Dear Vignesh,
> One key point to consider is that the lock on transaction objects will
> be held for a longer duration when using streaming in parallel. This
> occurs because the parallel apply worker initiates the transaction as
> soon as streaming begins, maintaining the lock until the transaction
> is fully completed. As a result, for long-running transactions, this
> extended lock can hinder concurrent access that requires a lock.
So, the current default is the most conservative and can run anywhere; your
proposal is a bit optimistic, right? Since long transactions should be avoided
in any case, and everyone knows it, I agree with your point.
One comment for your patch;
Shouldn't we add a streaming=off case for pg_dump test? You added lines but no one
passes it.
Best regards,
Hayato Kuroda
FUJITSU LIMITED
