Skip to content

Latest commit

 

History

History
212 lines (159 loc) · 7.9 KB

README.md

File metadata and controls

212 lines (159 loc) · 7.9 KB

Miniswarm - Docker Swarm cluster in one command

One command?

Yup, creating a 3 node cluster is simply:

miniswarm start 3

When you're done:

miniswarm delete

What is Miniswarm?

Miniswarm is a tool that intends to make creating and managing a local Docker Swarm cluster as easy as possible. Miniswarm was inspired by Minikube which does a similar thing for kubernetes clusters. See FAQ below for info on managing a remote Swarm cluster.

The tool takes less than 10 minutes to learn, see the tutorial section below, or watch this tutorial video:

Prerequisites

  • Docker >= 1.12
  • docker-machine >= 0.7.0
  • VirtualBox - Needed for local cluster, see FAQ for remote Swarm cluster

Miniswarm: Local Docker Swarm and Docker health check tutorial (less than 10min)

In this tutorial we'll install miniswarm, create a Swarm cluster, deploy some apps and learn all the features of miniswarm in the process.

Install

# As root
curl -sSL https://raw.githubusercontent.com/aelsabbahy/miniswarm/master/miniswarm -o /usr/local/bin/miniswarm
chmod +rx /usr/local/bin/miniswarm

Start a cluster - Pick your desired size

# 1 manager 2 workers
miniswarm start 3

# 1 manager cluster - if you want a smaller cluster
miniswarm start

# 2 managers 3 workers - nice laptop or desktop :)
miniswarm start 2 3

A couple of minutes later, you should get this message

INFO: Stack starup complete. To connect to your stack, run the following command:
INFO: eval $(docker-machine env ms-manager0)

Visualize your cluster

This will open a browser with a nice visualization of your Docker Swarm using docker-swarm-visualizer

miniswarm vis

Deploy your first service - This will initially be failing to showcase healtchecks

This service will be unhealthy due to failing Goss healthchecks and missing dependencies. See next few steps for how we can debug and remedy this.

# Connect to your cluster
eval $(docker-machine env ms-manager0)

# Create a network for your service
docker network create -d overlay healthyvote_net

# Ensure network is set to driver=overlay, scope=swarm
docker network ls

# Create your first service
docker service create -p 8080:80 --replicas 2 --network healthyvote_net --name vote aelsabbahy/healthyvote

Inspect the service health

# Wait for the service to finish preparing, but it won't ever be ready due to failing health
docker service ls
docker service ps vote

# Lets look at the healthchecks using miniswarm
# -a shows all containers, including exited/failed containers
miniswarm health vote -a

We should see something like this:

[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] ======
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] Start: 2016-08-07 22:39:03.748704565 +0000 UTC
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] ======
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] .F
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q]
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] Failures/Skipped:
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q]
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] Title: Redis backend is reachable
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] Meta:
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q]     remedy.1: Deploy redis service if you haven't already
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q]     remedy.2: ctrl-alt-delete
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q]     remedy.3: take a nap aka human ctrl-alt-delete
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] Addr: tcp://redis:6379: reachable:
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] Expected
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q]     <bool>: false
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] to equal
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q]     <bool>: true
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q]
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] Total Duration: 0.500s
[ms-worker0 vote.2.630mookf9pnc8dip5hl7yng3q] Count: 2, Failed: 1, Skipped: 0

Deploy missing dependencies

Lets remedy the healthcheck issue

docker service create --replicas 1 --network healthyvote_net --name redis redis

# Wait for it to show up in `miniswarm vis` or by using CLI
docker service ls
docker service ps redis

# Now that redis is deployed, the  vote service should now be running
docker service ls

# And the health..
miniswarm health vote

Open service in browser

# Open app in browser
miniswarm service vote

# print url, but don't open (--url has to be at the end for now)
miniswarm service vote --url

View the logs

miniswarm logs vote

# Tail the log file (-f has to be at the end for now)
miniswarm logs vote -f

Scale down your cluster

# Take a look at `miniswarm vis` GUI to see the services move around as we scale down
miniswarm scale 2

Delete your cluster

miniswarm delete

FAQ

Can this be used to manage a remote Swarm cluster?

The tool was written with local Swarm cluster in mind. That said, it can probably be used to manage a remote Swarm clusters, but that hasn't been tested. Take a look at MACHINE_DRIVER variable at the top of the script.

In theory, if you set the MACHINE_DRIVER variable to the driver you want + any required variables from the driver itself, see supported drivers it should just work. Feel free to submit a pull-request to improve this or add more support.

How do I disable the color output?

MS_NOCOLOR=1 miniswarm ..

# or
export MS_NOCOLOR=1
miniswarm ...

Does this work on Mac/OSX?

It should, but I don't own a Mac, so I depend on others to verify it. So.. if something is broke on mac, please submit a pull request.

Why did you use Goss for healthchecks?

Mostly shameless self-promotion, and while we're on the topic, check out:

  • Goss - Project page
  • blog post - Using Goss with Docker healthchecks and Kubernetes

Why the #$@^%$ is this written in Bash?

Two reasons:

  1. I though it was going to be ~100 lines of bash, I was wrong.. very wrong :(.
  2. I want users to be able to look at this script and see all the commands needed to set up a Swarm cluster.
  • Go would be great for this tool, especially by leveraging the Docker go packages directly, but then the tool will be more of a blackbox to new users

I'm getting intermittent network issues, why is this happening?

Honestly, I don't know.. I see them too. Either I'm doing something dumb, or Docker Swarm mode has intermittent DNS issues. Hopefully with more users we can get to the bottom of this.

Why is healthyvote not a Docker automated build?

Because Dockerhub doesn't support HEALTHCHECK in Dockerfile yet. The code for this image can be found in the healthyvote/ folder.

Why does this suck?

Because it's a quick hack I did over the weekend.. or I suck.. maybe both?

I tried to use it and got an error

Open an issue, create a pull request.. contribute! :)

Why is the CLI parsing so bad?

See the last two questions.

Social Media:

  • Twitter Follow - Stay updated on new releases
  • Twitter URL - If you like miniswarm, spread the word!
  • Blog - See what I'm ranting about