0801[k8s version up, eks ingress]

망지·2022년 8월 1일
0
post-custom-banner

📌 쿠버네티스 버전업

📙 저번에 만들어둔 centos7 복제(연결된 복제)해서 worker1,2만들기

✔️ centos 7 선택 - 복제 - 이름 : worker1 - 연결된 복제 -> 이름 worker2로 한번 더 진행
✔️ centos7이름 master1으로 바꿔주기
✔️ worker 1,2 : CPU1,RAM 1G로 스펙 바꿔주기 ( 설정 - 시스템)
✔️ 3개의 서버 그룹으로 묶어주고 이름 Kubernetes Cluster

📙 노드공통 명령어

# hostnamectl set-hostname master
# exit

# cat <<EOF >> /etc/hosts
192.168.56.106 master1
192.168.56.107 worker1
192.168.56.108 worker2
EOF

📙 노드 명령어

--master--
# kubeadm init --apiserver-advertise-address=192.168.56.106 --pod-network-cidr=10.244.0.0/16
# mkdir -p $HOME/.kube
# cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
# chown $(id -u):$(id -g) $HOME/.kube/config
# kubectl apply -f https://raw.githubusercontent.com/flannel-io/flannel/master/Documentation/kube-flannel.yml

-- worker 1,2에서 kubeadem init에서 맨 하단에 나온 kubeadm join (토큰정보) 입력 --
# kubeadm join 192.168.56.106:6443 --token 32eyw8.5l08iqkzvrg3q5i0 --discovery-token-ca-cert-hash sha256:5a1f6308637da409dc02418fc208c34caf343d52d857894ab13c7920543deb93

--다시 master로 돌아와서 확인하고 자동완성 명령어진행 --
[root@master ~]# kubectl get node
NAME      STATUS   ROLES    AGE   VERSION
master    Ready    master   6m    v1.19.16
worker1   Ready    <none>   71s   v1.19.16
worker2   Ready    <none>   44s   v1.19.16

# kubectl get pods --all-namespaces
# source <(kubectl completion bash) ## 자동완성 기능
# echo "source <(kubectl completion bash)" >> ~/.bashrc
# exit

세 노드 모두 poweroff 후 스냅샷 찍기 이름 : 1.19

📙 ingress / 경로기반라우팅 (AWS - EKS,ELB(ALB))

# yum install git -y
# git clone https://github.com/hali-linux/_Book_k8sInfra.git
# kubectl apply -f /root/_Book_k8sInfra/ch3/3.3.2/ingress-nginx.yaml ##controller 사용하기 위한 apply
# kubectl get pods -n ingress-nginx
# mkdir ingress && cd $_
# vi ingress-deploy.yaml
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: foods-deploy
spec:
  replicas: 1
  selector:
    matchLabels:
      app: foods-deploy
  template:
    metadata:
      labels:
        app: foods-deploy
    spec:
      containers:
      - name: foods-deploy
        image: halilinux/test-home:v1.0
---
apiVersion: v1
kind: Service
metadata:
  name: foods-svc
spec:
  type: ClusterIP
  selector:
    app: foods-deploy
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: sales-deploy
spec:
  replicas: 1
  selector:
    matchLabels:
      app: sales-deploy
  template:
    metadata:
      labels:
        app: sales-deploy
    spec:
      containers:
      - name: sales-deploy
        image: halilinux/test-home:v2.0
---
apiVersion: v1
kind: Service
metadata:
  name: sales-svc
spec:
  type: ClusterIP
  selector:
    app: sales-deploy
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: home-deploy
spec:
  replicas: 1
  selector:
    matchLabels:
      app: home-deploy
  template:
    metadata:
      labels:
        app: home-deploy
    spec:
      containers:
      - name: home-deploy
        image: halilinux/test-home:v0.0
---
apiVersion: v1
kind: Service
metadata:
  name: home-svc
spec:
  type: ClusterIP
  selector:
    app: home-deploy
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80

# kubectl apply -f ingress-deploy.yaml
# kubectl get all
# kubectl get pod -o wide
# kubectl get svc
# curl 10.98.225.168 ##다른 노드들에서도 접속 가능 
# kubectl get node -o wide


# vi ingress-config.yaml
apiVersion: networking.k8s.io/v1beta1
kind: Ingress
metadata:
  name: ingress-nginx
  annotations:
    nginx.ingress.kubernetes.io/rewrite-target: /
spec:
  rules:
  - http:
      paths:
      - path: /foods
        backend:
          serviceName: foods-svc
          servicePort: 80
      - path: /sales
        backend:
          serviceName: sales-svc
          servicePort: 80
      - path:
        backend:
          serviceName: home-svc
          servicePort: 80

