-
[Udemy] Deploymentskubernetes/udemy 2025. 2. 13. 10:28
Deployments
- ReplicaSet을 제어해주는 부모 역할
- ReplicaSet을 컨트롤해서 Pod 수 조절
- Deployment -> ReplicaSet -> Pod 순으로 컨트롤
- Deployment가 ReplicaSet을 Control하고, ReplicaSet이 Pod를 컨트롤
- 목적 : Pod 수 조절 & Rolling update & Rolling back
- Rolling update?
- 애플리케이션의 새로운 버전을 점진적으로 배포하는 방법으로, 서비스의 가용성을 유지하면서 업데이트를 수행할 수 있도록 돕습니다.
- kubectl set image deployment <deploy_name> <container_name> = <new_version_image>
- kubectl rollout history deployment <deploy_name> # 업데이트 히스토리 확인
- Rolling Back?
- 애플리케이션의 이전 버전으로 되돌리는 과정입니다. 이는 주로 새로운 버전의 배포 중에 문제가 발생했을 때 사용되며, 서비스의 가용성을 유지하면서 신속하게 안정적인 상태로 복구할 수 있도록 돕습니다.
- kubectl rollout undo deploy <deploy_name>
- Rolling update control
- kubectl rollout status deployment app-deploy # 업데이트 상태 확인
- kubectl rollout pause deployment app-deploy # 업데이트 중지
- kubectl rollout resume deployment app-deploy # 업데이트 재개
- Rolling update?
[따배쿠] Controller - Deployment
DeploymentReplicaSet을 제어해주는 부모 역할ReplicaSet을 컨트롤해서 Pod 수 조절Deployment -> ReplicaSet -> Pod 순으로 컨트롤Deployment가 ReplicaSet을 Control하고, ReplicaSet이 Pod를 컨트롤목적 : Pod 수 조절 & Rolling u
bbiyak-cloud.tistory.com
Deployment 생성
# Deployment 생성
# Create Deployment kubectl create deployment <Deplyment-Name> --image=<Container-Image> kubectl create deployment my-first-deployment --image=stacksimplify/kubenginx:1.0.0# Deployment 확인
# Verify Deployment kubectl get deployments kubectl get deploy --- NAME READY UP-TO-DATE AVAILABLE AGE my-first-deployment 1/1 1 1 3m22s# Deployment 자세히 확인
# Describe Deployment kubectl describe deployment <deployment-name> kubectl describe deployment my-first-deployment# ReplicaSet 확인
# Verify ReplicaSet kubectl get rs NAME DESIRED CURRENT READY AGE my-first-deployment-5bb6fc76f8 1 1 1 3m15s# Pod 확인
# Verify Pod kubectl get po --- NAME READY STATUS RESTARTS AGE my-first-deployment-5bb6fc76f8-6rdk5 1/1 Running 0 3m24sDeployment Scale up
# Deployment Scale up
# Scale Up the Deployment kubectl scale --replicas=10 deployment/<Deployment-Name> kubectl scale --replicas=10 deployment/my-first-deployment# Deployment 확인
# Verify Deployment kubectl get deploy --- NAME READY UP-TO-DATE AVAILABLE AGE my-first-deployment 10/10 10 10 15m# ReplicaSet 확인
# Verify ReplicaSet kubectl get rs --- NAME DESIRED CURRENT READY AGE my-first-deployment-5bb6fc76f8 10 10 10 15m# Pod 확인
# Verify Pods kubectl get po --- NAME READY STATUS RESTARTS AGE my-first-deployment-5bb6fc76f8-22pzl 1/1 Running 0 106s my-first-deployment-5bb6fc76f8-5vbhb 1/1 Running 0 106s my-first-deployment-5bb6fc76f8-6rdk5 1/1 Running 0 16m my-first-deployment-5bb6fc76f8-c5wg5 1/1 Running 0 12m my-first-deployment-5bb6fc76f8-gb9zh 1/1 Running 0 106s my-first-deployment-5bb6fc76f8-hn2vd 1/1 Running 0 106s my-first-deployment-5bb6fc76f8-l6266 1/1 Running 0 106s my-first-deployment-5bb6fc76f8-pmzwx 1/1 Running 0 106s my-first-deployment-5bb6fc76f8-v9ldr 1/1 Running 0 106s my-first-deployment-5bb6fc76f8-xxdfp 1/1 Running 0 106sDeployment Scale down
# Deployment Scale down
kubectl scale --replicas=2 deployment/my-first-deployment deployment.apps/my-first-deployment scaled# Deployment 확인
