NodePool NodeCount doesn't match MachineSet Replicas

[sjenning]

$ oc get nodepool
NAME                       NODECOUNT
example-cluster-external   

$ oc get machineset -n example example-c-c5ed10a5-example-cluster-external
NAME                                          REPLICAS   AVAILABLE   READY
example-c-c5ed10a5-example-cluster-external   2

This code should be syncing the two
https://github.com/openshift/hypershift/blob/main/hypershift-operator/controllers/nodepool_controller.go#L161-L165

[sjenning]

Note, I am not using the default NodePool. My OCS initialComputeReplicas is 0 and I'm creating my own NodePool.

Logs in the operator show this in a loop

2021-02-02T17:05:00.293Z	INFO	controller.nodepool	Reconciling	{"reconciler group": "hypershift.openshift.io", "reconciler kind": "NodePool", "name": "example-cluster-external", "namespace": "hypershift"}
2021-02-02T17:05:00.293Z	INFO	controller.nodepool	Reconcile nodePool	{"reconciler group": "hypershift.openshift.io", "reconciler kind": "NodePool", "name": "example-cluster-external", "namespace": "hypershift"}
2021-02-02T17:05:00.293Z	INFO	controller.nodepool	Requeueing nodePool	{"reconciler group": "hypershift.openshift.io", "reconciler kind": "NodePool", "name": "example-cluster-external", "namespace": "hypershift", "expected available nodes": 2, "current available nodes": 0}
2021-02-02T17:05:00.294Z	INFO	controller.nodepool	Successfully reconciled	{"reconciler group": "hypershift.openshift.io", "reconciler kind": "NodePool", "name": "example-cluster-external", "namespace": "hypershift"}

[sjenning]

Did some digging and this is because we are using AvailableReplicas in the MachineSet status which only includes Machines that have a NodeRef in their status. The phase for Machines created by our MachineSets (created by the NodePool) is persistently Provisioning with a NodeHealthy condition of False with reason WaitingForNodeRef.

[ironcladlou]

Pretty sure this has been solved.