Test failover job took longer time than expectation when CBT enabled on source VM

Products

• VMware Cloud Director Availability - Disaster Recovery 4.x

Issue/Introduction

• Test failover took around 10 mins or more to complete.
• Test failover benchmark takes a long time and it's impacting RTO.
• On source site, the VM was configured with Change Block Tracking (CBT) enabled.
• On target site vCenter Server, the "reconfigure VM" task took around 10 mins or more to complete, which is the most time-consuming contributor to the test failover job.
• On target site, on target ESXi host, in /var/log/hostd.log, there are entries as below, indicating CBT (Change Block Tracking) was configured from "false" to "true":

  YYYY-MM-DDTHH:MM:SS Db(167) Hostd[2104656]: [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/<datastore uuid>/C4-<uuid>/test/<VM Name>.vmx opID=UI-<opid> sid=<sid> user=vpxuser:<username>] disk cbt old state false -> new state true

Environment

VMware Cloud Director Availability 4.7.x

Cause

This is expected behavior. VMware Cloud Director Availability(VCDA) instances are not compatible with CBT in the source site.

Refer to Using instances where it states:  

"Instances cannot interoperate with Changed Block Tracking (CBT) enabled for the source virtual machine"

Resolution

Disable CBT on the virtual machine on the the source site. Refer to: Enabling or disabling Changed Block Tracking (CBT) on virtual machines