Getting started

Devspace is a Continuous Integration tool managed by Jenkins CI providing an automation framework that runs repeated jobs. The default deployment initializes a Jenkins CI master with a predefined set of jobs.

Running and maintaining Devspace requires brief understanding of:

• Docker engine
• Docker compose

Running Devspace requires access to SSH and Git configuration files used for fetching and pushing the Git repositories.

Devspace code depends on the following repositories:

• devagent-docker

Installation

The following instructions explain how to deploy a devspace on a Docker host.

• Log into the Docker host using ssh

• Create a directory /data/username and change ownership:

  $ sudo mkdir /data/username
  $ sudo chown username /data/username

• Clone the devspace Git repository:

  $ git clone https://github.com/ome/devspace.git
  $ cd devspace

• Generated self-signed SSL certificates for the Jenkins and NGINX containers:

  $ ./sslcert jenkins/sslcert HOST_IP
  $ ./sslcert nginx/sslcert HOST_IP

  alternatively put your own certificate .crt and .key in the above locations.

• Copy the SSH and Git configuration files used for fetching and pushing the Git repositories under slave/.ssh and slave/.gitconfig. This is usually your own SSH and Git configuration files. Make sure that the permissions of the key are not too open. If this is the case, change the permissions e.g. `chmod 400 YOUR_KEY You need to use a public key without a passphrase and a .gitconfig file containing the following sections:

  [user]
     name = YOUR_NAME
     email = YOUR_EMAIL
  [github]
      token = YOUR_GITHUB_TOKEN
      user = YOUR_GITHUB_NAME

  • Run rename.py to match your topic name. Specify the Git user corresponding to the confguration files used above. If you do not yet have topic branches available on origin, use develop/master or one of the main branches:

    $ ./rename.py USER MYTOPIC --user git_user

• This will also replace the USER_ID of the various Dockerfile with the ID of the user who will run the devspace, assumed to be: id -u, i.e. the current user.

• Set the environment variables in .env, especially:

  JENKINS_USERNAME=devspace
  JENKINS_PASSWORD=<password>

• Optionally, commit all the deployment changes above on the local clone of the devspace repository.

Start and configure:

• Build devspace using docker compose:

  $ docker compose -f docker-compose.yml build 

• Start devspace using docker compose:

  $ docker compose -f docker-compose.yml up -d

  By default, this will use the name of the directory as the project name. In the case of a shared Docker host, it is possible to override the project name using

  $ docker compose up -p my_project -d

• Depending on the ssh key, you might have to run the following comment in the test-integration container. For example: $ docker exec -it devspace-testintegration-1 bash $ ssh -T git@github.com

  A message should be returned after running the command: $ Hi snoopycrimecop! You've successfully authenticated, but GitHub does not provide shell access.

• Retrieve the dynamic port of the Jenkins NGINX container. You can access the Jenkins UI from https://HOST_IP:PORT after accepting the self-signed certificate:

  $ docker compose -p my_project port nginxjenkins 443

• Create the maven-internal Nexus repository:

  $ docker compose exec nexus /nexus-data/createRepoMavenInternal.sh

• [Optional] Turn on Basic HTTP authentication for Jenkins

  sudo htpasswd -c jenkins/conf.d/passwdfile nginx

  and update jenkins/conf.d/jenkins.conf:

  auth_basic "Restricted";
  auth_basic_user_file /etc/nginx/conf.d/passwdfile;

GitHub OAuth

You can optionally enable GitHub OAuth:

• Copy home/init.groovy.d/github-oauth.groovy.disabled to home/init.groovy.d/github-oauth.groovy
• Create a new GitHub app and edit the variables at the top of home/init.groovy.d/github-oauth.groovy. The script also gives details of the required GitHub OAuth callback.
• Restart Jenkins

Note: if you are modifying an existing devspace you are advised to backup home/config.xml. If there are errors in the GitHub setup you can restore home/config.xml to return to the default authentication.

After the script has completed you can either leave it in place so it will override any manual changes on restart, or delete it and make changes through the Jenkins UI.

Job configurations

• When running the OMERO-server job for the first time, select the PURGE_DATA option to create the database.

Job workflow

The default deployment initializes a Jenkins server with a predefined set of jobs.

The table below lists the job names, the Jenkins node labels and the associated docker they are associated with and a short description of the jobs.

This means that by default the following repositories need to be forked to your GitHub account:

• ome/openmiscrocopy
• ome/omero-documentation
• ome/bioformats

If you do not have some of the repositories forked, you will need to remove the jobs from the list of jobs to run either from the Trigger job configuration or directly from the Jenkins UI i.e. Trigger > Configure.

New jobs

It is recommended that new jobs should be defined using Jenkinsfile pipelines in the target repository as this makes it easier to maintain jobs. Most Jenkins Pipeline jobs can share the same configuration apart from the repository URL. If you do not require any special configuration use the TEMPLATE-pipeline-job-config.xml template by adding the job and parameters to pipeline-configs.yaml. Supported parameters are documented in that file.

The rename.py script will create the required job configurations from pipeline-configs.yaml as well as performing the renaming steps. If for some reason you want to create the new job without running rename.py you can just run createpipelinejobs.py.

Alternatively create a new job in the Jenkins web-interface in the usual way.

Default packages used

Troubleshooting

See Troubleshooting

Upgrade

See Changelog

Run BioFormats jobs

To run the BioFormats testing the various readers on sample data, you will need to activate the a private job

• In the devspace, create a directory home/jobs/DATA_REPO_CONFIG-merge/
• Download the job configuration from config.xml in https://github.com/openmicroscopy/management_tools/tree/master/ci/jobs/DATA_REPO_CONFIG-merge (private repository) and place it in the newly created directory
• Comment out the line build job: "DATA_REPO_CONFIG-merge" in the trigger job