ElastiCache For Redis Cluster Chuang

As a cache database, Redis is highly respected in the Internet industry. With many unique advantages, it can solve the caching needs of many business architectures.

For Redis products, AWS launched Elasticache for Redis for PAAS services. Today, let's take a brief look at how to create an Elasticache for Redis Cluster.

And test how much time it will take for redis to refragment.

About ElastiCache Scaling links

Https://docs.aws.amazon.com/zh_cn/AmazonElastiCache/latest/red-ug/redis-cluster-resharding-online.html

1. The cluster has a unified endpoint that provides client access

two。 Three sharding are created, followed by sharding, on which node, there are a total of 16384 hash slot.

3. Click to enter a sharding, you can see that there are three nodes below, a group of two slaves, but now the page does not show which is the master and that is the slave. We can also access it through the endpoint below, but if the machine goes down, or if the slot you need to access is not in this sharding, the access will fail. Therefore, be sure to use the total redis endpoint for access.

4. Add sharding instances

5. It takes 13 minutes to actually complete (this is an empty redis cluster, if it is a production environment, the more data, the slower the sharding)

6. Delete shard operation

7. Delete operation for 10 minutes, not completed

8. Delete operation completed, total time 17 minutes

To sum up, redis as a cache database, general scenarios, memory consumption, cache data will not be very large. In other words, the amount of cached data is relatively stable. In the process of add shards and delete shards, there is a long modifying time, which can not be ignored for the performance impact of redis such a cached database. Therefore, it is recommended that expansion and contraction operations should be performed during the idle period of the business.