AWS Redis Failover is the Redis cluster’s fault tolerance mechanism and one of its most important functions.

As part of our AWS Support Service, Bobcares responds to all inquiries, big and small.

Let’s look at AWS Redis Failover in more detail.

AWS Redis Failover

Failover is available in two different modes:

• Failure failover: An automatic recovery cluster is available.
• Artificial failover: A support cluster that is both operational and maintainable.

If we use Redis Cluster Mode to create an ElastiCache cluster, we specify the number of shards in the cluster. Each shard has one primary node (for reads and writes) and a number of replica nodes ranging from zero to five (for reads and failover protection). A cluster can range in size from a single shard with no replicas (1 node) to 15 shards with 5 replicas each (90 total nodes).

When a replica node fails, it quickly detect the failure and replaces the node. When a primary node in a Redis Cluster fails, the cluster detects the failure and promotes a replica node to become the shard’s new primary. The cluster notifies all cluster nodes and clients about the change. This procedure should take no more than 30 seconds. Then replaces the failed node with a replica node and returned to the cluster.

In the ElastiCache Management Console, select engine version 3.2.4 and Cluster Mode enabled to use Redis Cluster with ElastiCache. We can simulate a failure for any node in the ElastiCache cluster using the console or the AWS CLI, and see how the failover process works for our own applications.

We can test failover in both multi-shard and single-shard environments, as well as on Redis 2.8. To do so in the console, go to the Nodes view and select the cluster and shard of your choice, then Failover primary.

We should use this method with caution. Because AWS has no way of knowing whether a cluster is in a development, test, or production role, it works the same on all clusters. Failures on production nodes should only be caused if we’re certain we want to test that production cluster.

[Looking for a solution to another query? We are just a click away.]

Conclusion

To sum up, our Support team went over AWS Redis Failover in greater depth.