etcd-embedded

command module

v0.0.0-...-355bce0 Latest Latest Go to latest Published: Sep 18, 2021 License: MIT Imports: 1 Imported by: 0

Details

Valid go.mod file

The Go module system was introduced in Go 1.11 and is the official dependency management solution for Go.
Redistributable license

Redistributable licenses place minimal restrictions on how software can be used, modified, and redistributed.
Tagged version

Modules with tagged versions give importers more predictable builds.
Stable version

When a project reaches major version v1 it is considered stable.
Learn more about best practices

Repository

github.com/balchua/etcd-embedded

Links

Open Source Insights

README ¶

Embedded etcd

This project is to simulate embedding etcd with high availability.

How to

Before you start, you to satisfy these pre-requisites

etcdctl - https://github.com/etcd-io/etcd/releases
lxd - This will be used to simulate multi node etcd
Must have these available ports 2379, 2380

Build

Simply run go build .

Setting up the lxd

We will start a 3 node LXD instance. Launch an LXD instance

lxc launch -p default -p microk8s ubuntu:20.04 mk8s-0
lxc launch -p default -p microk8s ubuntu:20.04 mk8s-1
lxc launch -p default -p microk8s ubuntu:20.04 mk8s-2

Copy the default config and binary lxd instance

lxc file push default-config.yaml mk8s-0/home/ubuntu/config.yaml
lxc file push default-config.yaml mk8s-1/home/ubuntu/config.yaml
lxc file push default-config.yaml mk8s-2/home/ubuntu/config.yaml

lxc file push etcd-embedded mk8s-0/home/ubuntu/etcd-embedded
lxc file push etcd-embedded mk8s-1/home/ubuntu/etcd-embedded
lxc file push etcd-embedded mk8s-2/home/ubuntu/etcd-embedded

Get all lxd instance IPs

lxc list

+-----------------------+---------+-----------------------+------+-----------+-----------+
|         NAME          |  STATE  |         IPV4          | IPV6 |   TYPE    | SNAPSHOTS |
+-----------------------+---------+-----------------------+------+-----------+-----------+
| mk8s-0                | RUNNING | 10.124.129.137 (eth0) |      | CONTAINER | 0         |
+-----------------------+---------+-----------------------+------+-----------+-----------+
| mk8s-1                | RUNNING | 10.124.129.151 (eth0) |      | CONTAINER | 0         |
+-----------------------+---------+-----------------------+------+-----------+-----------+
| mk8s-2                | RUNNING | 10.124.129.8 (eth0)   |      | CONTAINER | 0         |
+-----------------------+---------+-----------------------+------+-----------+-----------+

First start (`mk8s-0`) the main server

lxc exec mk8s-0 -- /home/ubuntu/etcd-embedded server /home/ubuntu/config.yaml

Add the second node (`mk8s-1`) as a learner.

lxc exec mk8s-1 -- /home/ubuntu/etcd-embedded join http://10.124.129.137:2380 /home/ubuntu/config.yaml

Before starting the second node (mk8s-1), check that the node is added as a learner.


./etcdctl --endpoints=10.124.129.137:2379,10.124.129.151:2379,10.124.129.8:2379 member list -w table
+------------------+-----------+--------+----------------------------+----------------------------+------------+
|        ID        |  STATUS   |  NAME  |         PEER ADDRS         |        CLIENT ADDRS        | IS LEARNER |
+------------------+-----------+--------+----------------------------+----------------------------+------------+
| 58a3d8673d099781 |   started | mk8s-0 | http://10.124.129.137:2380 | http://10.124.129.137:2379 |      false |
| dd94fc49fe0983b8 | unstarted |        | http://10.124.129.151:2379 |                            |       true |
+------------------+-----------+--------+----------------------------+----------------------------+------------+

Then start the second node (mk8s-1)

lxc exec mk8s-1 -- /home/ubuntu/etcd-embedded server /home/ubuntu/config.yaml

Wait for a few seconds, the promotion to a voter node is automatic.

