Redis Cluster Building Management

Redis 3.0 did not support cluster functionality, generally by combining twemproxy (Twitter) or codis (Pea pod use) to achieve distributed clustering, performance and stability are good, but the configuration and management is still complicated. The newly released version 3.0 now has a default built-in support cluster feature. Redis ensures cluster availability through partitioning, and even if some nodes fail or fail to connect, the cluster continues to process customer requests.

Redis clusters are implemented using data shards rather than consistent hashes. The cluster contains a total of 16,384 hash slots, each of which allocates a portion of the hash slot, each of which is stored in one of the slots. This also means that if the cluster does not turn on master-slave replication, any one of these nodes is down or disconnected, and the keys stored in those slots will not be accessible. So when you open a Redis cluster, it's a good idea to turn on replication to ensure that when the primary node fails or the connection is unsuccessful, the slave node corresponding to the master node can be replaced by the master node.

The following is an introduction to the construction of Redis cluster management.

I. installing Redis, common scripting address: http://dongsong.blog.51cto.com/916653/1649590

Redis cluster configuration file redis.conf content, and Redis management scripts in the same directory:

# redis.confport 6380cluster-enabled yescluster-config-file nodes.confcluster-node-timeout 5000appendonly Yesdaemonize Yeslogfile Redis.log

Mkdir-p/app/srccd/app/srcwget http://download.redis.io/releases/redis-3.0.1.tar.gzsh Redis Install

This article is from the "Pine" blog, be sure to keep this source http://dongsong.blog.51cto.com/916653/1649680

Redis Cluster Building Management