Supervisor Upgrade Pre-Check Failure Guide

Products

• VMware vSphere 7.0 with Tanzu
• VMware vSphere with Tanzu

Issue/Introduction

This article is designed to assist customers who have run into some of the more common upgrade pre-check failures introduced in 8.0U3 and guide them through the next steps. 

Cause

Starting in 8.0U3 additional precheck improvements were added during the upgrade process to help the VI administrator troubleshoot any underlying resource or component incompatibility/issue before the upgrade, allowing the upgrade to go through smoothly. 

These pre-checks include various health checks, resource checks, and dependency checks on other software components. 

Resolution

Pre-check Type and Description	Error Code Reported to User when pre-check fails	Troubleshooting
Datastore requirement verified -Checks if the VC datastore has enough resource to create the new Supervisor control plane VM.	Not enough free space on chosen datastore to deploy one more Supervisor CPVM. Need at least {free space in GB}.   For example,  Not enough free space on chosen datastore to deploy one more Supervisor CPVM. Need at least 32 GB.	User needs to reclaim the unused space from the configured datastores for the Supervisor control plane VMs or add additional datastores to the control plane storage policy.  Storage capacity is pulled from the esxi host. If the esxi hosts are all showing the correct storage capacity, then the next step would be to restart the vpxd service on the vCenter. (vmon-cli -r vpxd)
ControlPlaneVMsHealthChecked -Checks if the target Supervisor control plane VM health is marked healthy for upgrade.	Cluster {cluster-moid} is unhealthy: {SYSTEMERROR}.   For example,  Cluster domain-c50 is unhealthy: {SYSTEMERROR}.	Refer to the following document on Resolving Errors Health Statuses on Supervisor During Initial Configuration Or Upgrade
ETCDClusterHealthChecked -Checks if the etcd cluster of the current Supervisor has all etcd members marked in a healthy state.	Etcd members are not healthy during cluster upgrade pre-check.	Refer to the following document on Tanzu Kubernetes Cluster Upgrade Stuck - EtcdMemberHealthy condition is unknown
NetworkCompatibilityVerified   Checks if NSX version is compatible with NSX Container Plugin (NCP) version.	NSX-T version {version-number} is not compatible with NCP version {version-number}. Please choose NSX-T version from compatible list [comma separated list of supported versions] but not in incompatible list [comma separated list of unsupported versions].	Choose a compatible NSX-T version from the list in the error message.
GuestClustersCompatibilityChecked -Checks if running workload clusters are incompatible after upgrade.	Proposed upgrade to Namespaces cluster {cluster moid} is incompatible with one or more Tanzu Kubernetes cluster(s).  Note:  The error message for TKCs will differ in format based on the source SV cluster that it is being upgraded from.	To resolve this error by upgrading the TKC/Cluster to a compatible version before the Supervisor upgrade.
SupervisorServiceCompatibilityChecked  -Checks all the installed Supervisor Services on the given Kube instance, and checks if each installed service version will be compatible with the desired Supervisor version.	The incoming Supervisor requires the Supervisor Service ({0}) version ({1}) but the installed Service is on version {2}. Upgrading the Supervisor will make the installed service incompatible. The Supervisor Service ({0}) version ({1}) requires Supervisor release version {2} but the Supervisor will be on version {3}. Upgrading the Supervisor will make the installed service incompatible.	To resolve the warning upgrade the Supervisor Service to a compatible version before the Supervisor upgrade.