GITLAB AND GITLAB RUNNER INSTALLATION ON KUBERNETES CLUSTER

tarafından
71
GITLAB AND GITLAB RUNNER INSTALLATION ON KUBERNETES CLUSTER

Gitlab and Gitlab Runner Installation will be explained under this topic. This installation tried on the kubernetes cluster which has a one worker node. Also if you have multiple worker node you can use the node selector to run application on the specific worker node. İf you will be choose to use node selector you must create the mount folder under the selected node.We will use persistent volume, persistent volume claim, deployment and service resources in there. Also we will use nginx reverse proxy and self signed certificate, to connect gitlab application after created all kubernetes resources.
You must follow below commands;

Step1:

Create mount folder under your selected worker node and create yaml file for persistent volume and volume claim. İf you want to use the node selector for deployment you must add label on your node and you should edit the your deployment yaml file. 

mkdir -p /opt/gitlab/logs /opt/gitlab/data /opt/gitlab/config
vim gitlabpv.yaml
apiVersion: v1
kind: PersistentVolume
metadata:
  name: gitlabcnfgpv
spec:
  capacity:
    storage: 1Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/opt/gitlab/config"
---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: gitlablgspv
spec:
  capacity:
    storage: 1Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/opt/gitlab/logs"
---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: gitlabdtpv
spec:
  capacity:
    storage: 1Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/opt/gitlab/data"
vim gitlabpvc.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: gitlabcnfpvc
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: gitlablgspvc
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: gitlabdtpvc
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
kubectl create -f gitlabpvc.yaml

Then, check your pv and pvc resources.

kubectl get pv

Then, we can see our pv and pvc resources above picture. PVC resources bounded with the PV resources.

Step2:
Create deployment file after the label node. We suppose the we have multiple worker node and we choose the specific node to run the gitlab application

kubectl label nodes <worker-node-name> app=gitlab
vim gitlab.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: gitlab
  labels:
    app: gitlab
spec:
  replicas: 1
  selector:
    matchLabels:
      app: gitlab
  template:
    metadata:
      labels:
        app: gitlab
    spec:
      hostAliases:
        - ip: "your_ip_address"
          hostnames:
          - gitlab-ingress.<your_ip_address>.nip.io
      containers:
      - name: gitlab
        image: gitlab/gitlab-ee:latest
        ports:
        - name: http
          containerPort: 80
        - name: https
          containerPort: 443
        - name: ssh
          containerPort: 22
        volumeMounts:
        - mountPath: "/etc/gitlab"
          name: gitlabconfig
        - mountPath: "/var/log/gitlab"
          name: gitlablogs
        - mountPath: "/var/opt/gitlab"
          name: gitlabdata
      volumes:
      - name: gitlabconfig
        persistentVolumeClaim:
          claimName: gitlabcnfpvc
      - name: gitlablogs
        persistentVolumeClaim:
          claimName: gitlablgspvc
      - name: gitlabdata
        persistentVolumeClaim:
          claimName: gitlabdtpvc
      nodeSelector:
        app: gitlab
kubectl create -f gitlab.yaml

Now we should go to worker node which gitlab is running and Open mount folder and add below lines for server name.

vim /opt/gitlab/config/gitlab.rb
external_url "http://gitlab-ingress.<192.168.56.111>.nip.io"

Then you must use below command to reconfigure gitlab.

pod=$(kubectl get pods | awk '{print $1}' | grep -e "gitlab")
kubectl exec -it $pod bash
gitlab-ctl reconfigure
exit

Step3:
Create Service file to establish the connection for  gitlab application.

vim gitlabsvc.yaml
apiVersion: v1
kind: Service
metadata:
  name: gitlab
  labels:
    app: gitlab
spec:
  ports:
  - port: 80
    name: http
    protocol: TCP
  - port: 443
    name: https
    protocol: TCP
  - port: 22
    name: ssh
    protocol: TCP
  selector:
    app: gitlab
  type: NodePort
kubectl create -f gitlabsvc.yaml

Now we can check the our created service and we should take the node port number to establish connection for gitlab with our browser.

kubectl get svc

If you access the gitlab application like picture below, you must chenge the default root password with connect the pod with “exec” command and use the following lines.

pod=$(kubectl get pods | awk '{print $1}' | grep -e "gitlab")
kubectl exec -it $pod bash
gitlab-rails console
 user = User.find_by(username: "root")
 user.password = 'hafifbilgiler123'
 user.password_confirmation = 'hafifbilgiler123'
 user.save
exit
exit

