Devops-05-Oct-2024

Question

###To create environment variables in pod manifest files: +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ syntax in spec block is: spec: containers: - name: image: env: - name: value: ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ # create a pod with below configurations. - pod name ==> tomcatpod3 , image tomee & container tomeecontainer3 - labels ==> type: webapptool & author: - environment variables==> sports football vi pod-definition2.yml --- apiVersion: v1 kind: Pod metadata: name: tomcatpod3 labels: type: webapptool spec: containers: - name: tomeecontainer3 image: tomee env: - name: sports value: football ... kubectl apply -f pod-definition2.yml to see env variables details in k8s master node use ==> kubectl describe pod (observe env variables defined here) to see env variables details inside the pod/container use ==> login into pod kubectl exec -it -- /bin/bash echo $cricketer (from output - observe variable substitutions) echo $password (from output - observe variable substitutions) ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Port Mapping: Syntax / Template to write pod manifest file +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ apiVersion: v1 kind: Pod metadata: name: labels: : spec: # containers: - name: image: ports: - containerPort: hostPort: ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ apiVersion: v1 kind: Pod metadata: name: apachetomcatpod spec: containers: - name: tomcat-container-v2 image: tomee # Replace with your desired Tomcat image ports: - containerPort: 8080 hostPort: 30010 ============================================================================= Commands recap: --------------- kubectl create -f .yaml --> To create a object with a definition file kubectl get pods --> To display all the pods on the cluster kubectl get pods -o wide --> To display more information of pods kubectl describe --> To see the information about a particular object kubectl delete --> To delete a k8s object ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Note: pod is smallest object which we can create in k8s, simillar to pods kubernetes also have other objects like replicasets, replication-controller , deployments , services etc... ReplicaSet: ---------- is a higher level object, using which we can run multiple instances (multiple pods) of our applications. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ***** important interview question ***** what are labels in k8s? - labels are user provided key-value pairs, using labels we can organize(group) Kubernetes objects - using labels we can filter different Kubernetes objects - Same label (key/value) can be assigned to multiple Kubernetes objects ex: create a pod name as my-pod-2, container name as c2 use image nginx, with labels as environment: production & app: nginx observation on filtering pods using labels: if we have multiple pods in our cluster & if we want to list only Pods which have label as ==> app:nginx Kubectl get pods -l app=nginx ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Controller- manager: -------------------- controller manager is the brain behind the orchestration, it always monitor actual state of objects with that of desired states. Controller manager ==> ( actual state == desired state) controller manager objects includes following 1. Replication-controller --> older feature (not used now a days) 2. Replicasets --> newer feature 3. Deployments. 4. Deamonsets. ReplicaSet: ----------- if we want to run multiple number of pods then we can use replicasets. Pod is the smallest kubernetes object, which we worked on. Next Level is replicasets. - ReplicaSets is used for creating multiple replicas(numbers) of a specific pod very easily. - ReplicaSets ensure that "NUMBER OF PODS (Replicas)" specified in manifest file always exists. - using ReplicaSets we can achieve high availablity, load balancing and scaling. - if a pod get stopped / deleted then Replicasets will automatically recreate the new pod with same configuration ( Auto-healing). Note: - ReplicaSet uses elements like "replicas","selector" & "template" field in its 'spec' section. - Template contains all pod related information so it can be also called as pod template - copy pod defintion file contents ==> under "template" ==> without apiversion & kind SYNTAX of ReplicaSet manifest file: ------------------------------------ apiVersion: apps/v1 kind: ReplicaSets metadata: name: labels: : spec: replicas: selector: matchLabels: : template: # POD Template metadata: name: labels: : spec: - containers: - name: image: ... ex1: create a replicaset object with 5 replicas of tomcat pod apiVersion: apps/v1 kind: ReplicaSet metadata: name: my-tomee-replicaset labels: author: bharath spec: replicas: 2 selector: matchLabels: author: