• Distributed key-value store that stores the cluster state. ETCD is distributed datastore, which means it replicates the data across all its instances. That way, if one of them crashes, the data is still available on the rest of them.
• When setting up the cluster from scratch, the etcd binary needs to be downloaded and run as a service on the master node. By default it listens on port 2379.
• When setting up the cluster using KubeAdmin, etcd is automatically deployed as a static pod on the master node.
• etcdctl utility can be used to interact with etcd to get and put items in it.
• In a cluster with multiple master nodes (for high availability), the etcd service will be running on each master node and will be communicating with each other on port 2380. This can be configured in the etcd config.

etcdctl

Installation

Follow the release notes for any etcd release to download and unzip the etcd and etcdctl binary from https://github.com/etcd-io/etcd/releases/. At the end, move the extracted binaries to /usr/local/bin or /usr/bin.

Commands

• Backup a cluster’s ETCD

  ETCDCTL_API=3 etcdctl \\
  --cacert=/etc/kubernetes/pki/etcd/ca.crt \\
  --cert=/etc/kubernetes/pki/etcd/server.crt \\
  --key=/etc/kubernetes/pki/etcd/server.key \\
  snapshot save <backup-filename>
  

• Restore a cluster from an ETCD backup

  ETCDCTL_API=3 etcdctl snapshot restore --data-dir="/var/lib/etcd-snapshot" <backup-filename>