Developer Deployment

Operating System

Any flavor of Linux should be acceptable for development, but all of the instructions below are written for Fedora 35 Workstation. You may need to make slight changes to the package names, depending on what is available from your system’s package repos, and your specific system configuration.

Getting the Source

You can browse the source code at https://github.com/candlepin/candlepin/.

For anonymous access to the Candlepin source code, feel free to clone the repository:

$ git clone https://github.com/candlepin/candlepin

Candlepin committers can clone the repository using the ssh url, which is required if you want to push changes into the repo (and of course you need permission to do so).

$ git clone git@github.com:candlepin/candlepin.git

For more information on working with Git, checkout the Spacewalk GitGuide.

Vagrant

Candlepin provides a Vagrant box for creating a virtualized development environment. This is the recommended way to do work on Candlepin, as it helps ensure everyone is working with the same the development environment.

Setup

  1. Install libvirt, Vagrant, Ansible, and the required vagrant plugins
     $ sudo dnf install @virtualization vagrant vagrant-libvirt vagrant-sshfs vagrant-hostmanager ansible
    
  2. Enable the libvirtd and virtnetworkd services
     $ sudo systemctl enable --now libvirtd virtnetworkd
    
  3. Bring up the Vagrant box
     $ vagrant up
    

    This will run through the process of starting the virtual machine, creating it as necessary, and provisioning the box with Ansible. On first run, this step may take some time, and may appear unresponsive at some steps. However, later runs will bring the box up very quickly.

  4. SSH into the Vagrant box
     $ vagrant ssh
    

    This will drop you into a shell at the root directory. The local Candlepin from which the Vagrant box was built has been synced into /vagrant which itself is symlinked from ~/devel/candlepin. At this point, Candlepin can be deployed following the instructions in the Deploying Candlepin section.

Post-setup Steps

Note that while Vagrant makes setting up the environment easier, some elements of the user space may not be fully configured:

  • Git configuration is not copied into the Vagrant box

    You will need to manually configure your git e-mail address and name in the Vagrant box. This step may be omitted, but you will be forced to drop out of the Vagrant box to commit or push changes.

These steps only need to be performed after initially bringing up the box for the first time, or if the box is ever recreated after destroying it.

These may eventually be resolved in future versions of the Ansible roles responsible for provisioning the Vagrant boxes.

Local Development

If you are unable or unwilling to use the Vagrant box for development tasks, you can, instead, set up your local environment for Candlepin development. Note that some Candlepin dependencies may not be available from your system’s package manager and/or may conflict with those that are. In such cases, you will need to manually resolve the conflict.

  1. Install and configure core dependencies

     $ sudo dnf install gcc make java-1.8.0-openjdk-devel java-11-openjdk-devel jss tomcat gettext openssl
    

    By default, the JAVA_HOME environment variable is not setup during this install. This should be configured to point to the home directory for the “active” JVM (Java 11 in most cases, Java 8 for some older Candlepin branches).

    On systems where Alternatives is available and configured (as is the case on Fedora 34), this should point to the Alternatives-managed JDK directory:

     $ echo 'export JAVA_HOME="/var/lib/jvm/java"' >> ~/.bashrc
    

    If Alternatives is not installed or enabled, JAVA_HOME will need to be configured to point to the explicit JDK directory as needed.

  2. Install Ruby and Ruby gems

    Candlepin currently requires Ruby 2.4 to run its spec test suite (rspec), which is unfortunately no longer available through the standard package repos in Fedora 34. For this reason, and to avoid conflicts with various gems, we highly recommend using RVM to install and manage Ruby and its gems.

    Instructions for installing RVM are available on the RVM website (http://rvm.io).

    Warning: Do not install RVM, Ruby, or the necessary gems as root

    Once RVM is installed, install Ruby 2.4 via rvm:

     $ rvm install ruby-2.4
    

    Finally, install bundler so we can install the required gems:

     $ gem install bundler
     $ bundle install
    
  3. Set up PostgreSQL

    PostgreSQL is the recommended database solution for Candlepin. PostgreSQL 9.6 is required at minimum, but newer versions are also supported and recommended.

