Content Sources

What is it?

Content Sources is an application for storing information about external content (currently YUM repositories) in a central location as well as creating snapshots of those repositories, backed by a Pulp server.

To read more about Content Sources use cases see:

1. Introspection
2. Snapshots

Developing

Requirements:

1. podman & podman-compose Do not use v1.3.0, or docker & docker-compose, installed and running (and Orbstack for Mac))
   • This is used to start a set of containers that are dependencies for content-sources-backend
2. yaml2json tool installed (pip install json2yaml).

Create your configuration

Create a config file from the example:

cp ./configs/config.yaml.example ./configs/config.yaml

Add pulp.content to /etc/hosts for Go integration tests and client access

sudo echo "<your-host-ip> pulp.content" | sudo tee -a /etc/hosts

You can find your host IP by running ifconfig or ip addr list. If you have a virtual bridge interface (virbr0, started up by running virt-manager), then you can use that IP.

Start dependency containers

make compose-up

If you want to start containers without initializing any data:

make compose-run

Import RHEL 9 Repos

make repos-import-rhel9

or import them all

make repos-import

Introspect and Snapshot all current repos

make process-repos

For local development

If you want less Red Hat repos:

OPTIONS_REPOSITORY_IMPORT_FILTER=small make repos-import

This will import and snapshot repos needed for the minimal viable environment. Useful for running Playwright tests.

make repos-minimal

Run the server!

make run

Hit the API:

curl -H "$( ./scripts/header.sh 9999 1111 )" http://localhost:8000/api/content-sources/v1.0/repositories/

Stop dependency containers

When its time to shut down the running containers:

make compose-down

Clean the volume that it uses (this stops the container before doing it if it were running):

make compose-clean

There are other make rules that could be helpful, run make help to list them. Some are highlighted below

Playwright testing

• Ensure that the backend server is running
• Ensure the correct node version, is installed and in use:
  cd _playwright-tests and nvm use
• Copy the env file and create a file at: _playwright-tests/.env For local development only the BASE_URL:http://127.0.0.1:8000 is required, which is already set in the example config.

make playwright

OR

cd _playwright-tests \
&& yarn install \
&& yarn playwright install \
&& yarn playwright test

HOW TO ADD NEW MIGRATION FILES

You can add new migration files, with the prefixed date attached to the file name, by running the following:

go run cmd/dbmigrate/main.go new <name of migration>

Database Commands

Migrate the Database

make db-migrate-up

Get an interactive shell:

make db-shell

Or open directly a postgres client by running:

make db-cli-connect

Kafka commands

You can open an interactive shell by:

make kafka-shell

You can run kafka-console-consumer.sh using KAFKA_TOPIC by:

make kafka-topic-consume KAFKA_TOPIC=my-kafka-topic
make kafka-topic-consume # Use the first topic at KAFKA_TOPICS list

There are other make rules that could be helpful, run make help to list them.

Start / Stop prometheus

Create the configuration for prometheus, getting started with the example one.

Update the configs/prometheus.yaml file to set your hostname instead of localhost at scrape_configs.job_name.targets:

# Note that the targets object cannot reference localhost, it needs the name of your host where
# the prometheus container is executed.
cat ./configs/prometheus.example.yaml | sed "s/localhost/$(hostname)/g" > ./configs/prometheus.yaml

To start prometheus run:

make prometheus-up

To stop prometheus container run:

make prometheus-down

To open the prometheus web UI, once the container is up, run the below:

make prometheus-ui

Start / Stop mock for rbac

Configuration requirements

• To use this you need to enable RBAC into config/configs.yaml file:

  clients:
    rbac_enabled: True
    rbac_base_url: http://localhost:8800/api/rbac/v1
    rbac_timeout: 30
  mocks:
    rbac:
      user_read_write: ["[email protected]", "jdoe"]
      user_read: ["[email protected]", "tdoe"]
    kessel:
      user_read_write: ["write-user"]
      user_read: ["read-user"]
      user_no_permissions: ["no-perms-user"]

Running it

• Run the application by: make run or ./release/content-sources api consumer instrumentation mock_rbac.
• Make some request using: ./scripts/header.sh 12345 [email protected] for admin or ./scripts/header.sh 12345 [email protected] for viewer.

