Raku Land

Sparky

cpan:MELEZHIK

SYNOPSIS

Sparky is a flexible and minimalist continuous integration server written in Raku.

Nice web UI

The essential features of Sparky:

Interested? Let's go ahead! (:

Build status

Build Status

Sparky workflow in 4 lines:

$ sparkyd # run Sparky daemon to build your projects
$ raku bin/sparky-web.raku # run Sparky web UI to see build statuses and reports
$ nano ~/.sparky/projects/my-project/sparrowfile  # write a build scenario
$ firefox 127.0.0.1:3000 # see what's happening

Installation

$ sudo apt-get install sqlite3
$ git clone https://github.com/melezhik/sparky.git
$ cd sparky && zef install .

Setup

First you should run database initialization script to populate database schema:

$ raku db-init.raku

Running daemon

Then you need to run (*) the sparky daemon

$ sparkyd
$ sparkyd --timeout=600 # sleep 10 minutes
$ SPARKY_TIMEOUT=30 sparkyd ...

Running sparky in demonized mode.

At the moment sparky can't demonize itself, as temporary workaround use linux nohup command:

$ nohup sparkyd &

Or you can use Sparrowdo installer, which install service as systemd unit:

$ nano utils/install-sparky-web-systemd.raku # change working directory and user
$ sparrowdo --sparrowfile=utils/install-sparkyd-systemd.raku --no_sudo --localhost

* sparkyd should be in your PATH, if not, fix it before going further.

You should have installed with zef install . listed above.

It could live for instance in $HOME/.raku/bin along with other scripts required (like sparrowdo)

Running Web UI

And finally sparky has simple web UI to show builds statuses and reports.

To run Sparky web UI launch sparky-web.raku script from the bin/ directory:

$ raku bin/sparky-web.raku

This is Bailador application, so you can set any Bailador related options here.

For example:

$ BAILADOR=host:0.0.0.0,port:5000 raku bin/sparky-web.raku

You can use Sparrowdo installer as well, which installs service as systemd unit:

$ nano utils/install-sparky-web-.raku # change working directory, user and root directory
$ sparrowdo --sparrowfile=utils/install-sparky-web-systemd.raku --no_sudo --localhost

Creating first sparky project

Sparky project is just a directory located at the sparky root directory:

$ mkdir ~/.sparky/projects/bailador-app

Build scenario

Sparky is built on Sparrowdo, read Sparrowdo to know how to write Sparky scenarios. Here is a short example.

Say, we want to check out a Raku project from from Git, install dependencies and then run unit tests:

$ nano ~/.sparky/projects/bailador-app/sparrowfile

And add content like this:

directory "project";

git-scm 'https://github.com/melezhik/rakudist-teddy-bear.git', %(
  to => "project",
);

zef "{%*ENV<PWD>}/project", %( depsonly => True );

zef 'TAP::Harness App::Prove6';

bash 'prove6 -l', %(
  debug => True,
  cwd => "{%*ENV<PWD>}/project/"
);

Configure Sparky workers

By default the build scenario gets executed on the same machine you run Sparky at, but you can change this to any remote host setting Sparrowdo related parameters in the sparky.yaml file:

$ nano ~/.sparky/projects/bailador-app/sparky.yaml

And define worker configuration:

sparrowdo:
  host: '192.168.0.1'
  ssh_private_key: /path/to/ssh_private/key.pem
  ssh_user: sparky
  no_index_update: true
  sync: /tmp/repo

You can read about the all available parameters in Sparrowdo documentation.

Skip bootstrap

Sparrowdo bootstrap takes a while, if you don't need bootstrap ( sparrow client is already installed at a target host ) use bootstrap: false option:

sparrowdo:
  bootstrap: false

Purging old builds

To remove old build set keep_builds parameter in sparky.yaml:

$ nano ~/.sparky/projects/bailador-app/sparky.yaml

Put number of past builds to keep:

keep_builds: 10

That makes Sparky remove old build and only keep last keep_builds builds.

Run by cron

It's possible to setup scheduler for Sparky builds, you should define crontab entry in sparky yaml file. for example to run a build every hour at 30,50 or 55 minute say this:

$ nano ~/.sparky/projects/bailador-app/sparky.yaml

With this schedule:

crontab: "30,50,55 * * * *"

Follow Time::Crontab documentation on crontab entries format.

Manual run

If you want to build a project from web UI, use allow_manual_run:

$ nano ~/.sparky/projects/bailador-app/sparky.yaml

And activate manual run:

allow_manual_run: true

Trigger build by SCM changes

** warning ** - the feature is not properly tested, feel free to post issues or suggestions

To trigger Sparky builds on SCM changes, define scm section in sparky.yaml file:

scm:
  url: $SCM_URL
  branch: $SCM_BRANCH

Where:

For example:

scm:
  url: https://github.com/melezhik/rakudist-teddy-bear.git
  branch: master

Once a build is triggered one needs to handle build environment leveraging tags()<SCM_*> objects:

directory "scm";

say "current commit is: {tags()<SCM_SHA>}";

git-scm tags()<SCM_URL>, %(
  to => "scm",
  branch => tags<SCM_BRANCH>
);

bash "ls -l {%*ENV<PWD>}/scm";

Disable project

You can disable project builds by setting disable option to true:

$ nano ~/.sparky/projects/bailador-app/sparky.yaml

disabled: true

It's handy when you start a new project and don't want to add it into build pipeline.

