When performing a Test Failover on a specific Virtual Machine in VMware Cloud Director Availability (vCDA), the task fails. The vCDA UI or task list displays an error similar to the following:
Failed to update RT VM <UUID1> mode to REPLICATION_MODE (Cannot update the replication tracking mode of virtual machine '<UUID1>' to 'REPLICATION_MODE')
Upon reviewing the vCDA / Cloud service backend logs, the following critical errors can be observed:
ERROR - [UI-...] com.vmware.h4.cloud.service.RtVmService : Failed to update RT VM <UUID1> mode to REPLICATION_MODEcom.vmware.vcloud.client.exception.VcloudException: (Major code = 500, minor code = INTERNAL_SERVER_ERROR) - [...] No row with the given identifier exists: [com.vmware.vcloud.common.model.logicalresource.VdcLogicalResourceModel#<UUID2>]
VMware Cloud Director 10.6.1
VMware Cloud Director Availability 4.7.2
This issue occurs due to an inconsistent Storage Policy record within the VMware Cloud Director (VCD) database.
Specifically, the storage policy ID (storage_class_lr_id) registered for the VM disk in the vapp_vm_disk_storage_class table has been deleted or no longer exists in the VCD logical resource table (vdc_logical_resource). Additionally, the datastore currently hosting the VM does not match this stale registered storage policy. When VCDA attempts to update the Replication Tracking (RT) VM mode, the system fails to locate the corresponding storage resource using this invalid ID, resulting in an INTERNAL_SERVER_ERROR.
Resolving this issue requires direct modifications to the underlying database that may carry a high risk of causing data corruption or severe system failure if executed incorrectly. To ensure the safety and integrity of environment, open a Support Request with Broadcom Technical Support for assisted remediation. see Creating and managing Broadcom support cases