VCF 9.0 vCenter Deployment Stalled during the vSAN Bootstrap phase Due to Partitioned Capacity Disk on ESXi Host
search cancel

VCF 9.0 vCenter Deployment Stalled during the vSAN Bootstrap phase Due to Partitioned Capacity Disk on ESXi Host

book

Article ID: 438975

calendar_today

Updated On:

Products

VMware vSAN

Issue/Introduction

During a greenfield VMware Cloud Foundation (VCF) 9.0 deployment, the vCenter installation fails during the vSAN Bootstrap phase.

  • SDDC Manager UI Error: vCenter installation failed. Check logs under /var/log/vmware/vcf/domainmanager/ci-installer...

  • Log Evidence: In the vcsa-cli-installer.log, the following fatal exception is observed:

tasking.taskflow.TaskExecutionFailureException: VSANBootstrapTask: The number of cache disks are 6 and the number of capacity disks are 3. The number of capacity disks must be greater than the number of cache disks.

Environment

VMware Cloud Foundation 9.x

VMware vSAN 9.x (Original Storage Architecture - OSA)

Cause

This issue occurs when one or more intended vSAN capacity disks on an ESXi host are partitioned (e.g., formatted as a local VMFS datastore).

VCF 9.0 automation requires all vSAN-eligible disks to be completely raw and untagged. If a disk contains an existing partition table:

  1. The ESXi storage stack flags the device as Ineligible for use by vSAN.

  2. The VCF automated disk-claiming algorithm skips the ineligible disk, resulting in an incorrect count of available capacity vs. cache drives.

  3. The algorithm then attempts an invalid disk group ratio (e.g., more cache than capacity), violating vSAN architectural requirements and causing the bootstrap task to fail.

In this specific scenario, a 2TB NVMe drive was formatted as "Local Datastore" and was actively hosting the VCF appliance VM.

Resolution

To resolve this, the ineligible disk must be cleared of all partitions. Note: If the disk is hosting a virtual machine, you must relocate the VM first.

Step 1: Relocate Active Virtual Machines

Before clearing the disk, verify if any VMs are residing on the local datastore (e.g., CCA Local Datastore).

  1. Log into the ESXi Host Client.

  2. Perform a Storage vMotion to migrate any active VMs to a different datastore (e.g., the local datastore1 on the boot drive).

Step 2: Delete the Local Datastore

  1. Navigate to Storage > Datastores.

  2. Right-click the offending datastore and select Delete.

Step 3: Clear the Partition Table

  1. Navigate to Storage > Devices.

  2. Select the affected NVMe/SSD device.

  3. Click Clear partition table.

    • Alternative (CLI): Run esxcli vsan storage remove -s <device_ID> via SSH.

Step 4: Verify Disk Eligibility

Run the following command to ensure the disk is now seen as eligible by vSAN: vdq -q -H The disk should now show State: Eligible for use by VSAN and Reason: None.

Step 5: Retry Deployment

  1. In SDDC Manager, cleanup the failed deployment attempt.

  2. Restart the vCenter deployment workflow. Do not manually tag the disks via CLI, as VCF must handle this automatically per KB 384486.

Additional Information

Powered by Wolken Software