Why should a riak cluster have at least five nodes and five riak clusters?

Why should a riak cluster have at least five nodes and five riak clusters?

Link: http://blog.csdn.net/ddjohn/article/details/43523983

First, ensure that you have correctly installed riak. When you use Riak Fast Track to quickly create three nodes in the development or test environment, we recommend that you deploy at least five nodes in all production environments, to ensure that you benefit from the high availability principle of riak.

 

Deploying five or more nodes will provide the best performance and increase as the cluster expands. because riak is scaled horizontally by adding nodes, you will find that as the cluster grows larger, the performance, reliability, and throughput will get better and better. small-scale deployment may compromise the system's fault tolerance and availability: "sound" replication requirements (three copies by default). In smaller clusters, single point of failure (spof) means that the replication requirements may not be met. This will cause performance degradation and increase the risk of data loss. in addition, clusters with less than five nodes will make the 75% to 100% nodes need to respond to each request, which will lead to unnecessary load and thus reduce performance.

 

Let's further look at the situation of three nodes and four nodes.

 

Fault Tolerance and performance are in clusters with three nodes:

To ensure that the cluster always responds to read/write requests, riak recommends a rational replication policy: three copies are retained on three different nodes. by default, Riak requires at least four nodes to ensure that no single point is used to store multiple copies of any specific data segment. However, it is entirely possible to change the settings to ensure that the three copies exist in the independent nodes in a cluster with three nodes. in addition, you may encounter duplicate placement problems caused by spof or network faults.

 

When a three-node cluster encounters node effectiveness or network partitioning, the default replication requirement is that there are at least three nodes to complete, but only two or one node is working at this time. this causes performance degradation and data loss risks.

 

Fault Tolerance and performance are in clusters with four nodes:

Because replication requires three nodes, in a cluster with four nodes, each request for specific data needs to be responded from 75% to 100% nodes, this may reduce the cluster performance. if you encounter a single point of failure or network partition, you will encounter the problem we mentioned above.

 

What if I change the default replication configuration?

If you want to use the correct number of replicas in your project, make sure that N + 2 nodes are used in the cluster (N is the number of replicas for the above reasons ).

 

If there are five nodes:

When the number of riak cluster nodes increases to five, the percentage of requests responded to by each node in the cluster decreases. this is the basis for horizontal scaling of riak. in case of single point of failure, the number of other nodes is large enough to ensure that your data will not be lost.

 

Therefore, when you use a cluster in a test or development environment, you can use a smaller number of nodes. However, in a production environment, the number of nodes must be greater than or equal to 5.

Http://basho.com/why-your-riak-cluster-should-have-at-least-five-nodes/: