After some more testing, we are going ahead with `streaming = parallel`, with the change to reset `lock_timeout` to the default.
No, as per my understanding it is because parallel apply worker exiting due to lock_timeout set in the test. Ideally, the patch proposed by Kuroda-San should show in LOGs that the parallel worker is exiting due to lock_timeout. Can you try that once?
I don't think I'll have time to compile a patched version and run the tests again, sorry. Have this information in the log probably would have made it easier to work this whole problem out, though, so I hope someone else can test out the change and get it merged.
Thanks for your help Amit, Hou-san, and Hayato-san.