 using the kubernetes cluster and cluster 37 00:02:08,633 --> 00:02:11,599 it's going to have master and worker and 38 00:02:11,600 --> 00:02:13,800 there are n number of components available 39 00:02:13,800 --> 00:02:16,466 as a part of master worker architecture to 40 00:02:16,466 --> 00:02:19,632 manage this particular pods. Assume I do 41 00:02:19,633 --> 00:02:23,833 have a master and n number of workers. As 42 00:02:23,833 --> 00:02:25,799 we discussed earlier the worker needs to 43 00:02:25,800 --> 00:02:28,466 report to the master. The workers will have 44 00:02:28,466 --> 00:02:32,799 a component called kubelet, a small 45 00:02:32,800 --> 00:02:36,866 process or a daemon which will continue to run in 46 00:02:36,866 --> 00:02:40,132 each worker machine and to manage all this 47 00:02:40,133 --> 00:02:43,466 worker or this kubelet I am going to have a 48 00:02:43,466 --> 00:02:48,066 controller within the master. There are n 49 00:02:48,066 --> 00:02:50,166 number of controllers available like node 50 00:02:50,166 --> 00:02:57,066 controller, replication controller, router, volume 51 00:02:57,066 --> 00:03:00,799 and many more. So this controller will be 52 00:03:00,800 --> 00:03:03,433 controlling how this particular node should 53 00:03:03,433 --> 00:03:05,366 behave, what are all the availability of the 54 00:03:05,366 --> 00:03:08,599 resources and what work currently it is doing 55 00:03:08,600 --> 00:03:10,600 all will be controlled by this particular 56 00:03:10,600 --> 00:03:14,600 controller and I'm going to start a pod within 57 00:03:14,600 --> 00:03:16,966 this worker machines and pod going to have 58 00:03:16,966 --> 00:03:20,166 a container. So I need a runtime to run the 59 00:03:20,166 --> 00:03:22,699 containers, so I need a container runtime 60 00:03:22,700 --> 00:03:26,600 in each worker machine. Container run time it 61 00:03:26,600 --> 00:03:35,333 could be of docker racket rkt or even containerd 62 00:03:35,333 --> 00:03:39,399 predominantly we will be using docker and 63 00:03:39,400 --> 00:03:41,933 depending on the type of runtime that 64 00:03:41,933 --> 00:03:43,799 corresponding containers would get created 65 00:03:43,800 --> 00:03:47,033 within the pod. So each worker machine will 66 00:03:47,033 --> 00:03:52,099 have a run time to start the containers and to 67 00:03:52,100 --> 00:03:54,633 interact with this particular cluster I need to 68 00:03:54,633 --> 00:03:59,533 have a component that is called API server. 69 00:03:59,533 --> 00:04:02,366 So the user will be interacting with this API 70 00:04:02,366 --> 00:04:06,299 server and provide the yaml file and say I 71 00:04:06,300 --> 00:04:09,133 wanted this type of infrastructure or this 72 00:04:09,133 --> 00:04:11,999 many number of pods to get created and 73 00:04:12,000 --> 00:04:14,800 depending on the availability of the resources 74 00:04:14,800 --> 00:04:17,966 the kube scheduler will schedule there is 75 00:04:17,966 --> 00:04:21,466 another component called kube scheduler 76 00:04:21,466 --> 00:04:23,666 where the pod should get created, what type 77 00:04:23,666 --> 00:04:25,599 of pod should get created, how it should get 78 00:04:25,600 --> 00:04:29,200 created. So correspondingly the pod would 79 00:04:29,200 --> 00:04:32,433 get created. So all these pods will be 80 00:04:32,433 --> 00:04:36,266 managed by kube scheduler and within the pod 81 00:04:36,266 --> 00:04:41,166 I am going to have the containers and I can 82 00:04:41,166 --> 00:04:43,499 consolidate and group all these containers 83 00:04:43,500 --> 00:04:47,200 together and I will be calling that as a service 84 00:04:47,200 --> 00:04:49,800 say I may deploy an application and that 85 00:04:49,800 --> 00:04:52,600 should have three web servers and a load 86 00:04:52,600 --> 00:04:55,633 balancer and all this will be consolidated and 87 00:04:55,633 --> 00:04:58,599 managed together as a service like this we 88 00:04:58,600 --> 00:05:00,700 have n number of components involved 89 00:05:00,700 --> 00:05:03,366 within kubernetes we will be seeing all of them 90 00:05:03,366 --> 00:05:06,399 one after the other and basically the core 91 00:05:06,400 --> 00:05:09,466 building block of kubernetes is a pod and the 92 00:05:09,466 --> 00:05:12,566 pod going to have n number of containers 93 00:05:12,566 --> 00:05:15,132 most of the time it is recommended to have 94 00:05:15,133 --> 00:05:17,799 just one containers but it can have other 95 00:05:17,800 --> 00:05:21,100 containers we call them as sidecar or the 96 00:05:21,100 --> 00:05:23,566 container that's going to work closely with 97 00:05:23,566 --> 00:05:27,499 the core container. Now I can go ahead and 98 00:05:27,500 --> 00:05:29,866 add n number of workers or remove the 99 00:05:29,866 --> 00:05:31,866 worker for any other maintenance purpose 100 00:05:31,866 --> 00:05:34,799 and do the configuration within the cluster all 101 00:05:34,800 --> 00:05:37,700 this needs to be stored and I'm going to use 102 00:05:37,700 --> 00:05:42,733 a database called etcd and this is a key 103 00:05:42,733 --> 00:05:45,199 value pair database and the configurations 104 00:05:45,200 --> 00:05:49,066 will be stored