kubectl get deploy --- NAME READY UP-TO-DATE AVAILABLE AGE my-first-deployment 2/2 2 2 17m# ReplicaSet 확인
kubectl get rs --- NAME DESIRED CURRENT READY AGE my-first-deployment-5bb6fc76f8 2 2 2 17m# Pod 확인
kubectl get pods --- NAME READY STATUS RESTARTS AGE my-first-deployment-5bb6fc76f8-6rdk5 1/1 Running 0 17m my-first-deployment-5bb6fc76f8-c5wg5 1/1 Running 0 13mDeployment를 Service를 통해 노출
# Deployment를 Service를 통해 노출
# Expose Deployment as a Service kubectl expose deployment <Deployment-Name> --type=LoadBalancer --port=80 --target-port=80 --name=<Service-Name-To-Be-Created> kubectl expose deployment my-first-deployment --type=LoadBalancer --port=80 --target-port=80 --name=my-first-deployment-service# Service 확인
# Get Service Info kubectl get svc --- NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE kubernetes ClusterIP 10.0.0.1 <none> 443/TCP 20h my-first-deployment-service LoadBalancer 10.0.232.64 20.214.196.89 80:30778/TCP 12m# Service의 External IP를 통해, Pod 내 애플리케이션 접근
http://<External-IP-from-get-service-output>
Service를 통해 Deployment에 접근하고, 그 안에 있는 Pod 내 애플리케이션까지 접근하였다.
Deployment Rolling Update
# Rolling Update (컨테이너 이미지 업데이트)
# Update Deployment - SHOULD WORK NOW kubectl set image deployment/<Deployment-Name> <Container-Name>=<Container-Image> --record=true kubectl set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:2.0.0 --record=truemy-first-deployment 내 컨테이너 이미지를 1.0.0 버전에서 2.0.0으로 업그레이드
더보기** container name 확인 방법
kubectl get deploy <Deployment-Name> -o yaml
kubectl get deploy my-first-deployment -o yaml
---
...
spec:
containers:
- image: stacksimplify/kubenginx:2.0.0
imagePullPolicy: IfNotPresent
name: kubenginx
resources: {}
terminationMessagePath: /dev/termination-log
terminationMessagePolicy: File
dnsPolicy: ClusterFirst
restartPolicy: Always
schedulerName: default-scheduler
securityContext: {}
terminationGracePeriodSeconds: 30
...여기서는 container name이 kubenginx고, container image는 stacksimplify/kubenginx:2.0.0임을 확인 가능
# Rolling Update 상태 확인
kubectl rollout status deployment/my-first-deployment --- deployment "my-first-deployment" successfully rolled out# ReplicaSet 확인
# Verify ReplicaSet kubectl get rs --- NAME DESIRED CURRENT READY AGE my-first-deployment-5bb6fc76f8 0 0 0 44m my-first-deployment-c4bf784bc 2 2 2 17mRolling Update와 동시에 기존 ReplicaSet은 0으로 변하고 신규 ReplicaSet이 생김
# Pod 확인
# Verify pods kubectl get pods --- NAME READY STATUS RESTARTS AGE my-first-deployment-c4bf784bc-78vzk 1/1 Running 0 19m my-first-deployment-c4bf784bc-hnrwk 1/1 Running 0 19m다시 Pod도 생성됨
# Rolling Update 히스토리 확인
# Check the Rollout History of a Deployment kubectl rollout history deployment/<Deployment-Name> kubectl rollout history deployment/my-first-deployment --- REVISION CHANGE-CAUSE 1 <none> 2 kubectl.exe set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:2.0.0 --record=true# Service의 External IP를 통해, Pod 내 애플리케이션 접근
kubectl get svc --- NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE kubernetes ClusterIP 10.0.0.1 <none> 443/TCP 20h my-first-deployment-service LoadBalancer 10.0.232.64 20.214.196.89 80:30778/TCP 44mExternal IP로 다시 접근해보자
http://<External-IP-from-get-service-output>/
이미지 버전 2.0.0으로 정상적으로 바뀐 것을 확인 가능하다.
