vmconfig-operator down, in error state, when deploying multiple node pools

Products

• VMware Telco Cloud Automation

Issue/Introduction

• When simultaneously deploying multiple nodepools, vmconfig crashed with nil pointer error:

INFO    Observed a panic in reconciler: runtime error: invalid memory address or nil pointer dereference    {"controller": "vmconfig", "controllerGroup": "acm.vmware.com", "controllerKind": "VMConfig", "VMConfig": {"name":"<wc_name>","namespace":"<wc_namespace>"}, "namespace": "<wc_namespace>", "name": "<wc_name>", "reconcileID": "########-####-####-####-############"}

• Nodepolicy successfully created but stuck in processing state.
• vmconfig-operator pod is down/in error state, on the TKG Mgmt cluster:

kubectl get pods -A | grep vmconfig-operator
tca-system   vmconfig-operator-############    1/2.   Error     1 (39s ago)       41s

Note: This issue may not occur in every case simply by deploying multiple node pools in parallel. It depends on how busy the vCenter is when populating the VM configuration for those objects before the TCA-CP inventory syncs with vCenter. If the vCenter is experiencing resource contention or is slow, the user may notice this issue.

• After a workload cluster upgrade, with the nodepool in a provisioned status, additional network interfaces are not renamed.

Environment

3.2

Cause

The node pool has a vmConfig null property on vCenter which is causing vmconfig-operator to crash.

Resolution

Resolved in TCA 3.3

Workaround

Apply the Patch Tool for TCA 3.2.0.1 KB.