Workload Domain Deployment in SDDC Manager Fails the Topology Test with error: "INVALID PARTNER COUNT"

Products

• VMware Cloud Foundation

Issue/Introduction

Symptoms:

The domain deployment fails with the following error details in the UI:

Workload Deployment failed "Fetched topology with 3 nodes, contains nodes with invalid partner count"

Running vdcrepadmin on the vCenters shows that a proper ring-topology is in place.

There is no evidence of replication errors and all vmdir instances have a status of 'NORMAL'.

There is no evidence of stale vCenters and/or PSCs in the environment.

The domainmanger log has logging similar to below:

YYYY-MM-TT:HH:MM.564+0000 DEBUG [vcf_dm,398737e9105a5b3e,1b73] [c.v.v.v.s.s.SsoRingTopologyServiceImpl,dm-exec-1] Found management nodes ["acme1.test.com","acme2.test.com"]
...
... 
YYYY-MM-TT:HH:MM.893+0000 DEBUG [vcf_dm,398737e9105a5b3e,1b73] [c.v.v.v.s.s.e.SsoRestCommandExecutor,dm-exec-1] Session creation to SSO host acme.test.com was successful. Status code was: 200 OK                      YYYY-MM-TT:HH:MM:17.893+0000 DEBUG [vcf_dm,398737e9105a5b3e,1b73] [c.v.v.v.s.s.SsoRingTopologyServiceImpl,dm-exec-1] No dead nodes found in the ring topology.                                                          YYYY-MM-TT:HH:MM:17.893+0000 ERROR [vcf_dm,398737e9105a5b3e,1b73] [c.v.e.s.o.model.error.ErrorFactory,dm-exec-1] [39UC4] SSO_RING_TOPOLOGY_INVALID_NODES Fetched topology with 3 nodes, contains nodes with invalid partner count, invalid ones (with unexpected respective partner counts) are: [host acme1.test.com(1), host acme2.test.com(1)]
com.vmware.evo.sddc.orchestrator.exceptions.OrchTaskException: Fetched topology with 3 nodes, contains nodes with invalid partner count, invalid ones (with unexpected respective partner counts) are: [host acme1.test.com(1), host acme2.test.com(1)]
        at com.vmware.vcf.vimanager.fsm.actions.HealthCheckRingTopologyVc70Action.execute(HealthCheckRingTopologyVc70Action.java:59)
        at com.vmware.vcf.vimanager.fsm.actions.HealthCheckRingTopologyVc70Action.execute(HealthCheckRingTopologyVc70Action.java:24)

Note the highlighted log entry - it suggests that the SDDC Manager has detected two Management vCenters when, of course, there should only be one.

 

Environment

VMware Cloud Foundation 4.5.x

Cause

This is caused by an erroneous advanced setting on one of the VI WLD vCenters which tags it as a Management node instead of a VI node.

Resolution

Check the Advanced Setting of each vCenter in the ELM:

vCenter -> Configure -> Advanced Settings -> config.SDDC.Deployed.WorkloadDomain

Ensure that the Management vCenter has a value of "Management" for this advanced Setting.

If a VI vCenter also has a value of "Management" this value must be changed to a value of "VirtualInfrastructure".

Save the settings and restart the workflow from the SDDC Manager.