Answer
Check for and kill any lingering IFS-related processes/sessions still running in the VM before attempting another restart — orphaned processes left over from a previous crash can silently prevent services from starting up cleanly even though the targets themselves report as 'running.' When restarting, bring services up in the correct dependency order rather than all at once or in an arbitrary sequence: start the Node Managers first, then the Admin Servers, then the Managed Servers, and finally the HTTP Servers last. Starting components out of this order, or restarting on top of stuck leftover processes from the previous crash, is a common cause of this kind of 'starts fine, then falls over again' instability after an upgrade.