./etcdctl --endpoints=10.124.129.137:2379,10.124.129.151:2379,10.124.129.8:2379 member list -w table
+------------------+---------+--------+----------------------------+----------------------------+------------+
|        ID        | STATUS  |  NAME  |         PEER ADDRS         |        CLIENT ADDRS        | IS LEARNER |
+------------------+---------+--------+----------------------------+----------------------------+------------+
| 58a3d8673d099781 | started | mk8s-0 | http://10.124.129.137:2380 | http://10.124.129.137:2379 |      false |
| dd94fc49fe0983b8 | started | mk8s-1 | http://10.124.129.151:2379 | http://10.124.129.151:2379 |      false |
+------------------+---------+--------+----------------------------+----------------------------+------------+

Add the third node (`mk8s-2`) as a learner.

lxc exec mk8s-2 -- /home/ubuntu/etcd-embedded join http://10.124.129.137:2380 /home/ubuntu/config.yaml
{"level":"info","ts":1631414981.2671242,"caller":"util/network.go:21","msg":"Interface is a loopback","name":"lo"}
{"level":"info","ts":1631414981.2682645,"caller":"util/network.go:30","msg":"Interface details","name":"eth0","isUp":true}
{"level":"info","ts":1631414981.2686086,"caller":"util/network.go:34","msg":"Interface name"}
{"level":"info","ts":1631414981.2686648,"caller":"util/network.go:37","msg":"IP found","ip":"10.124.129.8"}
{"level":"info","ts":1631414981.2689445,"caller":"etcd/learner.go:16","msg":"AddMemberAsLearner","peerUrl":"http://10.124.129.8:2379","LeaderEndpoint":"http://10.124.129.137:2380"}
{"level":"info","ts":1631414981.2927167,"caller":"cmd/join.go:41","msg":"Join Successful.","PeerURL":["http://10.124.129.8:2379"],"MemberId":2288430810285401478,"ClientURL":[],"IsLearner":true}

Before starting the third node (mk8s-2), check that the node is added as a learner.

./etcdctl --endpoints=10.124.129.137:2379,10.124.129.151:2379,10.124.129.8:2379 member list -w table
+------------------+-----------+--------+----------------------------+----------------------------+------------+
|        ID        |  STATUS   |  NAME  |         PEER ADDRS         |        CLIENT ADDRS        | IS LEARNER |
+------------------+-----------+--------+----------------------------+----------------------------+------------+
| 1fc223b2841ee586 | unstarted |        |   http://10.124.129.8:2379 |                            |       true |
| 58a3d8673d099781 |   started | mk8s-0 | http://10.124.129.137:2380 | http://10.124.129.137:2379 |      false |
| dd94fc49fe0983b8 |   started | mk8s-1 | http://10.124.129.151:2379 | http://10.124.129.151:2379 |      false |
+------------------+-----------+--------+----------------------------+----------------------------+------------+

Then start the third node (mk8s-2)

lxc exec mk8s-2 -- /home/ubuntu/etcd-embedded server /home/ubuntu/config.yaml

Check that the third node (mk8s-2) is all part of the cluster

./etcdctl --endpoints=10.124.129.137:2379,10.124.129.151:2379,10.124.129.8:2379 member list -w table
+------------------+---------+--------+----------------------------+----------------------------+------------+
|        ID        | STATUS  |  NAME  |         PEER ADDRS         |        CLIENT ADDRS        | IS LEARNER |
+------------------+---------+--------+----------------------------+----------------------------+------------+
| 1fc223b2841ee586 | started | mk8s-2 |   http://10.124.129.8:2379 |   http://10.124.129.8:2379 |      false |
| 58a3d8673d099781 | started | mk8s-0 | http://10.124.129.137:2380 | http://10.124.129.137:2379 |      false |
| dd94fc49fe0983b8 | started | mk8s-1 | http://10.124.129.151:2379 | http://10.124.129.151:2379 |      false |
+------------------+---------+--------+----------------------------+----------------------------+------------+

Now you have a working etcd HA cluster.

