VCF Management Component Deployment fails with "existing deployment" error in SDDC Manager
search cancel

VCF Management Component Deployment fails with "existing deployment" error in SDDC Manager

book

Article ID: 445490

calendar_today

Updated On:

Products

VCF Operations

Issue/Introduction

When attempting to deploy VCF Management components (VCF Operations, VCF Automation) on a VMware Cloud Foundation 9.x environment, the deployment fails.

  • The SDDC Manager UI or API returns the error: VCF_MANAGEMENT_COMPONENTS_DEPLOYMENT_STATUS_ERROR.
  • The deployment sub-task "Trigger VCF Operations fleet management deployment on SDDC Manager" shows a status of FAILED.
  • Subsequent attempts to retry or start a new deployment fail immediately with a message stating that an existing deployment is already present.

    SubTask (Trigger VCF Operations fleet management deployment on SDDC Manager) {

"completionTimestamp": "YYYY-MM-DDTHH:MM:SS".
"creationTimestamp": "YYYY-MM-DDTHH:MM:SS",
"description": "Trigger VCF Operations fleet management deployment on SDDC Manager",
"errors":

Error {
"arguments": [ ],
"causes":
  [
ErrorCause  { .. }.

"errorCode": "FAILED_TO_DEPLOY_VCF_OPS_MGMT",
"errorType": "ERROR",
"message": "Failed to deploy VCF Operations fleet management on SDDC Manager",
"referenceToken": "<Token_Ref>",

"name": "Trigger VCF Operations fleet management deployment on SDDC Manager",
"status": "FAILED",

Environment

VCF 9.x

Cause

This issue occurs when a previous deployment attempt is interrupted or fails during the initial orchestration phase.
SDDC Manager platform database retains stale entries in the management component tables, which effectively "locks" the deployment state to prevent concurrent installations.

You can verify the stale entries by running the following query in the SDDC Manager PostgreSQL database:

 Access the platform database: psql -h localhost -U postgres -d platform

platform=# select * from vcf management_component;
component_type                  |   fqdn              		 | deployment_type | deployment_status 
----------------------------------------------------------------------------------------------
VCF OPERATIONS:MASTER			| <vcf-ops-master-fqdn>		 | NEW	   		   | NOT STARTED
VCF OPERATIONS_CLOUD PROXY		| <vcf-ops-cloud-proxy-fqdn> | NEW             | NOT STARTED
VCF AUTOMATION					| <vcfa-fqsn>				 | NEW             | NOT STARTED
VCF OPERATIONS MANAGEMENT		| <fleet-manager-fqdn>       | NEW             | IN PROGRESS
(4 rows)

Resolution

[!CAUTION] IMPORTANT: Before performing any manual modifications or deletions within the SDDC Manager PostgreSQL database, you must take a virtual machine snapshot of the SDDC Manager appliance. This ensures a valid restore point is available in the event of unforeseen database corruption or service disruption.

  1. Take a Snapshot of the SDDC Manager VM.
  2. SSH into the SDDC Manager as the vcf user and switch to root (su -).
  3. Access the platform database: psql -h localhost -U postgres -d platform
  4. Run the following commands to clear the stale records: truncate vcf_management_component; truncate vrslcm;
  5. Exit the database utility: \q
  6. Restart the SDDC Manager services to ensure the state is refreshed: /opt/vmware/vcf/operationsmanager/scripts/cli/sddcmanager_restart_services.sh
  7. Retry the deployment using a fresh API request or through the SDDC Manager UI by following Deploy VCF Management Components on an NSX Overlay Segment

Additional Information

  •  vcf_management_component table tracks the registration and lifecycle status of integrated components like VCF Operations (formerly vRealize Operations) and VCF Automation. When a deployment is initiated, an entry is created with a status of IN PROGRESS. If the process crashes or is manually killed, the entry remains, blocking the "start" validation of any new deployment workflow.
  • Truncating these tables is safe in this specific scenario because the deployment has not yet completed and no valid configuration has been finalized. This essentially "blanks the slate" for the management component orchestration.
  • After performing the database cleanup and restarting services, it is highly recommended to use a new API request with a unique id or referenceToken if deploying via REST API. Attempting to "Resume" a failed task from the UI may still reference the old task ID which could lead to inconsistent results.
  • Log Locations for Further Troubleshooting:
    • /var/log/vmware/vcf/operationsmanager/operationsmanager.log
    • /var/log/vmware/vcf/sddc-manager-ui-app/sddc-manager-ui-app.log
Powered by Wolken Software