Unless a better solution comes along for N way locking in a future, I think this would remain useful for us.
Updating the architecture to where we avoid this synchronization would actually be more painful in many cases, and the alternative is to not allow many of our manager (disk/peer/handshake) sinks to be cloneable (which is painful for clients to deal with).
Unless a better solution comes along for N way locking in a future, I think this would remain useful for us.
Updating the architecture to where we avoid this synchronization would actually be more painful in many cases, and the alternative is to not allow many of our manager (disk/peer/handshake) sinks to be cloneable (which is painful for clients to deal with).