0

We are trying to deploy PingFed in a Kubernetes cluster. We have files deployment.yml and service.yml, but the pod is throwing CrashBackLoop error after trying to reach the server profile with a readiness probe error. We have tried everything with changing the service type to tinkering with deployment and service YAML files, but nothing seems to be working.

This is the deployment.yml file

Enter image description here

Service.yml

POD error

All the pods, services, deployment running in the namespace

All the pods, services, deployment running in the namespace

Jonas
  • 121,568
  • 97
  • 310
  • 388
Atmesh Acharya
  • 11
  • 1
  • 2
  • Service.yaml apiVersion: v1 kind: Service metadata: name: pingfederate labels: role: pingfederate tier: console version: 10.1.5 spec: ports: - name: console port: 9999 protocol: TCP targetPort: 9999 selector: role: pingfederate tier: console version: 10.1.5 type: LoadBalancer – Atmesh Acharya Jul 16 '21 at 18:42
  • 3
    Please do not post images of code. Please edit your question and include the code as text (formatted as a code block). – DaveL17 Jul 18 '21 at 11:38
  • Welcome to the community! First as already asked, please replace you screenshots with actual text/code. Then, you can try to add `initialDelaySeconds` to your readiness probe as @San suggested below. If it doesn't work, remove completely healthcheck to have pod running and then using `curl` test if this application based on this container actually serves on port `9999` and `/pingfederate/app`. – moonkotte Jul 19 '21 at 12:33
  • Thank you guys for the suggestions. I'm fairly new on this platform, will improve on that. @moonkotte Let me try if that works. – Atmesh Acharya Jul 19 '21 at 15:31
  • Can you please [make](https://stackoverflow.com/posts/68413879/edit) the requested changes? You can [edit](https://stackoverflow.com/posts/68413879/edit) (change) your question. Thanks in advance. – Peter Mortensen Aug 18 '21 at 06:33
  • https://idownvotedbecau.se/imageofcode – rkosegi Aug 18 '21 at 06:33

2 Answers2

2

I don't see initialDelaySeconds in the readiness probe. So, I guess that might be the reason the readiness probe is failing. I would set it to the time it takes for the application to come up.

initialDelaySeconds: Number of seconds after the container has started before liveness or readiness probes are initiated. Defaults to 0 seconds. Minimum value is 0.

readinessProbe:
      httpGet:
        path:
        port: 8080
      initialDelaySeconds: 5
      periodSeconds: 10
Peter Mortensen
  • 30,738
  • 21
  • 105
  • 131
San
  • 191
  • 4
1

I would check and make sure a couple of things:

  1. Check that 9999 port is in fact exposed on the container, check the dockerfile if you can, I can already see it's opened up on the deployment.
  2. If it is, I noticed your scheme on probes is set to https, should that be http instead?
Noah Dlugoszewski
  • 41
  • 3