STS: Mongodb goes into `CrashLoopBackOff`

[surajssd]

Describe the bug

I installed the app using aks-store-all-in-one.yaml config but the mongodb pod goes into CrashLoopBackOff because its liveness probe fails. Liveness check is failing with the following error:

pets 0s Warning Unhealthy pod/mongodb-0 Liveness probe failed: OCI runtime exec failed: exec failed: unable to start container process: exec: "mongosh": executable file not found in $PATH: unknown

To Reproduce
Steps to reproduce the behavior:

1. Follow steps from: https://github.com/Azure-Samples/aks-store-demo/tree/main?tab=readme-ov-file#run-on-any-kubernetes i.e.:

kubectl create ns pets

kubectl apply -f https://raw.githubusercontent.com/Azure-Samples/aks-store-demo/main/aks-store-all-in-one.yaml -n pets

2. Run the following to see that the mongodb-0 pod keeps restarting:

kubectl get events -w -n pets

Expected behavior

Pod should pass the liveness probe.

Source of the error

aks-store-demo/aks-store-all-in-one.yaml

Lines 32 to 36 in 5712ff5

	exec:
	command:
	- mongosh
	- --eval
	- db.runCommand('ping').ok

[pauldotyu]

@surajssd what type of kubernetes cluster were you running this on?

[surajssd]

@pauldotyu I was using Minikube.

But this seems like an issue regardless of the k8s kind. This looks like a container image issue.

[ITOpenwork]

The following settings seem to work good in minikube:

        livenessProbe:
          exec:
            command:
            - mongosh
            - --eval
            - db.runCommand('ping').ok
          initialDelaySeconds: 30
          periodSeconds: 20
          timeoutSeconds: 10
          failureThreshold: 6
          successThreshold: 1

[kszicsillag]

But the original error message looks valid: it seems that the image (mcr.microsoft.com/mirror/docker/library/mongo:4.2) referenced in the linked yaml does not contain the mongosh tool (only mongo CLI).