# edit 명령어를 통해 직접 파일을 수정하여 컨테이너를 3.0.0 이미지로 업데이트
# Edit Deployment kubectl edit deployment/<Deployment-Name> --record=true kubectl edit deployment/my-first-deployment --record=true해당 명령어를 수정하면 yml 파일이 열리는데,
이미지 버전을 3.0.0 버전으로 수정해주자
# Deployment 확인
kubectl get deploy --- NAME READY UP-TO-DATE AVAILABLE AGE my-first-deployment 2/2 2 2 59m# ReplicaSet 확인
kubectl get rs --- NAME DESIRED CURRENT READY AGE my-first-deployment-5bb6fc76f8 0 0 0 59m my-first-deployment-75df965dd 2 2 2 2m49s my-first-deployment-c4bf784bc 0 0 0 33m기존 ReplicaSet들은 0으로 변하고, 신규 ReplicaSet이 생성된 것을 확인
# Pod 확인
kubectl get pods --- NAME READY STATUS RESTARTS AGE my-first-deployment-75df965dd-2xcgq 1/1 Running 0 2m53s my-first-deployment-75df965dd-w9d6h 1/1 Running 0 3m새로운 Pod가 생성된 것을 확인
# Rolling Update 상태 & 히스토리 확인
kubectl rollout status deployment/my-first-deployment deployment "my-first-deployment" successfully rolled outkubectl rollout history deployment/my-first-deployment deployment.apps/my-first-deployment REVISION CHANGE-CAUSE 1 <none> 2 kubectl.exe set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:2.0.0 --record=true 3 kubectl.exe edit deployment/my-first-deployment --record=true# Service의 External IP로 Pod 내 애플리케이션에 접근
해당 컨테이너의 이미지가 3.0.0 버전으로 잘 변경된 것을 확인 가능하다.
Deployment Rolling back
# Rolling Update history 확인
# List Deployment Rollout History kubectl rollout history deployment/<Deployment-Name> kubectl rollout history deployment/my-first-deployment --- deployment.apps/my-first-deployment REVISION CHANGE-CAUSE 1 <none> 2 kubectl.exe set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:2.0.0 --record=true 3 kubectl.exe edit deployment/my-first-deployment --record=true# 특정 버전으로 Rollback을 해보자 (revision = 1)
kubectl rollout history deployment/my-first-deployment --revision=1 --- deployment.apps/my-first-deployment with revision #1 Pod Template: Labels: app=my-first-deployment pod-template-hash=5bb6fc76f8 Containers: kubenginx: Image: stacksimplify/kubenginx:1.0.0 Port: <none> Host Port: <none> Environment: <none> Mounts: <none> Volumes: <none>history의 revision=1로 rollback이 되었고,
하단에 Image: stacksimplify/kubenginx:1.0.0를 보면 어떤 이미지로 업데이트 되었는지 확인 가능하다.
# rollout undo 명령어로 Rollback
kubectl rollout history deployment/my-first-deployment --- deployment.apps/my-first-deployment REVISION CHANGE-CAUSE 1 <none> 4 kubectl.exe set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:2.0.0 --record=true 5 kubectl.exe edit deployment/my-first-deployment --record=truekubectl rollout undo deployment/my-first-deployment -- deployment.apps/my-first-deployment rolled backkubectl rollout history deployment/my-first-deployment --- deployment.apps/my-first-deployment REVISION CHANGE-CAUSE 1 <none> 5 kubectl.exe edit deployment/my-first-deployment --record=true 6 kubectl.exe set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:2.0.0 --record=true현재 2.0.0 버전으로 rollback 되었다.
# ReplicaSet 확인
kubectl get rs --- NAME DESIRED CURRENT READY AGE my-first-deployment-5bb6fc76f8 0 0 0 90m my-first-deployment-75df965dd 0 0 0 33m my-first-deployment-c4bf784bc 2 2 2 63m다시 2.0.0 버전의 컨테이너를 갖고 있는 ReplicaSet이 활성화되었다.
# Service의 External IP로 접근
Deployment Rollout Restart
# Deployment 재시작
# Rolling Restarts kubectl rollout restart deployment/<Deployment-Name> kubectl rollout restart deployment/my-first-deploymentKubernetes에서 특정 배포(Deployment)를 재시작하는 데 사용됩니다.
이 명령은 배포의 파드를 종료하고 새로운 파드를 생성하여 애플리케이션을 재시작하는 역할을 합니다.
# Pod 확인
kubectl get pods --- NAME READY STATUS RESTARTS AGE my-first-deployment-6c86bcddf6-8zn2f 1/1 Running 0 21s my-first-deployment-6c86bcddf6-jj4b4 1/1 Running 0 20sDeployment Rollout Pause & Resume
# Deployment의 Rollout pause
# Pause the Deployment kubectl rollout pause deployment/<Deployment-Name> kubectl rollout pause deployment/my-first-deploymentKubernetes에서 특정 배포(Deployment)의 롤아웃을 일시 중지하는 데 사용됩니다.