Removing a node from the cluster

Go to any node and execute the following:

lxc exec mk8s-0 -- /home/ubuntu/etcd-embedded remove http://10.124.129.137:2379 mk8s-2 /home/ubuntu/config.yaml

Arguments are:

the endpoint of the leader node
the name of the node you wish to remove
the configuration file

Inspect the member list using etcdctl

./etcdctl --endpoints=10.124.129.137:2379,10.124.129.151:2379,10.124.129.8:2379 endpoint status -w table
{"level":"warn","ts":"2021-09-12T11:34:38.345+0800","logger":"etcd-client","caller":"v3/retry_interceptor.go:62","msg":"retrying of unary invoker failed","target":"etcd-endpoints://0xc0000d2a80/#initially=[10.124.129.137:2379;10.124.129.151:2379;10.124.129.8:2379]","attempt":0,"error":"rpc error: code = DeadlineExceeded desc = latest balancer error: last connection error: connection error: desc = \"transport: Error while dialing dial tcp 10.124.129.8:2379: connect: connection refused\""}
Failed to get the status of endpoint 10.124.129.8:2379 (context deadline exceeded)
+---------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+
|      ENDPOINT       |        ID        | VERSION | DB SIZE | IS LEADER | IS LEARNER | RAFT TERM | RAFT INDEX | RAFT APPLIED INDEX | ERRORS |
+---------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+
| 10.124.129.137:2379 | 58a3d8673d099781 |   3.5.0 |   20 kB |      true |      false |         2 |         11 |                 11 |        |
| 10.124.129.151:2379 | 71c4859c30bb0e6e |   3.5.0 |   20 kB |     false |      false |         2 |         11 |                 11 |        |
+---------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+

The error you see above is normal, since the remove node is automatically goes down and it is no longer listening to the ports.

Endpoint

The app tries to add simple endpoint to show the etcd member nodes.

http://localhost:3000/members

etcdctl helpful commands

./etcdctl --endpoints=10.124.129.137:2379,10.124.129.151:2379,10.124.129.8:2379 member list -w table
+------------------+---------+--------+----------------------------+----------------------------+------------+
|        ID        | STATUS  |  NAME  |         PEER ADDRS         |        CLIENT ADDRS        | IS LEARNER |
+------------------+---------+--------+----------------------------+----------------------------+------------+
| 1fc223b2841ee586 | started | mk8s-2 |   http://10.124.129.8:2379 |   http://10.124.129.8:2379 |      false |
| 58a3d8673d099781 | started | mk8s-0 | http://10.124.129.137:2380 | http://10.124.129.137:2379 |      false |
| dd94fc49fe0983b8 | started | mk8s-1 | http://10.124.129.151:2379 | http://10.124.129.151:2379 |      false |
+------------------+---------+--------+----------------------------+----------------------------+------------+

Check the status of the endpoints, shows who is the leader

./etcdctl --endpoints=10.124.129.137:2379,10.124.129.151:2379,10.124.129.8:2379 endpoint status -w table
+---------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+
|      ENDPOINT       |        ID        | VERSION | DB SIZE | IS LEADER | IS LEARNER | RAFT TERM | RAFT INDEX | RAFT APPLIED INDEX | ERRORS |
+---------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+
| 10.124.129.137:2379 | 58a3d8673d099781 |   3.5.0 |   20 kB |      true |      false |         2 |         10 |                 10 |        |
| 10.124.129.151:2379 | dd94fc49fe0983b8 |   3.5.0 |   20 kB |     false |      false |         2 |         10 |                 10 |        |
|   10.124.129.8:2379 | 1fc223b2841ee586 |   3.5.0 |   20 kB |     false |      false |         2 |         10 |                 10 |        |
+---------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+