I can't get into the azure kubernetes pod

4/18/2020

I want to check inside the pod the flaw I'm having inside the image but the errors below appear. As I verified in describe, the name of the container is correct. What else can I do to get the connection in the cluster?

Command: kubectl exec -it -c airflow-console -n airflow airflow-console-xxxxxxx-xxxxx bash

error: unable to upgrade connection: container not found ("airflow-console")

Command: kubectl describe pod/airflow-console-xxxxxxx-xxxxx -n airflow

Events:
  Type     Reason     Age                    From                                           Message
  ----     ------     ----                   ----                                           -------
  Normal   Scheduled  37m                    default-scheduler                              Successfully assigned airflow/airflow-console-xxxxxxx-xxxxx to aks-test
  Normal   Pulling    37m                    kubelet, aks-test  Pulling image "test.azurecr.io/airflow:2"
  Normal   Pulled     37m                    kubelet, aks-test  Successfully pulled image "test.azurecr.io/airflow:2"
  Warning  BackOff    36m                    kubelet, aks-test  Back-off restarting failed container
  Normal   Pulled     36m (x3 over 37m)      kubelet, aks-test  Container image "k8s.gcr.io/git-sync:v3.1.2" already present on machine
  Normal   Created    36m (x3 over 37m)      kubelet, aks-test  Created container git-sync
  Normal   Started    36m (x3 over 37m)      kubelet, aks-test  Started container git-sync
  Normal   Created    36m (x3 over 36m)      kubelet, aks-test  Created container airflow-console
  Normal   Pulled     36m (x2 over 36m)      kubelet, aks-test  Container image "test.azurecr.io/airflow:2" already present on machine
  Normal   Started    36m (x3 over 36m)      kubelet, aks-test  Started container airflow-console
  Warning  BackOff    2m15s (x178 over 36m)  kubelet, aks-test  Back-off restarting failed container
-- Felipe FB
azure-kubernetes
kubectl
kubernetes

1 Answer

4/20/2020

this line

Warning  BackOff    2m15s (x178 over 36m)  kubelet, aks-test  Back-off restarting failed container

shows that your pod/container is in a failed state. This will prevent you from execution commands in the container due to it not being alive.

To learn why your pod/container is in a bad state, you should look at the logs of the failed container

kubectl logs -n airflow airflow-console-xxxxxxx-xxxxx -c airflow-console

or the logs off the previous container that failed. (sometimes it helps)

kubectl logs -n airflow airflow-console-xxxxxxx-xxxxx -c airflow-console -p

This explain the main reason why a user cannot exec into a container.

-- djsly
Source: StackOverflow