bharath template: metadata: name: tomcat-pod-v2 labels: author: bharath spec: containers: - name: tomcat-container-v2 image: tomee ... create replicaset with name as my-second-rs, image nginx:1.24.0-alpine --- apiVersion: apps/v1 kind: ReplicaSet metadata: name: my-second-rs labels: creator: mohan spec: replicas: 3 selector: matchLabels: creator: mohan template: #paste the podmanifest file without apiversion & kind metadata: name: mypodx labels: creator: mohan spec: containers: - name: c1 image: nginx:1.24.0-alpine ... kubectl create -f .yaml --> To create a rs object with a definition file kubectl get pods ( observe all pods created started by replicasets) to list replicasets: kubectl get replicasets (or) kubectl get rs to delete replicasets: kubectl delete rs Note: delete few pods created using replicasets & observe that new pods getting created automtically by controller manager component via replicasets ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Assignment: 1. create a pod manifest file with below configurations. - pod name ==> jenkinspod , image jenkins/jenkins & container jenkinscontainer - labels ==> use: cicdtool - environment variables==> stage1 build, stage2 test, stage3 deploy 2. create a pod manifest file - pod name ==> apachetomcatpod. - name the container as tomcat-container-v2. - also map container port 8080 to host machine on 4040 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Deployment Object: ------------------ This is also an high level k8s object , which can be used for running multiple replicas of a pod with other features like scalling, load balancing and perform rolling updates. Deployment makes sure that desired number of pods (replicas) specified in the manifest file, are always up and running. If a pod fails to run, deployment will remove that pod and replace it with a new one. deployments comes with advanced features like: - update application to newer versions without downtime. - rollback to older deployment versions. - scale deployment up or down. How does a Deployment ensure high availability? ---------------------------------------------- deployment ==creates==> replicasets ==creates==> pods ==creates==> Containers( our application runs inside the container) Deployments maintain high availability by managing replica sets. If a Pod fails due to any reason, the Deployment removes the failed pod with new pod & maintains the desired state ensuring high availablity. deployment startegies used in kubernetes? ----------------------------------------- i. Recreate strategy: - deleting all pods of old version at once & creating new pods with new version. - this startegy we can observe application downtime. ii. Rolling update strategy: - it will gradually delete 1 old version of a pod & bring 1 new version of a pod till all old pods gets replaced - in this startegy we will not see any applocitaion downtime. - this is the deafult startegy used in k8s Syntax / Template to write deployment manifest file --------------------------------------------- apiVersion: apps/v1 kind: Deployment metadata: name: labels: : spec: replicas: selector: matchLabels: : template: # POD Template metadata: name: labels: : spec: - containers: - name: image: ... create a deployment-defintion file with nginx:1.7.9 image with 4 replicas: -------------------------------------------------------------------------- apiVersion: apps/v1 kind: Deployment metadata: name: my-nginx-deployment labels: author: bharath spec: replicas: 4 selector: matchLabels: author: bharath template: # here paste content of pod manifest file without apiversion & kind metadata: name: nginx-pod labels: author: bharath spec: containers: - name: container2 image: nginx:1.7 kubectl create -f .yaml --> To create a deployment object with a definition file kubectl get pods ( observe all pods created started by deployment) to list deployment: kubectl get deployments (or) kubectl get deploy to delete deployment: kubectl delete deployment to know complete details about deployments: kubectl describe deployment it will show details about number of replicas used, stategy type, image used etc... Note: ----- 1. Auto-Healing feature With Deployment Controller: --------------------------------------------------- whenever we delete a pod which is running from any controller objects (i.e rs / deployment) new pod will get created this feature can be called as Auto-healing Deleting one the pods manually in (i.e rs / deployment) and observe the auto-healing behaviour of deployment *** important interview question *** 2. what all Deployment strategies used in kuberenets: ----------------------------------------------------- i. Recreate strategy: deleting all pods of old version at once & creating new pods with new version , this startegy we can observe application