Ruby 3.0

Description

Ruby 3.0 available as container is a base platform for building and running various Ruby 3.0 applications and frameworks. Ruby is the interpreted scripting language for quick and easy object-oriented programming. It has many features to process text files and to do system management tasks (as in Perl). It is simple, straight-forward, and extensible.

This container image includes an npm utility, so users can use it to install JavaScript modules for their web applications. There is no guarantee for any specific npm or nodejs version, that is included in the image; those versions can be changed anytime and the nodejs itself is included just to make the npm work.

Usage

For this, we will assume that you are using the ubi9/ruby-30 image, available via ruby:3.0 imagestream tag in Openshift. Building a simple ruby-sample-app application in Openshift can be achieved with the following step:

```
$ oc new-app ruby:3.0~https://github.com/sclorg/s2i-ruby-container.git --context-dir=3.0/test/puma-test-app/
```

The same application can also be built using the standalone S2I application on systems that have it available:

```
$ s2i build https://github.com/sclorg/s2i-ruby-container.git --context-dir=3.0/test/puma-test-app/ ubi9/ruby-30 ruby-sample-app
```

Accessing the application:

$ curl 127.0.0.1:8080

Environment variables

To set these environment variables, you can place them as a key value pair into a .s2i/environment file inside your source code repository.

• RACK_ENV

  This variable specifies the environment where the Ruby application will be deployed (unless overwritten) - production, development, test. Each level has different behaviors in terms of logging verbosity, error pages, ruby gem installation, etc.

  Note: Application assets will be compiled only if the RACK_ENV is set to production

• DISABLE_ASSET_COMPILATION

  This variable set to true indicates that the asset compilation process will be skipped. Since this only takes place when the application is run in the production environment, it should only be used when assets are already compiled.

• PUMA_MIN_THREADS, PUMA_MAX_THREADS

  These variables indicate the minimum and maximum threads that will be available in Puma's thread pool.

• PUMA_WORKERS

  This variable indicate the number of worker processes that will be launched. See documentation on Puma's clustered mode.

• RUBYGEM_MIRROR

  Set this variable to use a custom RubyGems mirror URL to download required gem packages during build process.

Hot deploy

In order to dynamically pick up changes made in your application source code, you need to make following steps:

• For Ruby on Rails applications

  Run the built Rails image with the RAILS_ENV=development environment variable passed to the podman -e run flag:

  $ podman run -e RAILS_ENV=development -p 8080:8080 rails-app
  

• For other types of Ruby applications (Sinatra, Padrino, etc.)

  Your application needs to be built with one of gems that reloads the server every time changes in source code are done inside the running container. Those gems are:

  • Shotgun
  • Rerun
  • Rack-livereload

  Please note that in order to be able to run your application in development mode, you need to modify the S2I run script, so the web server is launched by the chosen gem, which checks for changes in the source code.

  After you built your application image with your version of S2I run script, run the image with the RACK_ENV=development environment variable passed to the podman -e run flag:

  $ podman run -e RACK_ENV=development -p 8080:8080 sinatra-app
  

To change your source code in running container, use Podman's exec command:

$ podman exec -it <CONTAINER_ID> /bin/bash

After you podman exec into the running container, your current directory is set to /opt/app-root/src, where the source code is located.

Performance tuning

You can tune the number of threads per worker using the PUMA_MIN_THREADS and PUMA_MAX_THREADS environment variables. Additionally, the number of worker processes is determined by the number of CPU cores that the container has available, as recommended by Puma's documentation. This is determined using the cgroup cpusets subsystem. You can specify the cores that the container is allowed to use by passing the --cpuset-cpus parameter to the podman run command:

$ podman run -e PUMA_MAX_THREADS=32 --cpuset-cpus='0-2,3,5' -p 8080:8080 sinatra-app

The number of workers is also limited by the memory limit that is enforced using cgroups. The builder image assumes that you will need 50 MiB as a base and another 15 MiB for every worker process plus 128 KiB for each thread. Note that each worker has its own threads, so the total memory required for the whole container is computed using the following formula:

50 + 15 * WORKERS + 0.125 * WORKERS * PUMA_MAX_THREADS

You can specify a memory limit using the --memory flag:

$ podman run -e PUMA_MAX_THREADS=32 --memory=300m -p 8080:8080 sinatra-app

If memory is more limiting then the number of available cores, the number of workers is scaled down accordingly to fit the above formula. The number of workers can also be set explicitly by setting PUMA_WORKERS.

See also

Dockerfile and other sources are available on https://github.com/sclorg/s2i-ruby-container. In that repository you also can find another versions of Ruby environment Dockerfiles. Dockerfile for CentOS is called Dockerfile, Dockerfile for RHEL7 is called Dockerfile.rhel7, for RHEL8 it's Dockerfile.rhel8 and the Fedora Dockerfile is called Dockerfile.fedora.