将 Node/mongodb 应用程序部署到 GKE 并且它因 crashloopbackoff 不断失败

Question

I deployed my nodejs/typeScript container alongside mongodb image on GKE. But the nodejs app deployment keeps failing, and I know because I can't hit the server with the external IP address and port. When I inspect the deployment on GCP console, I see that it fails with error: Does not have minimum availability and crashloopbackoff . From the logs I could also see that the app fails to connect to the database I deployed alongside it, so it appears it can't also find the MongoDB service, could it be why it says crashloopbackoff ?

Here are the yaml files for the components I deployed:

MongoDB pod

apiVersion: v1
kind: Pod
metadata:
  name: mongodb
  labels:
    app: mongodb
spec:
  containers:
    - name: mongodb
      image: mongo
      ports:
        - containerPort: 27017
          name: mongodb-port

MongoDB service

apiVersion: v1
kind: Service
metadata:
  name: mongodb # Same service name I used in my db url
spec:
  ports:
    - port: 27017
      protocol: TCP
  selector:
    app: mongodb
  type: ClusterIP

Back-end pod's service

apiVersion: v1
kind: Service
metadata:
  name: journal-api-service
spec:
  ports:
    - port: 8080
      protocol: TCP
      targetPort: http-port
  selector:
    app: journal-api
    tier: backend
  type: LoadBalancer

Back-end deployment

apiVersion: apps/v1
kind: Deployment
metadata:
  name: journal-api-deployment
spec:
  selector:
    matchLabels:
      app: journal-api
      tier: backend
  replicas: 1
  template:
    metadata:
      labels:
        app: journal-api
        tier: backend
    spec:
      containers:
        - image: shulaa/journal-api:v1
          name: journal-api
          ports:
            - containerPort: 8080
              name: http-port

and I created them in the exact order I provided it here. The mongo and nodejs app containers run fine when I run them locally.

Answer 1

That's due to the

Does not have minimum availability

your container is not coming up or running due to a resource outage. No resources into the cluster to run the container.

Now you should add New node so K8s can allocate some resources to container, or delete running pods.

You check Node status using the

kubectl get nodes 

check resource of Nodes

kubectl top nodes  

if your nodes as resources left there could be a chance no resource left GCP side in zone or so

Wait. Google Cloud does not have enough resource available in the Region/Zone that you are trying to launch into. In some cases, this took an hour to an entire day.

Select a different Region/Zone.