How to use etcd (multi-machine cluster basic mode) in Ubuntu?

How to use etcd (multi-machine cluster basic mode) in Ubuntu?

Recently, I have found interesting solution. it is called as etcd. It is Distributed reliable key-value store. What is the key-value store?. It is similar with dictionary in Python. This etcd make dictionary items can be put and get on any hosts. 

1. Prerequesite

The etcd is used for cluster environment. Therefore, I will prepare 3 servers with ubuntu.

infra0     172.22.0.96 infra0.example.com infra1     172.22.0.33 infra1.example.com infra2     172.22.0.133 infra2.example.com

2. Get etcd release 

I have ubuntu server. I need the etcd release for ubuntu. In this documentation. There is installation guide for Linux.

ETCD_VER=v3.3.10 # choose either URL GOOGLE_URL=https://storage.googleapis.com/etcd GITHUB_URL=https://github.com/etcd-io/etcd/releases/download DOWNLOAD_URL=${GOOGLE_URL} rm -f /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz rm -rf /tmp/etcd-download-test mkdir -p /tmp/etcd-download-test curl -L ${DOWNLOAD_URL}/${ETCD_VER}/etcd-${ETCD_VER}-linux-amd64.tar.gz -o /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz tar xzvf /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz -C /tmp/etcd-download-test --strip-components=1 rm -f /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz /tmp/etcd-download-test/etcd --version ETCDCTL_API=3 /tmp/etcd-download-test/etcdctl version

I will run this command on every hosts. "/tmp" directory is used in command above. However, In my case, I will change the directory "/home/ubunut"

# mkdir -p /home/ubuntu/etcd-download-test # curl -L ${DOWNLOAD_URL}/${ETCD_VER}/etcd-${ETCD_VER}-linux-amd64.tar.gz -o  # /home/ubuntu/etcd-${ETCD_VER}-linux-amd64.tar.gz # tar xzvf /home/ubuntu/etcd-${ETCD_VER}-linux-amd64.tar.gz -C /home/ubuntu/etcd-download-test --strip-components=1 # rm -f /home/ubuntu/etcd-${ETCD_VER}-linux-amd64.tar.gz

I can check the installation version with both of commands.

/home/ubuntu/etcd-download-test/etcd --version ETCDCTL_API=3 /home/ubuntu/etcd-download-test/etcdctl version

# /home/ubuntu/etcd-download-test/etcd --version etcd Version: 3.3.10 Git SHA: 27fc7e2 Go Version: go1.10.4 Go OS/Arch: linux/amd64

More efficient, I will export the directory in PATH environment.

# export PATH=$PATH:/home/ubuntu/etcd-download-test # vi /root/.bashrc PATH=$PATH:/home/ubuntu/etcd-download-test # source /root/.bashrc

3. Running in multi-machine cluster basic mode

I can run in single standalone mode. The command line usage is here. At first, run in single mode

# etcd

I can put and get data. "ETCDCTL_API=3" is important parameter. It should be exist.

# ETCDCTL_API=3 etcdctl put mykey "this is awesome" OK # ETCDCTL_API=3 etcdctl get mykey mykey this is awesome

Also I can see the result with JSON format like below. If I want to find more example of command usage, look this documentation.

# export ETCDCTL_API=3 # etcdctl --endpoints=localhost:2379 put foo "Hello World!" OK # etcdctl --endpoints=localhost:2379 get foo --write-out="json" {"header":{"cluster_id":14841639068965178418,"member_id":10276657743932975437,"revision":3,"raft_term":2},"kvs":[{"key":"Zm9v","create_revision":3,"mod_revision":3,"version":1,"value":"SGVsbG8gV29ybGQh"}],"count":1} # etcdctl --endpoints=localhost:2379 get foo foo Hello World!

Even if this single mode works, but etcd is more powerful in mult-machin cluster mode. Look this documentation. To run cluster mode, there are 2 things important. IP address to listen and announce and IP address for peer hosts. 

etcd --name infra0 --initial-advertise-peer-urls http://172.22.0.96:2380 \   --listen-peer-urls http://172.22.0.96:2380 \   --listen-client-urls http://172.22.0.96:2379,http://127.0.0.1:2379 \   --advertise-client-urls http://172.22.0.96:2379 \   --initial-cluster-token etcd-cluster-1 \   --initial-cluster infra0=http://172.22.0.96:2380,infra1=http://172.22.0.33:2380,infra2=http://172.22.0.133:2380 \   --initial-cluster-state new etcd --name infra1 --initial-advertise-peer-urls http://172.22.0.33:2380 \   --listen-peer-urls http://172.22.0.33:2380 \   --listen-client-urls http://172.22.0.33:2379,http://127.0.0.1:2379 \   --advertise-client-urls http://172.22.0.33:2379 \   --initial-cluster-token etcd-cluster-1 \   --initial-cluster infra0=http://172.22.0.96:2380,infra1=http://172.22.0.33:2380,infra2=http://172.22.0.133:2380 \   --initial-cluster-state new etcd --name infra2 --initial-advertise-peer-urls http://172.22.0.133:2380 \   --listen-peer-urls http://172.22.0.133:2380 \   --listen-client-urls http://172.22.0.133:2379,http://127.0.0.1:2379 \   --advertise-client-urls http://172.22.0.133:2379 \   --initial-cluster-token etcd-cluster-1 \   --initial-cluster infra0=http://172.22.0.96:2380,infra1=http://172.22.0.33:2380,infra2=http://172.22.0.133:2380 \   --initial-cluster-state new

  

For peer communicate, the 2380 port is used. For data communication, the 2379 port is used. Therefore, "127.0.0.1:2379" is necessary to request to localhost. However, this is not necessary to announce. These command above should be run on each host. After running, I can get the status of clustering.

# ETCDCTL_API=3 etcdctl --endpoints=172.22.0.96:2379,172.22.0.33:2379,172.22.0.133:2379 endpoint health 172.22.0.96:2379 is healthy: successfully committed proposal: took = 2.385219ms 172.22.0.133:2379 is healthy: successfully committed proposal: took = 2.847419ms 172.22.0.33:2379 is healthy: successfully committed proposal: took = 2.292504ms

Now, I have done everything to cluster multi-hosts. In multi-hosts environments, I put data into anywhere, I can get data from anywhere.

# ETCDCTL_API=3 etcdctl --endpoints=172.22.0.96:2379 put foo "Hello World!" OK # ETCDCTL_API=3 etcdctl --endpoints=172.22.0.133:2379 get foo foo Hello World!

Finally, I can use the "etcd" little.

 Reference

[ 1 ] https://github.com/etcd-io/etcd

[ 2 ] https://github.com/etcd-io/etcd/releases

[ 3 ] https://github.com/etcd-io/etcd/blob/master/Documentation/op-guide/clustering.md

[ 4 ] https://coreos.com/etcd/docs/latest/demo.html