    • Install PostgreSQL server, client and the JDBC connector

        $ sudo dnf install postgresql-server postgresql postgresql-jdbc
      
    • Initialize the database

        $ sudo postgresql-setup --initdb --unit postgresql
      
    • Configure PostgreSQL to trust local connections

      Open /var/lib/pgsql/data/pg_hba.conf in a text editor and change the local connection methods to be “trust.”:

        # TYPE  DATABASE    USER        CIDR-ADDRESS          METHOD
        local   all         all                               trust
        host    all         all         127.0.0.1/32          trust
        host    all         all         ::1/128               trust
      

      Note: This is not advised on a production or external-facing environment. For such environments, this file should be configured to allow local connections to the “candlepin” database from the “candlepin” user. Exactly what this configuration will look like may vary depending on the requirements of the environment.

    • Enable and start the postgresql service

        $ sudo systemctl enable postgresql.service
        $ sudo systemctl start postgresql.service
      
    • Create the candlepin user

        $ sudo su - postgres -c 'createuser -dls candlepin'
      
    • Create the candlepin database

        $ createdb --user candlepin candlepin
      
  4. Set up MariaDB/MySQL

    In addition to PostgreSQL, Candlepin also supports using MariaDB or MySQL databases, so both must be available for testing purposes.

    • Install MariaDB server, MariaDB, and the JDBC connector

        $ sudo dnf install mariadb-server mariadb mysql-connector-java
      
    • Add the Candlepin-specific server configuration

      Create the file /etc/my.cnf.d/candlepin.cnf and the following:

        [mysqld]
        collation-server=utf8_general_ci
        character-set-server=utf8
      
        transaction-isolation=READ-COMMITTED
      
    • Set the default time zone

      Open /etc/my.cnf.d/mariadb-server.cnf in a text editor and add the following under [server]:

        default-time-zone="+00:00"
      
    • Enable and start the mariadb service

        $ sudo systemctl enable mariadb.service
        $ sudo systemctl start mariadb.service
      
    • Create the candlepin user and its necessary permissions

        $ sudo mysql --user=root mysql --execute="CREATE USER 'candlepin'@'localhost'; GRANT ALL PRIVILEGES on candlepin.* TO 'candlepin'@'localhost' WITH GRANT OPTION;"
      
    • Create the candlepin database

        $ mysqladmin --user candlepin create candlepin
      

At this point, Candlepin can be deployed following the instructions in the Deploying Candlepin section.

Deploying Candlepin

The preferred method for deploying Candlepin from source is using the supplied deploy script in the bin/deployment directory. This script will generate the database schema, create the necessary certificates for Tomcat, deploy the .war file, and restart Tomcat. It can also be used to generate candlepin.conf or generate data for manual testing.

Note: Older Candlepin branches may store the deploy script in server/bin/ instead.

For the first deployment, or whenever changing databases, it is recommended to use the -a option to regenerate the candlepin.conf file for the new environment. The -g flag can also be specified to tell the deploy script to drop the database (if it exists) and regenerate it.

$ ./bin/deployment/deploy -g -a

By default, the deploy script will target PostgreSQL, however, the -m flag can be set to target MariaDB/MySQL instead:

$ ./bin/deployment/deploy -g -m -a

For a typical redeployment once the environment has already been configured, no flags or options need to be provided. This will simply recompile and deploy Candlepin without changing the configuration or dropping the existing database.

$ ./bin/deployment/deploy

A full listing of the options available can be viewed by using -h or --help with the command:

$ ./bin/deployment/deploy -h
usage: deploy [options]

OPTIONS:
  -f          force cert regeneration
  -g          generate database
  -r          generate test repositories
  -t          import test data
  -T          import minimal test data, some owners, users, and roles
  -H          include test resources for hosted mode
  -m          use MySQL
  -a          auto-deploy a generated candlepin.conf
  -v          verbose output
  -d <name>   specify a database name to use when creating or updating the Candlepin database
  -b          deploy candlepin with an external Artemis message broker

Note: The above output may differ from the options available in your working branch.

