A warm VM migration from VMware is stuck for hours, logging the same message more than 10K times per second

$ oc logs importer-prime-22863568-343e-436e-99e2-8889fb71f944-checkpoint-snapshot-12585 -n test -c importer

I1031 11:53:44.972321 1 vddk-datasource_amd64.go:611] No new block status data at offset 6806175744

Source:

vSphere 7.0.3, build 21290409

1 RHEL 7.9 VM with 1 disk of 30 GB (using XFS).

CBT is enabled for both the VM and the disk.

The VM was running the following disk stressor for 5-10 min before the cutover:

stress-ng --hdd 4 --timeout=60m

Destination:

OCP 4.15.36

OCP-V 4.15.6

MTV 2.7.2

VDDK image version 8.0.3

Results:

The migration has been stuck in this state for more than 4 hours logging the same message all the time. The offset doesn't change.

The pod is using 100% of a CPU all the time.

Expected results:

I'd expect either that the migration finishes successfully or it fails gracefully.