DEV Community: Jonathan Walter

Deploy your containers to AWS or Azure in under 10 minutes (no, really!)

Jonathan Walter — Fri, 28 May 2021 12:57:16 +0000

Did you know you can deploy your existing docker-compose stack to AWS Elastic Container Service or Azure Container Instance using the docker command? Yeah, me neither.

Get the CLI

There is a tech preview of the docker CLI that includes the compose command. This is mostly compatible with the earlier docker-compose.

The tech preview is included in Docker Desktop for Mac and Windows 3.2.1 and above, but if you want to run it on Linux you'll have to install it yourself:

# Docker CLI tech preview linux install
curl -L https://raw.githubusercontent.com/docker/compose-cli/main/scripts/install/install_linux.sh | sh

Create context

To manage our cloud deployments we'll have to set up a new docker context.

Using AWS

docker context create ecs some-name-here

You can use your AWS credentials in one of three ways. If you alreade have the AWS CLI configured, you can use your credentials from there. Otherwise you can input them when asked or use environment variables. Make sure your IAM role has the correct premissions.

? Create a Docker context using:  [Use arrows to move, type to filter]
> An existing AWS profile
AWS secret and token credentials
AWS environment variables

Using Azure

The procedure for Azure is a little bit different. You must first tell docker to log in to Azure:

docker login azure

This will open a web page where you enter your credentials, or fall back to using the Azure device code flow. Note that this is separate from the Azure CLI login.

This can be done without interaction as well, which is handy for CI workflows:

docker login azure --client-id xx --client-secret yy --tenant-id zz

You can use the --tenant-id option alone to specify a tenant, if you have multiple ones.

To create the actual docker context we run:

docker context create aci some-name-here

This command will let you pick a resource group to use, or create a new one. Again, to run this without interaction we can specify --subscription-id, --resource-group, and --location.

Activate your context

Once your context is created you can view it by running docker context ls.

To activate the context run:

docker context use some-name-here

Deploy your stack

Now that we have the basics covered, we are ready to deploy stuff!

It is almost too simple, just run docker compose up as you normally would. Docker will automagically create all necessary resources in AWS or Azure and deploy your container(s). Similarly, docker compose down will remove everything.

But I want to do more advanced things!

Fret not. The Docker CLI has support for many more things, such as volumes, secrets, auto scaling, DNS labels, health checks and so on...

I highly recommend you read the official documentation/guides for AWS and Azure deployment. But that will probably take more than 10 minutes.

Jonathan Walter is a Media Consultant at Eyevinn Tecnology, the European leading independent consultancy firm specializing in video technology and media distribution.

Image is a montage of photos by Pero Kalimero and Todd Cravens on Unsplash

asdf - the final(?) version manager

Jonathan Walter — Mon, 26 Apr 2021 12:43:21 +0000

Introduction

Have you ever had the need to run multiple different versions of Nodejs? -Of course, silly! We have nvm for that.

How about python, then? -Yeah, we use pyenv.

Java? -Yup. SDKman.

Ruby? -rvm.

Tmux? -Um...

Vim? -But why?!

If you are (or pretending to be) a polyglot software developer, you'll need a whole lot of versioning managers and a whole lot of different commands and syntax to remember.

Wouldn't it be nice if there was one tool that could handle everything? There is - asdf!

Installation

asdf is available on both macOS and Linux. The examples in this article will use macOS.

Dependencies

asdf requires you to have corutils, curl and git installed. We can install them using Homebrew:

brew install coreutils curl git

We'll install asdf itself with Homebrew too:

brew install asdf

Next we need to add asdf.sh to our .zshrc:

echo -e "\n. $(brew --prefix asdf)/asdf.sh" >> ${ZDOTDIR:-~}/.zshrc

(if you're using oh-my-zsh there is a plugin for asdf)

Now we're ready to start using it!

Getting started with `asdf`

These are the basic principles for using asdf. Be sure to check out the official documentation as well.

Adding plugins and versions

asdf has a long list of plugins for different languages and tools.

To add python we simply execute:

asdf plugin add python

asdf is now python-aware, and we can install multiple versions of it.

To get the latest one (3.9.4 when I wrote this):

asdf install python latest

To get a specific version:

asdf install python 3.8.7

Setting versions

Now that we have a couple of versions of Python installed we should set the default global version:

asdf global python 3.9.4

We can also set the global version to the operating systems version of python (not managed by asdf):

asdf global python system

To set a version to be used in the current directory (and any sub-directories) we run:

asdf local python 3.8.7

We can even set a specific version for the current shell:

asdf shell python 3.8.7

Finding out what versions are active is as simple as running:

asdf current

The `.tool-versions` file

To keep track of what versions to use, asdf creates a .tool-versions file. For global settings this will be located at $HOME/.tool-versions and for local versions it'll be put in your current directory.

This is just a text file describing what versions to use and it can look something like this:

nodejs 16.0.0
python 3.9.4

It is also possible to have several versions defined:

python 3.9.4 3.8.7 system

Coming from other version managers

"I already use nvm and pyenv, and have a million .nvmrc, .node-version and .python-version files sprinkled through-out my system!", you might say. asdf can take advantage of these files if we tell it to.

We need to create a file in your home directory called .asdfrc with the following content:

legacy_version_file = yes

This will enable support for reading version files used by other version managers.

Conclusion

That is pretty much all there is to it. It is super handy to just cd into a directory and have asdf automatically switch to the correct version of the tools needed for that particular project.

It also integrates very nicely with oh-my-zsh and powerline10k.