downtime. ii. Rolling update strategy: - it will delete 1 old version of a pod & bring 1 new version of a pod - this startegy we will not see any downtime. - this is the deafult startegy used in k8s 3. How to scale the deployment in k8s (scale up --> increasing the replicas of a pod, scale down --> decreasing the replicas of a pod): ------------------------------------------------------------------------------------------------------------------------------------- kubectl scale deployment --replicas if traffic coming to our applications increases, then we need to increase the number of replicas & when traffic reduces we can descrease the number of replicas SERVICES Agenda: ---------- - Deployment objects -- increase replicas for a deployment -- rolling update deployments -- rollback a deployment ++++++++++++++++++++++++++++++++++++++++++++++++ replica:10 strategy: rollingUpdate: maxSurge: 25% maxUnavailable: 30% type: RollingUpdate -------------------------------------------------- maxSurge: 1.Controls the maximum additional pods during a rolling update. 2.Specifies the number or percentage above the desired replica count. 3.Prevents the total pod count from exceeding a specified limit. maxUnavailable: 1.Determines the maximum unavailable pods during a rolling update. 2.Specifies the maximum number or percentage simultaneously removed from service. 3.Default behavior terminates one pod at a time while creating new pods to maintain the desired replica count. ++++++++++++++++++++++++++++++++++++++++++++++++++++ if we are updating our application to next version means, we are create new docker image, we use the updated docker image in our manifest files nginx:1.7 ==> nginx:1.8 ==> nginx:1.9 ==> nginx:1.10 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Deployment ( continued......) ----------------------------- create deployment with image ==> nginx:1.7 & 3 replicas apiVersion: apps/v1 kind: Deployment metadata: name: nginx-deployment labels: author: bharath environments: staging spec: replicas: 3 selector: matchLabels: author: bharath environments: staging template: metadata: name: nginx-deployment-v1 labels: author: bharath environments: staging spec: containers: - name: my-nginx-container image: nginx:1.7 ports: - containerPort: 80 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ RollingUpdate startegy is default startegy type used in kubernetes Deployment based Practical scenarios: case 1: deploy a new application version (new image ==> nginx:1.8) using rolling deployment startegy? --------------------------------------------------------------------------------------- developers ==> nginx:1.7 ==> nginx:1.8 ==> nginx:1.9 if we update from version to another version of my application, only image name & tag (version) will be updated in deployment manifest file. syntax: kubectl set image deployment = kubectl set image deployment nginx-deployment my-nginx-container=nginx:1.8 --record=true #To check the revision hitory kubectl rollout history deployment kubectl rollout history deployment nginx-deployment case 2: deploy a new application version again (from nginx:1.8 to new image ==> nginx:1.9) using rolling deployment startegy? --------------------------------------------------------------------------------------- kubectl set image deployment nginx-deployment my-nginx-container=nginx:1.9 --record=true case 3: code deplyed in Version 1.9 has issues, we need to rollback to previous version (from latest version nginx:1.9 to old version ==> nginx:1.8) --------------------------------------------------------------------------------------- ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Kubectl rollout options: ======================= 1.How to check a deployments history ? ------------------------------------------- kubectl rollout history deployment kubectl rollout history deployment nginx-deployment 2. How to rollback deployment to previous version? ----------------------------------------------- kubectl rollout undo deployment kubectl rollout undo deployment nginx-deployment (or) < to rollback to specific version > kubectl rollout undo deployment --to-revision= kubectl rollout undo deployment nginx-deployment --to-revision=1 3. How to Check the status of the rollout using status command ? ------------------------------------------------------------- kubectl rollout status deployment ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Note: ---- How to expose GUI based application in docker In Dockerfile ==> EXPOSE 8080 docker run --name c1 -p : ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Service Object: --------------- service objects is used to expose our application (port mapping) service object will provide a static ip address to our application running in pods. Service Object we use 3 ports in service manifest