Insights operator controller

• Description
• How to build the tool
• Start
• Configuration
  • HTTPS instead of HTTP
  • Configuration file
  • Environment variables
  • LDAP Authentication
• Data storage
• ER Diagram
  • SQLite
  • PostgreSQL
  • RDS AWS PostgreSQL
• Testing
  • Unit tests
  • REST API tests
  • BDD tests
• Code style and cyclomatic complexity checks
• CI
• Contribution
• Package manifest

Description

A service to be used to store Insights operator configuration and to offer the configuration to selected operator.

How to build the tool

Go version 1.14 or higher is required to build the client.

Use the standard Go command:

go build

This command should create an executable file named insights-operator-controller.

Alternatively you can use GNU make to perform the same operation:

make build

Start

Just run the executable file created by go build:

./insights-operator-controller

If you need to start the cleanly built controller, use the following command:

make run

Configuration

HTTPS instead of HTTP

Change the following lines in config.toml:

use_https=true
address=":4443"
tls_cert="certs/cert.pem"
tls_key="certs/key.pem"

Please note that the service (when run locally) use the self-signed certificate. You'd need to use certs.pem file on client side (curl, web browser etc.)

• use_https is boolean flag that defines if connection should be secured
• address is address of controller server
• tls_cert is path to certificate, can be used only if use_https == true
• tls_key is path to key of certificate, can be used only if use_https == true

Configuration file

Default configuration file is config.toml. It is possible to specify config file via environment variable named INSIGHTS_CONTROLLER_CONFIG_FILE. For example:

export INSIGHTS_CONTROLLER_CONFIG_FILE=~/config.toml
./insights-operator-controller

Environment variables

• INSIGHTS_CONTROLLER_CONFIG_FILE - custom path to config file (default: ./config.toml)
• CONTROLLER_ENV - specify environment (development, test, production. Default: development)
• CONTROLLER_PREFIX - specify URL path prefix (Default: /api/v1/)

LDAP Authentication

For authentication we using Insights operator LDAP Auth that are working as proxy between client and controller. For turning on authentication need set export CONTROLLER_ENV=production, by default for development and test purposes it turned off.

Data storage

Data storage used by the service is configurable via the command line parameters. Currently it is possible to configure the following data storages:

• SQLite local database: controller.db for the local deployment and test.db for functional tests
• PostgreSQL database: for local deployment and to be able to deploy the application to developer development

ER Diagram

Insights operator database

SQLite

Use the following scripts from the local_storage subdirectory to work with SQLite database:

• create_database_sqlite.sh to create new database stored in file controller.db
• create_test_database_sqlite.sh to create new database stored in file test.db, this database will be used by tests

PostgreSQL

PostgreSQL needs to be setup correctly:

• User postgres should have password set to postgres
• In the configuration file /var/lib/pgsql/data/pg_hba.conf, the method md5 needs to be selected for user postgres and all
• The PostgreSQL daemon (service) has to be started, of course: sudo systemctl start postgresql

For more information how to install PostgreSQL on Fedora (or RHEL) machine, please follow this guide: https://computingforgeeks.com/how-to-install-postgresql-on-fedora/

Use the following scripts from the local_storage subdirectory to work with PostgreSQL database:

• create_database_postgres.sh to create new database named controller
• create_test_database_postgres.sh to create new database named test_db

The following two scripts can be used to drop existing database(s):

• drop_database_postgres.sh to drop database named controller
• drop_test_database_postgres.sh to drop database named test_db

---

You can also run PostgreSQL containerized with the prepared Dockerfile local_storage/Dockerfile.postgres

• there is a simple shortcut script local_storage/dockerize_postgres.sh which stops the container if it exists, then builds and runs a detached image.
• to automatically fill the database with test data, set an environment variable MOCK_DATA=true and pass it to docker run
• the environment variables from the original postgres image, such as PGDATA also work

RDS AWS PostgreSQL

To set up a database on RDS AWS, an AWS account is needed. To set up the AWS account follow instructions: https://aws.amazon.com/premiumsupport/knowledge-center/create-and-activate-aws-account/

After AWS account is set up, follow instructions to set up a PostreSQL database instance here : https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_CreatePostgreSQLInstance.html Set up the database instance as needed. Set the name of the database as postgres. (Not the DB instance identifier.)

When database instance status becomes available, set environment variables $RDS_MASTERUSER, $RDS_MASTERPASSWORD and $RDS_ENDPOINT with values master username, master password, and endpoint including port, of your database instance. Run create_RDS_database_postgres.sh to create database.

To drop previously created database, run drop_RDS_database_postgres.sh.

Testing

Unit tests

The following command runs all unit tests against newly built controller:

make test

In case you just need to start unit tests without the clean+build step, use following command:

go test ./...

It is also possible to increase verbosity level:

go test -v ./...

REST API tests

REST API tests need the running service and the test database to be prepared. In order to perform REST API tests, start the following script:

./test.sh

Please note that the service should not be running at the same moment (as it used the same port).

BDD tests

Behaviour tests for this service are included in Insights Behavioral Spec repository. In order to run these tests, the following steps need to be made:

1. clone the Insights Behavioral Spec repository
2. go into the cloned subdirectory insights-behavioral-spec

Code style and cyclomatic complexity checks

All code style checks, cyclomatic complexity measurement etc. can be started from command line by using:

make style

CI

Travis CI is configured for this repository. Several tests and checks are started for all pull requests:

• Unit tests that use the standard tool go test
• go fmt tool to check code formatting. That tool is run with -s flag to perform following transformations
• go vet to report likely mistakes in source code, for example suspicious constructs, such as Printf calls whose arguments do not align with the format string.
• golint as a linter for all Go sources stored in this repository
• gocyclo to report all functions and methods with too high cyclomatic complexity. The cyclomatic complexity of a function is calculated according to the following rules: 1 is the base complexity of a function +1 for each 'if', 'for', 'case', '&&' or '||' Go Report Card warns on functions with cyclomatic complexity > 9
• gosec to inspect source code for security problems by scanning the Go AST
• ineffassign to detect and print all ineffectual assignments in Go code
• errcheck for checking for all unchecked errors in go programs
• shellcheck to perform static analysis for all shell scripts used in this repository
• abcgo to measure ABC metrics for Go source code and check if the metrics does not exceed specified threshold

History of checks done by CI is available at RedHatInsights / insights-operator-controller.

Contribution

Please look into document CONTRIBUTING.md that contains all information about how to contribute to this project.

Please look also at Definition of Done document with further informations.

Package manifest

Package manifest is available at docs/manifest.txt.