Redis Cluster Features
All Redis nodes are interconnected with each other, using binary protocols to optimize bandwidth and transfer speeds internally
The fail of a node is only effective if it is not monitored by more than half of the nodes in the cluster
Client and Redis node direct connection does not need proxy, link any node in the cluster can
Redis-cluster. All nodes are mapped to "0-16383" (hash slots), and cluster is responsible for maintaining node
Cluster fault tolerance and election
(1) Node fault tolerant election when all master is elected, if more than half of the master communication node hangs, the master cluster is considered to be hung off
(2) If the master of any cluster is hung and the master is not slave, the cluster enters the fail state
(3) More than half of the master in the cluster is not available
This article is from the "Brick Blog" blog, please be sure to keep this source http://wsxxsl.blog.51cto.com/9085838/1886041
Redis cluster election