/gcp-concourse

Primary LanguageShell

PCF on GCP

Concourse Pipeline

This pipeline uses Terraform to create all the infrastructure required to run a 3 AZ PCF deployment on GCP.

To use it, you'll need to change all of the CHANGEME values in params.yml.

You'll also need to enable:

  • GCP Compute API here
  • GCP Storage API here
  • GCP SQL API here
  • GCP DNS API here
  • GCP Cloud Resource Manager API here

This pipeline downloads artifacts from DockerHub, GitHub, and the configured S3-compatible object store, and as such the Concourse instance must have access to those. You can use AWS S3 as your S3-compatible object store, but note that Terraform outputs a .tfstate file that contains plaintext secrets. For this reason Minio is preferrable to keep the visibility of the .tfstate local to Concourse.

If you don't already have Concourse running:

vagrant init concourse/lite
vagrant up
fly -t lite login -c http://192.168.100.4:8080

If you want to use Minio as your S3-compatible object store:

OSX:

docker run -e MINIO_ACCESS_KEY="example-access-key" \
           -e MINIO_SECRET_KEY="example-secret-key" \
           minio/minio server /tmp

Linux:

docker run -e MINIO_ACCESS_KEY="example-access-key" \
           -e MINIO_SECRET_KEY="example-secret-key" \
           --detach \
           --network host \
           minio/minio server /tmp

Set the pipeline:

fly -t lite set-pipeline -p deploy-pcf -c pipeline.yml -l params.yml

Usage

Unpause the pipeline if you haven't already.

upload-opsman-image will automatically upload the latest matching version of Operations Manager.

Once that is complete you can trigger the create-infrastructure job. create-infrastructure will output at the end the DNS settings that you must configure before continuing.

Once DNS is set up you can run configure-director. From there the pipeline should automatically run through to the end.

Tearing down the environment

There is a job, wipe-env, which you can run to destroy the infrastructure that was created by create-infrastructure. If you want to bring the environment up again, run create-infrastructure. This can also be used if create-infrastructure fails for some reason, where Terraform creates only some of the infrastructure.