# kubectl apply -f ingress-config.yaml
# kubectl get ingress
# kubectl describe ingress ingress-nginx

# vi ingress-service.yaml
apiVersion: v1
kind: Service
metadata:
  name: nginx-ingress-controller
  namespace: ingress-nginx
spec:
  ports:
  - name: http
    protocol: TCP
    port: 80
    targetPort: 80
  - name: https
    protocol: TCP
    port: 443
    targetPort: 443
  selector:
    app.kubernetes.io/name: ingress-nginx
  type: LoadBalancer
  externalIPs:
  - 192.168.56.106 ##master IP

# kubectl apply -f ingress-service.yaml

📙 taint와 toleration

✔️taint

# kubectl taint node worker1 tier=dev:NoSchedule
# kubectl describe nodes worker1
Taints:             tier=dev:NoSchedule

# vi ingress-deploy.yaml
replicas를 다 2개로 (3개의 디플로이 모두)

[root@master ingress]# kubectl get pod -o wide
NAME                            READY   STATUS    RESTARTS   AGE   IP           NODE      NOMINATED NODE   READINESS GATES
foods-deploy-7ffcb8f58-d8b6w    1/1     Running   0          13s   10.244.2.6   worker2   <none>           <none>
foods-deploy-7ffcb8f58-lrpvd    1/1     Running   0          16m   10.244.2.5   worker2   <none>           <none>
home-deploy-688558dc79-48xcs    1/1     Running   0          13s   10.244.2.8   worker2   <none>           <none>
home-deploy-688558dc79-7qhhk    1/1     Running   0          68m   10.244.1.3   worker1   <none>           <none>
sales-deploy-7cdbd9848c-4rqdv   1/1     Running   0          68m   10.244.2.3   worker2   <none>           <none>
sales-deploy-7cdbd9848c-tpdch   1/1     Running   0          13s   10.244.2.7   worker2   <none>           <none>

-> worker2에만 생성된 모습

✔️toleration

# vi pod-taint.yaml
apiVersion: v1
kind: Pod
metadata:
  name: pod-taint-metadata
  labels:
    app: pod-taint-labels
spec:
  containers:
  - name: pod-taint-containers
    image: nginx
  tolerations:
  - key: "tier"
    operator: "Equal"
    value: "dev"
    effect: "NoSchedule"
---
apiVersion: v1
kind: Service
metadata:
  name: pod-taint-service
spec:
  type: NodePort
  selector:
    app: pod-taint-labels
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80
    
 [root@master ingress]# kubectl apply -f pod-taint.yaml
pod/pod-taint-metadata created
service/pod-taint-service created
[root@master ingress]# kubectl get po -o wide
NAME                            READY   STATUS    RESTARTS   AGE     IP           NODE      NOMINATED NODE   READINESS GATES
foods-deploy-7ffcb8f58-d8b6w    1/1     Running   0          9m45s   10.244.2.6   worker2   <none>           <none>
foods-deploy-7ffcb8f58-lrpvd    1/1     Running   0          26m     10.244.2.5   worker2   <none>           <none>
home-deploy-688558dc79-48xcs    1/1     Running   0          9m45s   10.244.2.8   worker2   <none>           <none>
home-deploy-688558dc79-7qhhk    1/1     Running   0          77m     10.244.1.3   worker1   <none>           <none>
pod-taint-metadata              1/1     Running   0          22s     10.244.1.5   worker1   <none>           <none>
sales-deploy-7cdbd9848c-4rqdv   1/1     Running   0          77m     10.244.2.3   worker2   <none>           <none>
sales-deploy-7cdbd9848c-tpdch   1/1     Running   0          9m45s   10.244.2.7   worker2   <none>           <none>

