Thread spins logging "Error while waiting for resource to be scaled"

[k-wall]

Using 4.6.2, I am seeing a situation where a kubernetes-client thread spins in a tight loop with the following exception logged, every 50ms:

2019-11-06T14:59:28.679Z ERROR [StatefulSetOperationsImpl] Error while waiting for resource to be scaled. java.lang.NullPointerException: null at io.fabric8.kubernetes.client.dsl.internal.StatefulSetOperationsImpl.getCurrentReplicas(StatefulSetOperationsImpl.java:72) at io.fabric8.kubernetes.client.dsl.internal.StatefulSetOperationsImpl.getCurrentReplicas(StatefulSetOperationsImpl.java:35) at io.fabric8.kubernetes.client.dsl.internal.RollableScalableResourceOperation.lambda$waitUntilScaled$0(RollableScalableResourceOperation.java:104) at java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515) at java.base/java.util.concurrent.FutureTask.runAndReset(FutureTask.java:305) at java.base/java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:305) at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128) at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628) at java.base/java.lang.Thread.run(Thread.java:834) 201

I notice that StatefulSetOperationsImpl#getObservedGeneration is careful to guard against a null current.getStatus(), but StatefulSetOperationsImpl.getCurrentReplicas does not.

[k-wall]

@rohanKanojia

I understand more about this issue now. It comes about when you've io.fabric8.kubernetes.client.DefaultKubernetesClient#load a resource such as a statefulset then call io.fabric8.kubernetes.client.dsl.Deletable#delete on the returned object.

In this case, the object is never fetched from the server (reloadingFromServer is false so io.fabric8.kubernetes.client.dsl.base.BaseOperation#getMandatory returns the existing object). In the case of a scaleable resource, prior to deletion the code tries to scale to zero, but this fails with the NPE if the resource lacks a Status section.

My initial PR doesn't really help. The code still just spins comparing -1 != 0 rather than throwing the NPE.

I think the code ought to omit the waitUntilScaled part if reloadingFromServer is false.

[rohanKanojia]

@k-wall : apologies for the late reply, let me go through your new PR and see if can be merged.