0

I've deployed a bare-metal Kubernetes cluster with Calico networking. I'm successfully able to communicate between pods and from external network to pods using Ingress. When I use Dashboard EXEC function the pod is able to fetch external resource (DNS works, PING works, CURL works), but the pod process itself is failing.

What should I check next? Do you need more info to analyze it?

UPDATE

Issue is related to Gitlab Runner only.

Jan Zahradník
  • 2,417
  • 2
  • 33
  • 44
  • What do you mean by Dashboard EXEC? You execute a shell command in the running container? – weibeld Sep 23 '19 at 11:06
  • @weibeld At Kubernetes Dashboard at Pod detail there is EXEC button at top right corner. I assume it opens bash in the container. – Jan Zahradník Sep 23 '19 at 12:47
  • @JanZahradník what do you mean by "pod process itself is failing"? Could you provide output of `kubectl describe pod ` – PjoterS Sep 23 '19 at 13:28

1 Answers1

1

Finally it turned out that the problem was not with whole Kubernetes cluster, but related only to Gitlab Runner. Thanks to this issue it works again:

Successfully able to complete builds by adding the network to the build command --network host

build:
  script: 
    - docker build --network host .

This requires a .gitlab-ci.yml and excludes autodevops.

Jan Zahradník
  • 2,417
  • 2
  • 33
  • 44