이 명령은 배포의 업데이트 프로세스를 중단하여, 현재 진행 중인 변경 사항이 적용되지 않도록 합니다.
# Rolling Update (컨테이너 이미지 업데이트)
# Update Deployment - Application Version from V3 to V4 kubectl set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:4.0.0 --record=true컨테이너의 이미지를 3.0.0에서 4.0.0으로 변경해보자
명령을 실행해도 지금은 rollout이 pause 상태이기 때문에, 즉시 변경되지 않는다.
kubectl rollout history deployment/my-first-deployment --- deployment.apps/my-first-deployment REVISION CHANGE-CAUSE 1 <none> 6 kubectl.exe set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:2.0.0 --record=true 7 kubectl.exe edit deployment/my-first-deployment --record=true 8 kubectl.exe edit deployment/my-first-deployment --record=true# Get list of ReplicaSets kubectl get rs --- NAME DESIRED CURRENT READY AGE my-first-deployment-5bb6fc76f8 0 0 0 3h40m my-first-deployment-6c86bcddf6 2 2 2 10m my-first-deployment-75df965dd 0 0 0 163m my-first-deployment-c4bf784bc 0 0 0 3h13m# 컨테이너의 리소스 사용량 변경
kubectl set resources deployment/my-first-deployment -c=kubenginx --limits=cpu=20m,memory=30Mi테스트를 위해, 컨테이너의 리소스 사용량도 변경해보자
해당 작업도 즉시 변경되지 않는다. rollout pause 상테이기 때문이다.
# Deployment의 Rollout Resume
kubectl rollout resume deployment/my-first-deployment --- deployment.apps/my-first-deployment resumedRollout을 재개해보자
# Deployment의 Rollout Resume 확인
kubectl rollout history deployment/my-first-deployment --- deployment.apps/my-first-deployment REVISION CHANGE-CAUSE 1 <none> 6 kubectl.exe set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:2.0.0 --record=true 7 kubectl.exe edit deployment/my-first-deployment --record=true 8 kubectl.exe edit deployment/my-first-deployment --record=true 9 kubectl.exe set image deployment/my-first-deployment kubenginx=stacksimplify/kubenginx:4.0.0 --record=true재개하자마자 아까 실행했던 4.0.0 버전 이미지 업데이트가 진행되었다.
kubectl get rs --- NAME DESIRED CURRENT READY AGE my-first-deployment-5bb6fc76f8 0 0 0 3h48m my-first-deployment-6c86bcddf6 0 0 0 19m my-first-deployment-75df965dd 0 0 0 171m my-first-deployment-8dfd9f847 2 2 2 85s my-first-deployment-c4bf784bc 0 0 0 3h21m재개하자마자 replicaSet도 추가된 것을 볼 수 있고
kubectl get pods --- NAME READY STATUS RESTARTS AGE my-first-deployment-8dfd9f847-hpvmn 1/1 Running 0 78s my-first-deployment-8dfd9f847-mlb68 1/1 Running 0 89sPod도 이미지에 맞게 다시 생성되었다.
# Service의 External IP로 접근
Pod 내 애플리케이션 버전이 V4로 정상적으로 바뀐 것을 확인 가능하다.
Delete
# Deployment 삭제
# Delete Deployment kubectl delete deployment my-first-deployment# Service 삭제
# Delete Service kubectl delete svc my-first-deployment-service
[참고 영상]
Udemy - Azure Kubernetes Service with Azure DevOps and Terraform
섹션 4 : Kubernetes Fundamentals with kubectl - Imperative Approach
29. Step-12 : Introduction ot Kubernetes Deployments
30. Step-13 : Create Deployment, Expose with a Service, Scale Up and Down Replicas
31. Step-14: Understand how to Update Deployments in Kubernetes
32. Step-15: Understand how to rollback deployments in Kubernetes
33. Step-16: Understand how to pause and resume deployments in kubernetes
[참고 문서]
'kubernetes > udemy' 카테고리의 다른 글
[Udemy] Services Demo with Cluster IP and Load Balancer Services (Declarative) (0) 2025.02.14 [Udemy] Services Demo with Cluster IP and Load Balancer Services (Imperative) (0) 2025.02.13 [Udemy] ReplicaSet (0) 2025.02.12 [Udemy] Interact with pods, logs, connect to pod and cleanup (0) 2025.02.12 [Udemy] Service (0) 2025.02.12