Test-failover for a VM is stuck in "Failover test ready state" and failed in vCDA 4.7.x

Products

• VMware Cloud Director

Issue/Introduction

• Test-failover for a VM gets stuck at "Failover test ready state" for more than 8 hours and fails with an error:

"Failed to reload vm 'vm-7xxx8' from path '[Cxxx2_DR_Tx_001] C4-axxxxxxxxxxc-4xxd-bxx5-4xxxxxxxx8/PxD-Dx-0x.vmx' for vc '4xxxxxxxxx2-4xxb-axx9-exxxxxxxxxx4'." 

• It also starts to perform the cleanup on its own and fails with an error -

"The specified key ,name or identifier 'VM_NAME' already exists." 

Environment

VMware Cloud Director Availability 4.7.x

Cause

This is caused due to the CBT enabled at the source vCDA applaince. VMware Cloud Director Availability instances are not compatible with CBT in the source site. 

 

Resolution

Disable the CBT flag in VCDA configuration so that it does not try to enable CBT during test failover.

The steps to disable CBT flag is as follows:

     (a) SSH to the destination replicator appliance & navigate to directory : /opt/vmware/h4/replicator/config
     (b) Backup application.properties file using command: cp application.properties application.properties.backup
     (c) Run command: vi application.properties
     (d) Add parameter: vm.failover.disable_cbt=true (NOTE: no comma, quotes etc & all small case).
     (e) Restart the replicator service using command: systemctl restart replicator

Additional Information

Instances cannot interoperate with Changed Block Tracking (CBT) enabled for the source virtual machine. For more information, see Replicating Other Storage.

For information about the instances, see Using instances.