From: Jon M. <jon...@er...> - 2019-11-05 16:30:23
|
Hi, I searched through the tipc-discussion archives, and found a patch from Ying that explains the problem, and try to remedy it. https://sourceforge.net/p/tipc/mailman/message/29490037/. However, we never moved binding table updates to be sent over broadcast in upstream TIPC, simply because it is working well now, and we saw no reason to add any extra risks. If we are scaling up clusters to > 200 nodes this may become very different, so maybe it is worth trying. I think the chances for doing this successfully are much better now. ///jon |