within the etcd database and 105 00:05:49,066 --> 00:05:51,199 for high availability I can go for multiple 106 00:05:51,200 --> 00:05:54,466 masters as well. Now we do have all the 107 00:05:54,466 --> 00:05:56,999 components put in a single place, now let's 108 00:05:57,000 --> 00:05:59,400 explore the components available as a part 109 00:05:59,400 --> 00:06:03,033 of the master as well as the workers. So 110 00:06:03,033 --> 00:06:05,199 within master I'm going to have API server 111 00:06:05,200 --> 00:06:08,566 that will be used for the users to interact with 112 00:06:08,566 --> 00:06:10,566 and I am going to have a kube scheduler to 113 00:06:10,566 --> 00:06:13,766 manage the pods and which node should do 114 00:06:13,766 --> 00:06:16,166 that particular work of the pod in 115 00:06:16,166 --> 00:06:20,966 which node the pod should get started and where it should not get started, how the load 116 00:06:20,966 --> 00:06:26,199 is within the nodes all will be managed by the kube scheduler and what are all the nodes 117 00:06:26,200 --> 00:06:29,066 available and how the replication can happen 118 00:06:29,066 --> 00:06:31,366 where it can happen all will be controlled by 119 00:06:31,366 --> 00:06:33,799 this controllers and there are n number of 120 00:06:33,800 --> 00:06:36,133 controllers available, as of now assume I do 121 00:06:36,133 --> 00:06:38,333 have a controller to do different type of 122 00:06:38,333 --> 00:06:42,399 controls like replications routing volumes and 123 00:06:42,400 --> 00:06:45,800 other activities. In the same way I do have a 124 00:06:45,800 --> 00:06:50,733 key value pair database called etcd database. The similar 125 00:06:50,733 --> 00:06:55,099 way if I get into the worker machines each 126 00:06:55,100 --> 00:06:57,100 worker machine it's going to have a 127 00:06:57,100 --> 00:06:59,933 component called kubelet and it's going to 128 00:06:59,933 --> 00:07:02,199 have a container runtime environment it 129 00:07:02,200 --> 00:07:05,633 could be of docker rkt or containerd 130 00:07:05,633 --> 00:07:09,133 Containerd is open sourced by docker and rkt 131 00:07:09,133 --> 00:07:11,433 is another type of container provided by red 132 00:07:11,433 --> 00:07:14,333 hat and I will be having another small 133 00:07:14,333 --> 00:07:16,933 component called kube proxy to take care of 134 00:07:16,933 --> 00:07:19,766 the networking. All the pods and the 135 00:07:19,766 --> 00:07:22,332 containers across the nodes needs to be 136 00:07:22,333 --> 00:07:24,333 communicated within each other and that 137 00:07:24,333 --> 00:07:28,133 will be facilitated by the kube proxy. So just to 138 00:07:28,133 --> 00:07:31,433 recall how the workflow will happen the user 139 00:07:31,433 --> 00:07:34,299 will be having a yaml file that is declarative 140 00:07:34,300 --> 00:07:37,533 way of defining how the services or the pod 141 00:07:37,533 --> 00:07:40,999 should exist and that will be provided to the 142 00:07:41,000 --> 00:07:43,400 API server and that is done using a 143 00:07:43,400 --> 00:07:46,833 component called kubectl and API server will 144 00:07:46,833 --> 00:07:49,199 process this yml file with the help of 145 00:07:49,200 --> 00:07:51,666 kube scheduler and the controller it will 146 00:07:51,666 --> 00:07:53,999 define which are all the worker machine 147 00:07:54,000 --> 00:07:56,566 should have the pods and how many pods 148 00:07:56,566 --> 00:07:58,932 and where it should get created. So the pods 149 00:07:58,933 --> 00:08:01,466 will be assigned, so that's going to have X 150 00:08:01,466 --> 00:08:03,932 amount of capacity of memory processing 151 00:08:03,933 --> 00:08:06,366 capacity within that the containers can get 152 00:08:06,366 --> 00:08:09,066 started. So the worker machines will be 153 00:08:09,066 --> 00:08:11,066 communicating with the registry or the 154 00:08:11,066 --> 00:08:14,132 repository to get the image of the container 155 00:08:14,133 --> 00:08:16,933 get the container started and the proxy 156 00:08:16,933 --> 00:08:19,233 available will facilitate the networking within 157 00:08:19,233 --> 00:08:21,999 the worker machines and the kubelet it's also 158 00:08:22,000 --> 00:08:25,100 called as an agent. So that's going to keep 159 00:08:25,100 --> 00:08:26,966 providing the report to the controller how 160 00:08:26,966 --> 00:08:29,099 much amount of resources available and 161 00:08:29,100 --> 00:08:31,300 where the future pods can be started that 162 00:08:31,300 --> 00:08:33,299 can be controlled and managed by the 163 00:08:33,299 --> 00:08:35,600 controller. So in a quick summary we had 164 00:08:35,600 --> 00:08:37,400 seen what are all the different type of 165 00:08:37,400 --> 00:08:40,265 components available within a kubernetes 166 00:08:40,265 --> 00:08:43,599 cluster and this is a high-level overview on 167 00:08:43,600 --> 00:08:46,200 how all this components going to come 168 00:08:46,200 --> 00:08:48,033 together we will be discussing each and 169 00:08:48,033 --> 00:08:52,899 every component in detail this is just a quick introduction on various 170 00:08:52,900 --> 00:08:58,066 terminologies and the components involved within kubernetes.