Hi Peter,
Would you mind reporting a bug on this issue you saw?
To me, it looks like sync/timing issue between snapshot state('OK) and disk state('Locked').
Can you also please check if the disk that is reported as locked is indeed locked at the time the error is reported?
Please add this info to the bug and all relevant scripts.
As a workaround, please consider adding an additional wait for all the VM disks to also reach 'OK' state.
Thanks,
Avihai