Downstream projects

You can run downstream projects by setting downstream field at the upstream project sparky.yaml file:

$ nano ~/.sparky/projects/main/sparky.yaml

downstream: downstream-project

File triggering protocol (FTP)

Sparky FTP allows to trigger builds automatically by just creating files with build parameters in special format:

$ nano $project/.triggers/foo-bar-baz.pl6

File should have a *.pl6 extension and be located in project .trigger directory.

A content of the file should Raku code returning a Hash:

{
  description => "Build app",
  cwd => "/path/to/working/directory",
  sparrowdo => %(
    localhost => True,
    no_sudo   => True,
    conf      => "/path/to/file.conf"
  )
}

Sparky daemon parses files in .triggers and launch build per every file, removing file afterwards, this process is called file triggering.

To separate different builds just create trigger files with unique names inside $project/.trigger directory.

FTP allows to create supplemental APIs to implement more complex and custom build logic, while keeping Sparky itself simple.

Trigger attributes

Those keys could be used in trigger Hash. All they are optional.

%(
  host  => "foo.bar",
  ssh_user  => "admin",
  tags => "prod,backend"
)

Should follow the format of sparky.yaml, sparrowdo section

A unique key

Sparky plugins

Sparky plugins are extensions points to add extra functionality to Sparky builds.

These are Raku modules get run after a Sparky project finishes or in other words when a build is completed.

To use Sparky plugins you should:

Install Sparky plugins

You should install a module on the same server where you run Sparky at. For instance:

$ zef install Sparky::Plugin::Email # Sparky plugin to send email notifications

Configure Sparky

In project's sparky.yaml file define plugins section, it should be list of Plugins and its configurations.

For instance:

$ cat sparky.yaml

That contains:

plugins:
  - Sparky::Plugin::Email:
    parameters:
      subject: "I finished"
      to: "happy@user.email"
      text: "here will be log"
  - Sparky::Plugin::Hello:
    parameters:
      name: Sparrow

Creating Sparky plugins

Technically speaking Sparky plugins should be just Raku modules.

For instance, for mentioned module Sparky::Plugin::Email we might have this header lines:

use v6;

unit module Sparky::Plugin::Hello;

That is it.

The module should have run routine which is invoked when Sparky processes a plugin:

our sub run ( %config, %parameters ) {

}

As we can see the run routine consumes its parameters as Raku Hash, these parameters are defined at mentioned sparky.yaml file, at plugin parameters: section, so this is how you might handle them:

sub run ( %config, %parameters ) {

  say "Hello " ~ %parameters<name>;

}

You can use %config Hash to access Sparky guts:

For example:

sub run ( %config, %parameters ) {

  say "build id is: " ~ %parameters<build-id>;

}

Alternatively you may pass some predefined parameters plugins:

For example:

$ cat sparky.yaml

That contains:

plugins:
  - Sparky::Plugin::Hello:
    parameters:
      name: Sparrow from project %PROJECT%

Limit plugin run scope

You can defined when to run plugin, here are 3 run scopes:

Scopes are defined at run_scope: parameter:

- Sparky::Plugin::Hello:
  run_scope: fail
  parameters:
    name: Sparrow

An example of Sparky plugins

An example Sparky plugins are:

Command line client

You can build the certain project using sparky command client called sparky-runner.raku:

$ sparky-runner.raku --dir=/home/user/.sparky/projects/bailador-app

Or just:

$ cd ~/.sparky/projects/bailador-app && sparky-runner.raku

Sparky runtime parameters

All this parameters could be overridden by command line ( --root, --work-root )

Root directory

This is sparky root directory, or directory where Sparky looks for the projects to get built:

~/.sparky/projects/

Work directory

This is working directory where sparky might place some stuff, useless at the moment:

~/.sparky/work

Environment variables

SPARKY_SKIP_CRON

You can disable cron check to run project forcefully, by setting SPARKY_SKIP_CRON environment variable:

$ export SPARKY_SKIP_CRON=1 && sparkyd

SPARKY_ROOT

Sets the sparky root directory

SPARKY_HTTP_ROOT

Set Sparky web application http root. Useful when proxy application through Nginx.

SPARKY_TIMEOUT

Sets timeout for sparky workers, see Running daemon section.

Running under other databases engines (MySQL, PostgreSQL)

By default Sparky uses sqlite as database engine, which makes it easy to use when developing. However sqlite has limitation on transactions locking whole database when doing inserts/updates (Database Is Locked errors).

if you prefer other databases here is guideline.

Create sparky configuration file

You should defined database engine and connection parameters, say we want to use MySQL:

$ nano ~/sparky.yaml

With content:

database:
  engine: mysql
  host: $dbhost
  port: $dbport
  name: $dbname
  user: $dbuser
  pass: $dbpassword

For example:

database:
  engine: mysql
  host: "127.0.0.1"
  port: 3306
  name: sparky
  user: sparky
  pass: "123"

Installs dependencies

Depending on platform it should be client needed for your database API, for example for Debian we have to:

$ sudo yum install mysql-client

Creating database user, password and schema

DB init script will generate database schema, provided that user defined and sparky configuration file has access to the database:

$ raku db-init.raku

That is it, now sparky runs under MySQL!

Change UI theme

Sparky uses Bulma as a CSS framework, you can easily change the theme through sparky configuration file:

$ nano ~/sparky.yaml

And choose your theme:

ui:
  theme: cosmo

The list of available themes is on https://jenil.github.io/bulmaswatch/

Trigger jobs from HTTP API

POST /build/project/$project

Examples

You can see examples of sparky scenarios in examples/ folder.

See also

Bailador - A light-weight route-based web application framework for Perl 6. Sparky-docker - Run Sparky as Docker container.

Author

Alexey Melezhik