RBAC mock service is started for make run To use it running directly the service: ./release/content-sources api consumer instrumentation mock_rbac Add the option mock_rbac

Kessel

• The kessel permissions will be used if the kessel feature is enabled.
• Kessel will be used alongside rbac when only enabled for specific users, orgs, or accounts.

Migrate your database (and seed it if desired)

make db-migrate-up

Run the server!

make run

Hit the API:

curl -H "$( ./scripts/header.sh 9999 1111 )" http://localhost:8000/api/content-sources/v1.0/repositories/

Generating new openapi docs:

make openapi

Generating new mocks:

make mock

Live Reloading Server

This is completely optional way of running the server that is useful for local development. It rebuilds the project after every change you make, so you always have the most up-to-date server running. To set this up, all you need to do is install the "Air" go tool, here is how. The recommended way is doing:

go install github.com/air-verse/air@latest

After that, all that needs to be done is just running air, it should automatically use the defined config for this project (.air.toml).

air

Testing with Image Builder Frontend

While working on features in the Image Builder frontend that integrate with this service, it can be useful to develop and test those alongside each other locally. To do that there are 2 changes required in the IB FE repo:

1. Changing the path to the content sources API config. From which will Redux generate the clients.
   api/config/contentSources.ts: schemaFile: '/YOUR_LOCAL_PATH_TO_THIS_REPO/content-sources-backend/api/openapi.json',
2. Adding a route to the FEC config, which will be picked by the frontend dev proxy.
   fec.config.js: ['/api/content-sources']: { host: 'http://127.0.0.1:8000' }

Example Git Patch:

diff --git a/api/config/contentSources.ts b/api/config/contentSources.ts
index 7d4db495..32a00e3f 100644
--- a/api/config/contentSources.ts
+++ b/api/config/contentSources.ts
@@ -1,7 +1,7 @@
 import type { ConfigFile } from '@rtk-query/codegen-openapi';

 const config: ConfigFile = {
-  schemaFile: 'https://console.redhat.com/api/content-sources/v1/openapi.json',
+  schemaFile: 'YOUR_LOCAL_PATH_TO_THE_CONTENT_SOURCES_BACKEND_REPO/content-sources-backend/api/openapi.json',
   apiFile: '../../src/store/service/emptyContentSourcesApi.ts',
   apiImport: 'emptyContentSourcesApi',
   outputFile: '../../src/store/service/contentSourcesApi.ts',
diff --git a/fec.config.js b/fec.config.js
index 3767689d..cba37234 100644
--- a/fec.config.js
+++ b/fec.config.js
@@ -102,6 +102,7 @@ module.exports = {
         };
       }, {}),
     }),
+    ['/api/content-sources']: { host: 'http://127.0.0.1:8000' }
   },
   plugins: plugins,
   moduleFederation: {

Configuration

The default configuration file in ./configs/config.yaml.example shows all available config options. Any of these can be overridden with an environment variable. For example "database.name" can be passed in via an environment variable named "DATABASE_NAME".

Linting

To use golangci-lint:

1. make install-golangci-lint
2. make lint

To use pre-commit linter: make install-pre-commit

Code Layout

Path	Description
api	Openapi docs and doc generation code
db/migrations	Database Migrations
pkg/api	API Structures that are used for handling data within our API Handlers
pkg/config	Config loading and application bootstrapping code
pkg/dao	Database Access Object. Abstraction layer that provides an interface and implements it for our default database provider (postgresql). It is separated out for abstraction and easier testing
pkg/db	Database connection and migration related code
pkg/handler	Methods that directly handle API requests
pkg/middleware	Holds all the middleware components created for the service
pkg/event	Event message logic. More info here
pkg/models	Structs that represent database models (Gorm)
pkg/seeds	Code to help seed the database for both development and testing
pkg/candlepin_client	Candlepin client
pkg/pulp_client	Pulp client
pkg/tasks	Tasking system. More info here
scripts	Helper scripts for identity header generation and testing

More info

• Architecture
• OpenApi Docs