Step4:
Create Self Signed Certficate and secret for Secure connection  and  Ingress yaml file to access the gitlab application for without to using node port ip address. Also if you don’t have a Ingress Controller, you can get a lot of information about the Nginx Controller Installation to click hereAlso I have been used the nip.io to DNS solving.

openssl req -x509 -nodes -days 365 -newkey rsa:2048     -out gitlab-ingress-ssl.crt     -keyout gitlab-ingress-ssl.key     -subj "/CN=gitlab-ingress.<your_ip_addrress>.nip.io/O=gitlab-ingress.<your_ip_addrress>.nip.io"
kubectl create secret tls gitlab-ingress-secret --key ./gitlab-ingress-ssl.key     --cert ./gitlab-ingress-ssl.crt
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: gitlab-ingress
spec:
  tls:
  - hosts:
    - gitlab-ingress.<your_ip_addrress>.nip.io
    secretName: gitlab-ingress-secret
  ingressClassName: nginx
  rules:
  - host: gitlab-ingress.<your_ip_addrress>.nip.io
    http:
      paths:
      - backend:
          service:
            name: gitlab
            port:
              number: 80
        path: /
        pathType: Prefix
status:
  loadBalancer: {}
vim ingress.yaml
kubectl create -f ingress.yaml

Step5:
Create  volume folders and yaml files for persistentent volume and persistent volume claim.

mkdir -p /opt/gitlab/gitlab-runner/config /opt/gitlab/gitlab-runner/docker.sock /opt/gitlab/gitlab-runner/SSL
vim gitlabrunner_pv_pvc.yaml
apiVersion: v1
kind: PersistentVolume
metadata:
  name: gitlabrunnercnfgpv
spec:
  capacity:
    storage: 1Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/opt/gitlab/gitlab-runner/config"
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: gitlabrunnercnfgpvc
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: gitlabrunnerdockerpv
spec:
  capacity:
    storage: 1Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/opt/gitlab/gitlab-runner/docker.sock"
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: gitlabrunnerdockerpvc
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: gitlabrunnerssl
spec:
  capacity:
    storage: 1Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/opt/gitlab/gitlab-runner/SSL"
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: gitlabrunnerssl
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
kubectl create -f gitlabrunner_pv_pvc.yaml

Step6:
Create yaml file for gitlab runner application.

vim gitlabrunner.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: gitlab-runner
  labels:
    app: gitlab-runner
spec:
  replicas: 1
  selector:
    matchLabels:
      app: gitlab-runner
  template:
    metadata:
      labels:
        app: gitlab-runner
    spec:
      hostAliases:
        - ip: "your-ip-address"
          hostnames:
          - gitlab-ingress.your-ip-address
      containers:
      - name: gitlab-runner
        image: gitlab/gitlab-runner:latest
        #command: ['sh', '-c', 'mkdir /opt/SSL']
        ports:
        - name: http
          containerPort: 80
        - name: https
          containerPort: 443
        - name: ssh
          containerPort: 22
        volumeMounts:
        - mountPath: "/etc/gitlab-runner"
          name: config
        - mountPath: "/var/run/docker.sock"
          name: dockersock
        - mountPath: "/opt/SSL"
          name: sslcertificate
      volumes:
      - name: config
        persistentVolumeClaim:
          claimName: gitlabrunnercnfgpvc
      - name: dockersock
        persistentVolumeClaim:
          claimName: gitlabrunnerdockerpvc
      - name: sslcertificate
        persistentVolumeClaim:
          claimName: gitlabrunnerssl
      nodeSelector:
        app2: gitlab-runner
kubectl create -f gitlabrunner.yaml

Step7:
Now We can test it. Create basic project with  echo messages and some steps. Create bash  Template .gitlab-ci.yml file for this project on gitlab like below. Take your ip address and token from project CI/CD Settings. And you must write your gitlab url address with https.

Step8: 
Go to your worker node which Gitlab-Runner running and put your certificate under volume path “/opt/gitlab/gitlab-runner/SSL”. Then you will use the this certficate for trusted connection during  gitlab runner register.

pod=$(kubectl get pods | awk '{print $1}' | grep -e "gitlab")
kubectl exec -it $pod bash
gitlab-runner register --tls-ca-file=/opt/SSL/gitlab-ingress-ssl.crt

Step9:
Then, you can check the your project if it does register or does not from under CI / CD Settings.

That’s it. If you need to run bash command or other scripts on remote computer  you can add gitlab runner as a container like above steps.