After running the deploy script, you can verify Candlepin is running by testing the “status” endpoint:

$ curl -k "https://localhost:8443/candlepin/status"

If Candlepin is running properly, this will return a block of JSON data describing Candlepin’s current state and some capabilities:

{
  "mode" : "NORMAL",
  "modeReason" : null,
  "modeChangeTime" : null,
  "result" : true,
  "version" : "4.1.1",
  "release" : "1",
  "standalone" : true,
  "timeUTC" : "2021-06-25T19:08:39+0000",
  "rulesSource" : "default",
  "rulesVersion" : "5.43",
  "managerCapabilities" : [ "instance_multiplier", "derived_product", "vcpu", "cert_v3", "hypervisors_heartbeat", "remove_by_pool_id", "syspurpose", "storage_band", "cores", "hypervisors_async", "org_level_content_access", "guest_limit", "ram", "batch_bind" ],
  "keycloakRealm" : null,
  "keycloakAuthUrl" : null,
  "keycloakResource" : null
}

Running Tests

Candlepin provides two tests suites for proper functionality and stability: unit tests and specification tests, via JUnit and rspec, respectively. Whenever changes are made to Candlepin, both of these suites should be run to verify existing functionality has not be affected by the changes, and new tests should be added to each suite as appropriate for the change.

Unit tests

Unit testing is provided by JUnit, and can be invoked through Gradle with the test task from the root of the Candlepin repository.

$ ./gradlew test

This will run the entire unit test suite, and write the results in HTML form to the build directory. To run tests for a specific suite of tests, the --tests option can be used with the class name of the suite to run.

For example, to run the tests in the JobManagerTest class, the following command can be used:

$ ./gradlew test --tests JobManagerTest

Lastly, an individual test in a given suite can be run using the name of the method, following Java dot-notation:

$ ./gradlew test --tests "JobManagerTest.jobStatusFound"

Specification Tests

Specification tests are run through rspec, and can be invoked through Gradle with the rspec task from the root of the Candlepin repository.

$ ./gradlew rspec

This will run the entire spec test suite, and write the results to the console. To run tests for a specific spec test suite, the --spec option can be used with the partial file name of the spec file to run, without the “_spec.rb” suffix.

For example, to run the tests in the activation_key_spec.rb file, the following command can be used:

$ ./gradlew rspec --spec activation_key

Individual tests or groups of tests within a given spec file can be run using the --test option and providing a string to match against the name of the desired tests.

For instance, to run the tests in the activation_key_spec file containing the word “updating”, the following command can be used:

$ ./gradlew rspec --spec activation_key --test 'updating'

Troubleshooting

There are a few things that some folks have hit while deploying Candlepin.

Database Initialization

IMPORTANT: No matter which database you are using, you will most likely need to initialize it once and only once after it is created. This process creates some core entries in the database required for Candlepin to operate properly. If you are using our deploy script, this will be handled for you automatically. Otherwise you can trigger this by hitting the following URL:

$ wget -qO- http://localhost:8080/candlepin/admin/init

Repeated calls to this URL are not required, but will be harmless.

Tomcat not listening on the right ports

If you see something amiss, an easy first thing to check is to ensure that tomcat is running and listening on the right ports.

# netstat -anlp | grep java
tcp        0      0 :::8080                     :::*                        LISTEN      21864/java
tcp        0      0 :::8443                     :::*                        LISTEN      21864/java
tcp        0      0 ::ffff:127.0.0.1:8005       :::*                        LISTEN      21864/java
tcp        0      0 :::8009                     :::*                        LISTEN      21864/java
unix  2      [ ]         STREAM     CONNECTED     793153 21864/java

The important things here are 8080 (the default HTTP listener), and 8443 (the SSL listener). If you don’t have both of those, try setting FORCECERT=1 and trying again. Consulting catalina.out is always useful as well.

HornetQ Upgrade Errors

The hornetq journal doesn’t seem to like being used by new versions. If you see an exception on startup like:

SEVERE: Failure in initialisation
java.lang.IllegalStateException: Invalid record type 23
at org.hornetq.core.persistence.impl.journal.JournalStorageManager.loadBindingJournal(JournalStorageManager.java:1527)

You need to clear your journal:

$ rm -rf /var/lib/candlepin/hornetq*
Last modified on 1 March 2024