I applied the patch for the parallel hash and ran that against my production data. The query worked with parallel workers. That patch was applied on top of the earlier patch for the invalid Susa memory alloc that started this thread.
From my view as a user, these are both fixed and can be marked as such. We will wait for the patches to role through the release cycles.
If you change the patch and need another test with real data please let me know.
Thank you for your time and effort on these issues.