SadServers: medium task - Kubernetes Pod Crashing

Sadservers is a platform that delivers solving problem tasks . You receive a server from them and simply solve the problem. It’s an intriguing platform for learning debugging.

I encourage you to first attempt to solve the task on your own, then observe how I approached it. If you find yourself stuck, you can always refer back here.

Website: https://sadservers.com/ Task: “Buenos Aires”: Kubernetes Pod Crashing

Solution:

I begin by reading the task description attentively. Description: There are two brother pods, logger and logshipper, residing in the default namespace. Unfortunately, logshipper is experiencing an issue (crashlooping) and is unable to see what logger is trying to communicate. Could you assist in fixing Logshipper? You can check the status of the pods with ‘kubectl get pods’.

As the task type is ‘Fix’, we need to resolve the problem in the Kubernetes cluster.

Okay so I’ve written description. Now I am checking the status of the pods

sudo kubectl get pods

Reults:

admin@i-0b93d760028a401a4:~$ sudo kubectl get pod 
NAME                         READY   STATUS    RESTARTS       AGE
logger-574bd75fd9-wpg4w      1/1     Running   5 (75s ago)    56d
logshipper-7d97c8659-npscd   1/1     Running   20 (56d ago)   56d
admin@i-0b93d760028a401a4:~$ 

Now I am wanting to check what is happening in logger so:

sudo kubectl logs logger-574bd75fd9-wpg4w

Result:

Hi, My brother logshipper cannot see what I am saying. Can you fix him

Okay so let’s focus now on logshipper and investigate if service account attached to the pod

sudo kubectl get pod logshipper-7d97c8659-npscd -o yaml | grep serviceAccount

Result:

serviceAccount: logshipper-sa

Okay so we have service account. Let’s check what is happening with it ClusterRole.

sudo kubectl get ClusterRole logshipper-sa -o yaml

Result:

apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  annotations:
    descripton: Think about what verbs you need to add.
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"rbac.authorization.k8s.io/v1","kind":"ClusterRole","metadata":{"annotations":{"descripton":"Think about what verbs you need to add."},"name":"logshipper-cluster-role"},"rules":[{"apiGroups":[""],"resources":["namespaces","pods","pods/log"],"verbs":["list"]}]}
  creationTimestamp: "2023-10-23T02:40:43Z"
  name: logshipper-cluster-role
  resourceVersion: "1286"
  uid: eb5541f8-870c-40e0-ac57-17f7c4d24fd5
rules:
- apiGroups:
  - ""
  resources:
  - namespaces
  - pods
  - pods/log
  verbs:
  - list

Here we need add get verb and watch in case of reading

Then you need to restart deployment via rollout deployment

sudo kubectl rollout restart deployment logshipper

Result:

deployment.apps/logshipper restarted

Now we can check pod once again and everything should be OK now.

Summary

In this article, I showed you how to solve the problem of the medium from sadservers. It’s a good exercise in Kubernetes problem solving. I hope you enjoyed this article and see you next time.