On nextgenrepl, only one vnode response is required on both fetch and push. This allows the sink workers to cycle through the work much quicker.
This can make it more likely that if there are too many sink workers, they can overload the individual vnode mailboxes (especially when using range_repl). More conservative configurations are now therefore allowed, which will wait for more vnodes and reduce the chance of such overload occurring.
On nextgenrepl, only one vnode response is required on both fetch and push. This allows the sink workers to cycle through the work much quicker.
This can make it more likely that if there are too many sink workers, they can overload the individual vnode mailboxes (especially when using range_repl). More conservative configurations are now therefore allowed, which will wait for more vnodes and reduce the chance of such overload occurring.