AWS EKS Kubernetes 和 DockerHub

Question

I have a cluster and node creates in AWS EKS. I applied the deployment to that cluster as under

kubectl apply -f deployment.yaml

Where deployment.yaml contains the containers' specification along with DockerHub repo and image

However, I did a mistake in deployment.yaml and I need to re-apply it to the configuration

My question is:

1 - How do I reapply a deployment.yaml to the AWS EKS cluster using kubectl? Just running the above command is not working ( kubectl apply -f deployment.yaml )

2- After I re-apply the deployment.yaml, will the node will go an pick up the DockerHub image or do I still need to do something else( supposing all the other details are ok)

Some outputs below:

>> kubectl get pods

my-app-786dc95d8f-b6w4h   0/1     ImagePullBackOff   0          9h
my-app-786dc95d8f-w8hkg   0/1     ImagePullBackOff   0          9h

kubectl describe pod my-app-786dc95d8f-b6w4h                                                                            
Name:         my-app-786dc95d8f-b6w4h
Namespace:    default
Priority:     0
Node:         ip-192-168-24-13.ec2.internal/192.168.24.13
Start Time:   Fri, 10 Jul 2020 12:54:38 -0400
Labels:       app=my-app
              pod-template-hash=786dc95d8f
Annotations:  kubernetes.io/psp: eks.privileged
Status:       Pending
IP:           192.168.7.235
IPs:
  IP:           192.168.7.235
Controlled By:  ReplicaSet/my-app-786dc95d8f
Containers:
  simple-node:
    Container ID:
    Image:          BAD_REPO/simple-node
    Image ID:
    Port:           80/TCP
    Host Port:      0/TCP
    State:          Waiting
      Reason:       ImagePullBackOff
    Ready:          False
    Restart Count:  0
    Environment:    <none>
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-mwwvl (ro)
Conditions:
  Type              Status
  Initialized       True
  Ready             False
  ContainersReady   False
  PodScheduled      True
Volumes:
  default-token-mwwvl:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-mwwvl
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type     Reason   Age                    From                                    Message
  ----     ------   ----                   ----                                    -------
  Normal   BackOff  17m (x2570 over 9h)    kubelet, ip-192-168-24-13.ec2.internal  Back-off pulling image "BAD_REPO/simple-node"
  Warning  Failed   2m48s (x2634 over 9h)  kubelet, ip-192-168-24-13.ec2.internal  Error: ImagePullBackOff

BR

Answer 1

if you need to change image:

kubectl set image deployment.v1.apps/{your_deployment_name} image_name:tag

but you always can do

kubectl delete -f deployment.yaml
kubectl create -f deployment.yaml

since your image is in ImagePullBackOff - it doesn't work anyway and you can just recreate deployment. Usually you don't do drop/create on prod. that is why i am using image change all the time. just have to change tag on every new image.

Answer 2

ImagePullBackOff means that kubernetes is not able to pull the image.

Specially, the service account "default" is not able to pull the image.

To fix this issue, you need two checks:

• Check that you don't have typo in the image name and tag. And that image is available publically.
• If the Docker registry is private, make sure to create secret with dockerlogin type, and then patch the service account "default" by this secret.