Rake tasks for developers

Rake tasks are available for developers and others contributing to GitLab.

Set up database with developer seeds

Note that if your database user does not have advanced privileges, you must create the database manually before running this command. 

bundle exec rake setup

The setup task is an alias for gitlab:setup. This tasks calls db:reset to create the database, and calls db:seed_fu to seed the database. db:setup calls db:seed but this does nothing.

Environment variables

MASS_INSERT: Create millions of users (2m), projects (5m) and its relations. It’s highly recommended to run the seed with it to catch slow queries while developing. Expect the process to take up to 20 extra minutes.

See also Mass inserting Rails models.

LARGE_PROJECTS: Create large projects (through import) from a predefined set of URLs.

Seeding issues for all or a given project

You can seed issues for all or a given project with the gitlab:seed:issues task: 

# All projects
bin/rake gitlab:seed:issues

# A specific project
bin/rake "gitlab:seed:issues[group-path/project-path]"

By default, this seeds an average of 2 issues per week for the last 5 weeks per project.

Seeding issues for Insights charts

You can seed issues specifically for working with the Insights charts with the gitlab:seed:insights:issues task: 

# All projects
bin/rake gitlab:seed:insights:issues

# A specific project
bin/rake "gitlab:seed:insights:issues[group-path/project-path]"

By default, this seeds an average of 10 issues per week for the last 52 weeks per project. All issues are also randomly labeled with team, type, severity, and priority.

Seeding groups with subgroups

You can seed groups with subgroups that contain milestones/projects/issues with the gitlab:seed:group_seed task: 

bin/rake "gitlab:seed:group_seed[subgroup_depth, username]"

Group are additionally seeded with epics if GitLab instance has epics feature available.

Seeding custom metrics for the monitoring dashboard

A lot of different types of metrics are supported in the monitoring dashboard.

To import these metrics, you can run: 

bundle exec rake 'gitlab:seed:development_metrics[your_project_id]'

Automation

If you’re very sure that you want to wipe the current database and refill seeds, you can set the FORCE environment variable to yes: 

FORCE=yes bundle exec rake setup

This will skip the action confirmation/safety check, saving you from answering yes manually.

Discard stdout

Since the script would print a lot of information, it could be slowing down your terminal, and it would generate more than 20G logs if you just redirect it to a file. If we don’t care about the output, we could just redirect it to /dev/null: 

echo 'yes' | bundle exec rake setup > /dev/null

Note that since you can’t see the questions from stdout, you might just want to echo 'yes' to keep it running. It would still print the errors on stderr so no worries about missing errors.

Extra Project seed options

There are a few environment flags you can pass to change how projects are seeded

  • SIZE: defaults to 8, max: 32. Amount of projects to create.
  • LARGE_PROJECTS: defaults to false. If set, clones 6 large projects to help with testing.
  • FORK: defaults to false. If set to true, forks torvalds/linux five times. Can also be set to an existing project full_path to fork that instead.

Run tests

In order to run the test you can use the following commands:

  • bin/rake spec to run the RSpec suite
  • bin/rake spec:unit to run only the unit tests
  • bin/rake spec:integration to run only the integration tests
  • bin/rake spec:system to run only the system tests
  • bin/rake karma to run the Karma test suite

bin/rake spec takes significant time to pass. Instead of running the full test suite locally, you can save a lot of time by running a single test or directory related to your changes. After you submit a merge request, CI runs full test suite for you. Green CI status in the merge request means full test suite is passed.

You can’t run rspec . since this tries to run all the _spec.rb files it can find, also the ones in /tmp

You can pass RSpec command line options to the spec:unit, spec:integration, and spec:system tasks. For example, bin/rake "spec:unit[--tag ~geo --dry-run]".

For an RSpec test, to run a single test file you can run: 

bin/rspec spec/controllers/commit_controller_spec.rb

To run several tests inside one directory:

  • bin/rspec spec/requests/api/ for the RSpec tests if you want to test API only

Run RSpec tests which failed in Merge Request pipeline on your machine

If your Merge Request pipeline failed with RSpec test failures, you can run all the failed tests on your machine with the following Rake task: 

bin/rake spec:merge_request_rspec_failure

There are a few caveats for this Rake task:

  • You need to be on the same branch on your machine as the source branch of the Merge Request.
  • The pipeline must have been completed.
  • You may need to wait for the test report to be parsed and retry again.

This Rake task depends on the unit test reports feature, which only gets parsed when it is requested for the first time.

Speed up tests, Rake tasks, and migrations

Spring is a Rails application pre-loader. It speeds up development by keeping your application running in the background so you don’t need to boot it every time you run a test, Rake task or migration.