->worker1에 생성됨

  158  kubectl delete -f pod-taint.yaml
[root@master ingress]# kubectl taint node worker2 tier=dev:NoSchedule
[root@master ingress]# kubectl apply -f pod-taint.yaml
[root@master ingress]# kubectl get pod -o wide
NAME                            READY   STATUS    RESTARTS   AGE    IP           NODE      NOMINATED NODE   READINESS GATES
foods-deploy-7ffcb8f58-d8b6w    1/1     Running   0          96m    10.244.2.6   worker2   <none>           <none>
foods-deploy-7ffcb8f58-lrpvd    1/1     Running   0          112m   10.244.2.5   worker2   <none>           <none>
home-deploy-688558dc79-48xcs    1/1     Running   0          96m    10.244.2.8   worker2   <none>           <none>
home-deploy-688558dc79-7qhhk    1/1     Running   0          164m   10.244.1.3   worker1   <none>           <none>
pod-taint-metadata              1/1     Running   0          97s    10.244.1.6   worker1   <none>           <none>
sales-deploy-7cdbd9848c-4rqdv   1/1     Running   0          164m   10.244.2.3   worker2   <none>           <none>
sales-deploy-7cdbd9848c-tpdch   1/1     Running   0          96m    10.244.2.7   worker2   <none>           <none>

worker2에도 테인드, 1or2중에 생성.
또다시 1에 생성된 모습 (2에 할당된 pod가 많아서 그런듯)

📙 클러스터 업그레이드 1

- master 노드
# cd ~
# yum list --showduplicates kubeadm --disableexcludes=kubernetes
# yum install -y kubeadm-1.20.15-0 --disableexcludes=kubernetes
# kubeadm version
# kubeadm upgrade plan
# kubeadm upgrade apply v1.20.15
# yum install -y kubelet-1.20.15-0 kubectl-1.20.15-0 --disableexcludes=kubernetes
# systemctl daemon-reload
# systemctl restart kubelet

- worker 노드
# yum install -y kubeadm-1.20.15-0 --disableexcludes=kubernetes
# kubeadm upgrade node

- master 노드
# kubectl drain worker1 --ignore-daemonsets --force
# kubectl drain worker2 --ignore-daemonsets --force

- worker 노드
# yum install -y kubelet-1.20.15-0 kubectl-1.20.15-0 --disableexcludes=kubernetes
# systemctl daemon-reload
# systemctl restart kubelet

- master 노드
# kubectl uncordon worker1
# kubectl uncordon worker2

[root@master ~]# kubectl get node
NAME      STATUS   ROLES                  AGE     VERSION
master    Ready    control-plane,master   4h10m   v1.20.15
worker1   Ready    <none>                 4h5m    v1.20.15
worker2   Ready    <none>                 4h5m    v1.20.15


-> 버전 바뀐 모습 확인 가능

📙 클러스터 업그레이드 2

- master 노드
# cd ~
# yum list --showduplicates kubeadm --disableexcludes=kubernetes
# yum install -y kubeadm-1.21.14-0 --disableexcludes=kubernetes
# kubeadm version
# kubeadm upgrade plan
# kubeadm upgrade apply v1.21.14
# yum install -y kubelet-1.21.14-0 kubectl-1.21.14-0 --disableexcludes=kubernetes
# systemctl daemon-reload
# systemctl restart kubelet

- worker 노드
# yum install -y kubeadm-1.21.14-0 --disableexcludes=kubernetes
# kubeadm upgrade node

- master 노드
# kubectl drain worker1 --ignore-daemonsets --force
# kubectl drain worker2 --ignore-daemonsets --force

- worker 노드
# yum install -y kubelet-1.21.14-0 kubectl-1.21.14-0 --disableexcludes=kubernetes
# systemctl daemon-reload
# systemctl restart kubelet

- master 노드
# kubectl uncordon worker1
# kubectl uncordon worker2

[root@master ~]# kubectl get node
NAME      STATUS   ROLES                  AGE     VERSION
master    Ready    control-plane,master   4h23m   v1.21.14
worker1   Ready    <none>                 4h18m   v1.21.14
worker2   Ready    <none>                 4h18m   v1.21.14

📙 ✔️✏️📢⭐️📌

📌기타

⭐️ 프로젝트

https://saramin.github.io/2020-05-01-k8s-cicd/

⭐️ 뭔가 설치하다가 잘못됐거나 안됐을 때 리셋하는 명령어

kubeadm reset

⭐️ kebeadm init 에러

systemctl enable --now docker 다시 해주기.

⭐️ yaml파일 s3에 넣어두고 apply 가능.

# kubectl apply -f  s3주소

⭐️ dockerhub에서 이미지 가져온 경우 경로 문제때문에 웹화면 안나올 수 있음

-> 이미지 이름 자체를 foods 하면 문제가 발생하지 않을 것.

profile
망지
꾸준히, 차근차근
이전 포스트

0729[CI/CD;gtilab,jenkins,tomcat]

다음 포스트

🙂0802[CI/CD-AWS codecommit,Jenkins]

post-custom-banner

0개의 댓글