file, which are 1. Target port - Its is port on container 2. port - Refers to service reference port. 3. hostPort - Refers to host machine port to make it accessable from external network. *** important interview question *** What are different types of Service & difference between those ?: ---------------------------------------------------------------- 1. clusterIP: - It exposes the service within the Kubernetes cluster only. - ClusterIP is used when we want the pods in the cluster to communicate with each other and not with external network(from internet or browser). - This is default type of service object used in kubermetes. * - using clusterip type service pod-to-pod communication will happen within the cluster 2. nodePort: - It exposes the service both in and outside the cluster - It exposes the service on each Worker Node’s IP at a static port (i.e., which is called NodePort). - nodePort can be used, if we want to access the pods from an external network (Internet or broswer). - NodePort must be within the range from 30000-32767 3. LoadBalancer: - It exposes the service both in and outside the cluster, its the most preffered way of exposing a service in k8s. - It exposes the service externally using cloud providers load balancer ( AWS - ELB-> Elastic load balaNcers). - whenever the LoadBalancer service gets created it will also automtically create NodePort and ClusterIP services. how does pods-to-pod communication happens in kubernetes cluster? ---------------------------------------------------------------- using services (type-->clusterip) pod-to-pod communication will happen within the cluster +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Ex: create a deployment using jenkins image with 1 replicas & label environment: staging apiVersion: apps/v1 kind: Deployment metadata: name: jenkins-deployment labels: environment: staging spec: replicas: 1 selector: matchLabels: environment: staging template: metadata: labels: environment: staging spec: containers: - name: jenkins-container image: jenkins/jenkins Expose the jenkins pods created as a part of above deployment ? create a servce oject of type nodePort & expose the application vi my-jenkins-service.yaml --- apiVersion: v1 kind: Service metadata: name: my-jenkins-service spec: type: NodePort ports: - targetPort: 8080 #this is container port ==> jenkins container port: 8080 #this is service objects refernce port nodePort: 30001 #this is port assigning to worker node selector: environment: staging Important Note on service: -------------------------- service object will never creates pods. service object searches for all pods which have labels mentioned under selector keyword of service manifest file & expose those pods. in above example it will look for pods which have label as environment: staging, as jenkins pods have those labels (environment: staging) , it will expose all the jenkins pods on nodeport 30001. How to access application from gui? ----------------------------------- http://: http://3.90.88.139:30001 ********************************************************************************************************** deploying zomato-like application as k8s deployment ********************************************************************************************************** NotE: before starting this assignment delete all pods /deployments present in your cluster, as this application needs more resources # create deployment named -- zomato deployment, replicas 1 , image ==> acecloudacademy/zomato-app-image:latest --- apiVersion: apps/v1 kind: Deployment metadata: name: zomato-deployment labels: type: food-delivery-app spec: replicas: 1 selector: matchLabels: type: food-delivery-app template: metadata: name: zomato-pod labels: type: food-delivery-app spec: containers: - name: my-zomato-container image: acecloudacademy/zomato-app-image:latest ... create a service of type nodeport to expose zomato application apiVersion: v1 kind: Service metadata: name: my-zomato-service spec: type: NodePort ports: - targetPort: 3000 #this is container port ==> zomato listening port on container port: 3000 #this is service objects refernce port nodePort: 30003 #this is node port assigning to all worker node selector: type: food-delivery-app create service, once service matches pods & pods get exposed, access from GUI for zomato application Assignment: ---------- - Create a deployment of acecloudacademy webapplication image=acecloudacademy/myelevendevopsimage replicas=3 acecloudacademy webapplication runs on port 8080 expose this application to internet @ port 30008 - Create a deplyment of netflix app image=acecloudacademy/netflix-clone-app:v1 replicas=2 netflix application runs on port 80 (container port) expose this application to internet @ port 31111

Forum

Devops-05-Oct-2024