etcd stores network information and states in Kubernetes. Any data loss could be crucial. Clustering etcd is strongly recommended in a production environment. etcd comes with support for clustering; a cluster of N members can tolerate up to (N-1)/2 failures. Typically, there are three mechanisms for creating an etcd cluster. They are as follows:

• Static
• etcd discovery
• DNS discovery

Static is a simple way to bootstrap an etcd cluster if we have all etcd members provisioned before starting. However, it's more common if we use an existing etcd cluster to bootstrap a new member. Then, the discovery method comes into play. The discovery service uses an existing cluster to bootstrap itself. It allows a new member in an etcd cluster to find other existing members. In this recipe, we will discuss how to bootstrap an etcd cluster via static and etcd discovery manually.

We learned how to use kubeadm and kubespray in Chapter 1, Building Your Own Kubernetes Cluster. At the time of writing, HA work in kubeadm is still in progress. Regularly backing up your etcd node is recommended in the official documentation. The other tool we introduced, kubespray, on the other hand, supports multi-nodes etcd natively. In this chapter, we'll also describe how to configure etcd in kubespray.