If you want to use it, you must export the ENABLE_SPRING environment variable to 1: 

export ENABLE_SPRING=1

Alternatively you can use the following on each spec run, 

bundle exec spring rspec some_spec.rb

Compile Frontend Assets

You shouldn’t ever need to compile frontend assets manually in development, but if you ever need to test how the assets get compiled in a production environment you can do so with the following command: 

RAILS_ENV=production NODE_ENV=production bundle exec rake gitlab:assets:compile

This compiles and minifies all JavaScript and CSS assets and copy them along with all other frontend assets (images, fonts, etc) into /public/assets where they can be easily inspected.

Emoji tasks

To update the Emoji aliases file (used for Emoji autocomplete), run the following: 

bundle exec rake gemojione:aliases

To update the Emoji digests file (used for Emoji autocomplete), run the following: 

bundle exec rake gemojione:digests

This updates the file fixtures/emojis/digests.json based on the currently available Emoji.

To generate a sprite file containing all the Emoji, run: 

bundle exec rake gemojione:sprite

If new emoji are added, the sprite sheet may change size. To compensate for such changes, first generate the emoji.png sprite sheet with the above Rake task, then check the dimensions of the new sprite sheet and update the SPRITESHEET_WIDTH and SPRITESHEET_HEIGHT constants accordingly.

Update project templates

Starting a project from a template needs this project to be exported. On a up to date main branch run: 

gdk start
bundle exec rake gitlab:update_project_templates
git checkout -b update-project-templates
git add vendor/project_templates
git commit
git push -u origin update-project-templates

Now create a merge request and merge that to main.

To update just a single template instead of all of them, specify the template name between square brackets. For example, for the cluster_management template, run: 

bundle exec rake gitlab:update_project_templates\[cluster_management\]

Generate route lists

To see the full list of API routes, you can run: 

bundle exec rake grape:path_helpers

The generated list includes a full list of API endpoints and functional RESTful API verbs.

For the Rails controllers, run: 

bundle exec rails routes

Since these take some time to create, it’s often helpful to save the output to a file for quick reference.

Show obsolete ignored_columns

To see a list of all obsolete ignored_columns run: 

bundle exec rake db:obsolete_ignored_columns

Feel free to remove their definitions from their ignored_columns definitions.

Validate GraphQL queries

To check the validity of one or more of our front-end GraphQL queries, run: 

# Validate all queries
bundle exec rake gitlab::graphql:validate
# Validate one query
bundle exec rake gitlab::graphql:validate[path/to/query.graphql]
# Validate a directory
bundle exec rake gitlab::graphql:validate[path/to/queries]

This prints out a report with an entry for each query, explaining why each query is invalid if it fails to pass validation.

We strip out @client fields during validation so it is important to mark client fields with the @client directive to avoid false positives.

Analyze GraphQL queries

Analogous to ANALYZE in SQL, we can run gitlab:graphql:analyze to estimate the of the cost of running a query.

Usage: 

# Analyze all queries
bundle exec rake gitlab::graphql:analyze
# Analyze one query
bundle exec rake gitlab::graphql:analyze[path/to/query.graphql]
# Analyze a directory
bundle exec rake gitlab::graphql:analyze[path/to/queries]

This prints out a report for each query, including the complexity of the query if it is valid.

The complexity depends on the arguments in some cases, so the reported complexity is a best-effort assessment of the upper bound.

Update GraphQL documentation and schema definitions

To generate GraphQL documentation based on the GitLab schema, run: 

bundle exec rake gitlab:graphql:compile_docs

In its current state, the Rake task:

  • Generates output for GraphQL objects.
  • Places the output at doc/api/graphql/reference/index.md.

This uses some features from graphql-docs gem like its schema parser and helper methods. The docs generator code comes from our side giving us more flexibility, like using Haml templates and generating Markdown files.

To edit the content, you may need to edit the following:

  • The template. You can edit the template at lib/gitlab/graphql/docs/templates/default.md.haml. The actual renderer is at Gitlab::Graphql::Docs::Renderer.
  • The applicable description field in the code, which Updates machine-readable schema files, which is then used by the rake task described earlier.

@parsed_schema is an instance variable that the graphql-docs gem expects to have available. Gitlab::Graphql::Docs::Helper defines the object method we currently use. This is also where you should implement any new methods for new types you’d like to display.

Update machine-readable schema files

To generate GraphQL schema files based on the GitLab schema, run: 

bundle exec rake gitlab:graphql:schema:dump

This uses GraphQL Ruby’s built-in Rake tasks to generate files in both IDL and JSON formats.

Update documentation and schema definitions

The following command combines the intent of Update GraphQL documentation and schema definitions and Update machine-readable schema files: 

bundle exec rake gitlab:graphql:update_all