Are CSV paths for readinessProbe and livenessProbe reversed?

[jwalcorn]

Bug Report

Looking at the generated CSV yaml, under spec.install.spec.deployments, there are two containers: kube-rbac-proxy
and manager. The latter has a livenessProbe and a readinessProbe defined. But the paths look swapped: the /readyz path is on the livenessProbe, not on the readinessProbe as you'd expect from a path named "readyz". The path on the readinessProbe is /healthz. I'm wondering if a copy/paste typo got made here, swapping the two paths.

What did you do?

Generated a helm-based operator and looked at the yaml for the CSV.

What did you expect to see?

                livenessProbe:
                  httpGet:
                    path: /healthz
                    port: 8081
                  initialDelaySeconds: 15
                  periodSeconds: 20
                readinessProbe:
                  httpGet:
                    path: /readyz
                    port: 8081
                  initialDelaySeconds: 5
                  periodSeconds: 10

What did you see instead? Under which circumstances?

                livenessProbe:
                  httpGet:
                    path: /readyz
                    port: 8081
                  initialDelaySeconds: 15
                  periodSeconds: 20
                readinessProbe:
                  httpGet:
                    path: /healthz
                    port: 8081
                  initialDelaySeconds: 5
                  periodSeconds: 10

Environment

Operator type:

/language helm

Kubernetes cluster type:

OpenShift

$ operator-sdk version

v1.4.0

$ go version (if language is Go)

$ kubectl version

v1.17.2

Possible Solution

Swap the paths for the readinessProbe and the livenessProbe

Additional context

If the values are somehow correct, perhaps at least generate a comment explaining the confusion

[jwalcorn]

BTW, I'm getting an image pull error on that manager container. The image field is set to just controller:latest, without a registry or org specified. When I install my operator, I see the following in the events for the pod for my operator:

Failed to pull image "controller:latest": rpc error: code = Unknown desc = Error reading manifest latest in docker.io/library/controller: errors: denied: requested access to the resource is denied unauthorized: authentication required

Do we have to have our users create an ImagePullSecret in order to get our operators working? If so, is there a way I can get my install plan to make that for the user? Thanks.

My CSV is here: https://github.com/IBMStockTrader/stocktrader-operator/blob/master/bundle/manifests/stocktrader-operator.clusterserviceversion.yaml

[estroz]

@jwalcorn the endpoint mixup was fixed in #4546. The v1.5 migration guide, when released, will discuss swapping them.

The intent is to run make bundle IMG=<registry>/<repo>/your-operator:vX.Y.Z so kustomize can substitute the default image name controller:latest for your remote image. If you still have questions about this, open another issue. Thanks!

[jwalcorn]

Thanks, glad to hear the swapped endpoints will be fixed in 1.5. And yes, I had separately figured out what value needed to be in that image field for the manager container, by looking at the sample at https://github.com/operator-framework/operator-sdk/tree/master/testdata/helm/memcached-operator.