Debug School: Yogeshwari

Assignment #3 Day3 Learning Sumary

Yogeshwari — Wed, 18 Oct 2023 12:11:21 +0000

Ways to deploy the POD

POD
Replication Controller
Replica Set
Deployment
DaemonSet
StatefulSet
DaemonSet:
Make sure only 1 pod will be running on each node in the cluster
Job & CronJob:
Job: Create the pod, do the job and complete the activity
Cron job: Create the pod, do the job and complete the activity and repeats the same periodically based on the cron rules
ConfigMap:
Helps to store the data in the cluster.
Service:
Help for loadbalancing
Types:
Network Load balancer
Application Load balancer
Pods where the labels are matching with the labels of the selector of the service will be loadbalanced.They load balance only to the healthy pods.
They can be configured for session affinity.
Random Load Balancing is the algorithm used in the service
They use TCP by default
Types of service:
Cluster IP
Externalname
Loadbalancer
NodePort
Kubernetes Ingress:
They are the application load balancer. Ingress help to access the application in kubernetes.
They do
Enabling traffic
Traffic routing
Ensure traffic is reliable
Components of ingress:
Rule -> Ingress object definition
Controller -> Ingress controller deployment

Assignment Day#2 Learning Summary

Yogeshwari — Tue, 17 Oct 2023 12:07:47 +0000

Namespace:
Namespace is the logical separation of a cluster. It helps to isolate the pods. Scope of the pod is namespace. "default" is the default namespace in the cluster. All the pods will be created in default namespace when namespace is not specified.
To create the namespace:
kubectl create ns
Below are two ways to perform CRUD (Create,Read,Update and Delete) operations of the resource in cluster:
Create:

YAML - create -f yaml
Command - create Read:
YAML - get-f yaml
Command - get, describe Update:
YAML - apply -f yaml
Command - edit Delete:
YAML - delete -f yaml
Command - delete

POD
It is the basic unit of the cluster
Create the Pod:
kubectl create -y pod.yaml -nyogi
Display the pods:
kubectl get pods -nyogi
kubectl get pods -o wide -nyogi
To see the container logs:
kubectl logs yogi-pod -nyogi
Portforwarding:
kubectl port-forward --address 0.0.0.0 pod/yogi-pod -nyogi 8999:80
Replication Controller and ReplicaSet
Controls the number of pods to be running in the cluster
Replication controller is now deprecated and replaced by ReplicaSet
Deployment

Replication
Controller
Versioning
Rollout
Rollback Below are the deployment strategy followed in Kubernetes: Recreate RollingUpdate RollingUpdate is the default deployment strategy in kubernetes.

Assignment #2 What is Pod?

Yogeshwari — Mon, 16 Oct 2023 11:47:41 +0000

Pod is the atomic unit of scheduling. We can have one or more containers in the pod. It is the most basic unit of work. They are ephemeral meaning they can't be redeployed. Below is the lifecycle of a Pod:

Pending
Running
Succeeded \ Failed

Kubernetes job is to keep the pods in running state with the desired state.

Assignment #1

Yogeshwari — Mon, 16 Oct 2023 07:18:46 +0000

1)What is kubernetes?
Kubernetes is a container orchestration tool. It supports high availability, workload placements, Infrastructure abstraction, maintains desired state. .
2) why do you need kubernetes?
Scaling is the main limitation for a container.
Kubernetes supports high availability and scalability.Docker can't manage multiple nodes. To overcome this, we have kubernetes. It does self healing, automated Rollbacks, Storage orchestration, secret and configuration management etc
3) Explain in 10 lines how kubernetes works?
It manages the clusters.
We have one or multiple master nodes.
We have a worker node.
One or more pods will be running in the worker nodes. Each pod may contain one or more containers. All pods will be communicating to each other and to master node as well. Master node controls all the worker nodes.
4) kubernetes architecture

5) Explain each component one line
Components in master node:
1) API server - controls and manages the cluster
2) ETCD - It is the cluster storage
3) Kube controller manager - Controls one or more controllers. Eg Node controller etc
4) Kube Scheduler - Responsible for scheduling the work to the containers.
Components in worker node:
1) Kubelet - Communicates to master node and get the instructions
2) Container runtime - Container engine eg. Docker, ContainerD etc
3) kube proxy - for configuring the network in the cluster