VCF upgrade pre-check from 5.2 to 9.0.2 fails with error "BREAK_ELM_ONLY_SUPPORTED_ON_JOINED_MGMT_SSO"

Products

• VMware Cloud Foundation
• VMware SDDC Manager / VCF Installer

Issue/Introduction

• An upgrade of VMware Cloud Foundation (VCF) from version 5.2 to version 9.0.2 fails during the SDDC Manager Lifecycle Management upgrade prechecks. A blocking error occurs with the following message structure:

  "errorCode": "BREAK_ELM_ONLY_SUPPORTED_ON_JOINED_MGMT_SSO",
"message": "Domain ##### is not joined to the management SSO (ELM)."


• Enhanced Linked Mode (ELM) may have already been deactivated prior to this event via standard procedures - Deactivate Enhanced Link Mode from vCenter Using the cmsso-util break-elm Utililty

• The /var/log/vmware/vcf/domainmanager/domainmanager.log file records the following error trace:
  
  YYYY-MM-DDTHH:MM:SS.###+0000 ERROR [vcf_dm,################,####] [c.v.e.s.e.h.LocalizableRuntimeExceptionHandler,http-nio-127.0.0.1-7200-exec-7] [######] BREAK_ELM_ONLY_SUPPORTED_ON_JOINED_MGMT_SSO Domain ##### is not joined to the management SSO (ELM). com.vmware.evo.sddc.common.services.error.SddcManagerServicesIsException: Domain ##### is not joined to the management SSO (ELM).

Environment

• VMware Cloud Foundation 5.2
  
  
• VMware Cloud Foundation 9.0.2

Cause

The SDDC Manager inventory contains null or empty values for the join_sso_status field for vCenter Server entries in the platform database. This database discrepancy prevents the upgrade engine from processing the correct SSO tracking data following ELM deactivation.

Querying the vcenter table in the platform database confirms the empty join_sso_status attribute:

platform=# select id, status, type, version, vm_hostname, join_sso_status from vcenter;
-[ RECORD 1 ]------------+-------------------------------------
id                       | ########-####-####-####-############
status                   | ACTIVE
type                     | MANAGEMENT
version                  | 9.0.2.0.########
vm_hostname              | <vc_fqdn>
join_sso_status          |
-[ RECORD 2 ]------------+-------------------------------------
id                       | ########-####-####-####-############
status                   | ACTIVE
type                     | VI
version                  | 9.0.2.0.########
vm_hostname              | <vc_fqdn>
join_sso_status          |

Resolution

To resolve this issue, the inventory records within the platform database must be updated to align with the expected state, followed by triggering the drift resolution task.

1. A snapshot of the SDDC Manager appliance is captured before proceeding.

2. An SSH session to the SDDC Manager appliance is initiated.

3. Log in is completed with the user vcf, and the shell is transitioned to the root user using the su - command.

4. The platform database tool is opened with the following instruction:

   psql -h localhost -U postgres -d platform
   

5. The active vCenter records and their blank statuses are verified by running:

   select id, status, type, version, vm_hostname, join_sso_status from vcenter;
   

6. The join_sso_status columns are populated with 'JOINED' for the affected vCenter UUIDs by executing:

   update vcenter set join_sso_status = 'JOINED' where id in ('<INSERT_UUID_1>','<INSERT_UUID_2>');
   

7. The database prompt is closed by typing \q.

8. The SDDC Manager API Enhanced Linked Mode (ELM) drift resolution task is invoked to reconcile state configurations.

9. The VMware Cloud Foundation 9.0.2 upgrade precheck is re-run.