Getting Started with Spinnaker

These instructions cover pulling Spinnaker from source and setting up to run locally against Amazon Web Services and/or Google Cloud Platform accounts. ( If you would prefer to try a container based solution, see docker-compose installation )

We will clone into $SPINNAKER_HOME and create that as our working directory, including this repo for configuration scripts, as well as the various service repos.

Note If you are going to create a Virtual Machine in Amazon EC2 or Google Compute Engine for your development, then a reasonable machine type is m4.2xlarge (Amazon) or n1-standard-8 (Google). If using Google, you will need to add "Read Write" Compute access scope when creating the instance, and may also wish to add "Full" Storage scope to later write releases to Google Cloud Storage buckets. The Amazon credentials are discussed below in Configure your AWS Account.

These scripts are tested on: * Ubuntu 14.04 LTS * Mac OS X 10.11

Get the bootstrap / configuration repo

#export SPINNAKER_HOME=/path/to/your/Spinnaker/workspace
mkdir -p $SPINNAKER_HOME
cd $SPINNAKER_HOME
git clone git@github.com:spinnaker/spinnaker.git

Configure your environment

The Spinnaker platform has a few prerequisites, which are installed as a part of this configuration process. They are:

MacOSX

You need to have homebrew installed and ensure your version of git is above 2.0.

brew install redis cassandra brew-cask packer
brew cask install java
cd $SPINNAKER_HOME
mkdir build
cd build
../spinnaker/dev/refresh_source.sh --pull_origin --use_ssh --github_user default

Debian-linux

cd $SPINNAKER_HOME
spinnaker/dev/install_development.sh --package_manager
spinnaker/dev/bootstrap_dev.sh

The --package_manager argument requests the Debian Package Manager be used to the greatest extent possible. This permits adding new source repositories where package dependencies are distributed from non-standard repositories (e.g. cassandra). The alternative is --nopackage_manager to use the Package Manager for standard dependencies but explicitly install packages that come from other sources. --nopackage_manager is intended for installations where IT policies discourage or forbid adding additional source locations.

The bootstrap_dev.sh script will ask to install additional components. packer is only needed if you plan on building VM images. gcloud is only needed to write releases to Google Cloud Storage, but is convenient to have if you plan on using or accessing Google Cloud Platform resources from your development environment. These could be installed at a system level rather than user level, but the default install requires updating your path so is performed here.

Configure Spinnaker

We will create a directory for Spinnaker configuration overrides, copy the default configuration template there, and edit to select the appropriate cloud provider(s).

cd $SPINNAKER_HOME
mkdir -p $HOME/.spinnaker
cp spinnaker/config/default-spinnaker-local.yml $HOME/.spinnaker/spinnaker-local.yml
chmod 600 $HOME/.spinnaker/spinnaker-local.yml

Edit $HOME/.spinnaker/spinnaker-local.yml and set the enabled option for the cloud provider(s) of your choice.

Configure your AWS Account

If you enabled AWS for Spinnaker, there are some requirements for the AWS account:

Decide which region you want Spinnaker to index. In $HOME/.spinnaker/spinnaker-local.yml fill in that value in providers.aws.defaultRegion. (The default is us-east-1).

Sign into the AWS console, and select the region Spinnaker will manage.

Click on Networking > VPC
Name your vpc (edit the name tag, and give it a value with no spaces or dots in the name) (e.g. defaultvpc)
Name your subnets (edit the name tag and name following the pattern vpcName.internal.<availabilityZone>)
- e.g. defaultvpc.internal.us-east-1a, defaultvpc.internal.us-east-1b, defaultvpc-internal.us-east-1c
  - Create an EC2 role called BaseIAMRole
- Console > Identity & Access Management > Roles > Create New Role. Select Amazon EC2.
- You don’t have to apply any policies to this role. EC2 instances launched with Spinnaker will have this role associated.
  - Create an EC2 keyPair for connecting to your instances.
- Console > EC2 > Key Pairs > Create Key Pair. Name the key pair my-aws-account-keypair (my-aws-account should match the aws.primaryCredentials.name in $HOME/.spinnaker/spinnaker-local.yml, so if the name is awsprod, the name for the key pair should be awsprod-keyspace).
  - Create AWS credentials for Spinnaker
- Console > Identity & Access Management > Users > Create New Users. Enter a username.
- Create an Access Key for the user. Save the access key and secret key into ~/.aws/credentials as shown here. Alternatively, add the keys to $HOME/.spinnaker/spinnaker-local.yml
- Edit the users Permissions.
- Attach a Policy to the user granting PowerUserAccess.
- Create an inline policy for IAM granting PassRole on the resource '*' <img width="500" src="https://cloud.githubusercontent.com/assets/74310/11158316/5dd1e86a-8a0c-11e5-8dc9-c06a6ec616cf.png">

Configure your Google Cloud Platform Account

If you enabled Google for Spinnaker, perform the following steps for your project in the Google Developer’s Console:

Enable APIs (click 'Enable API' button for each):
- Compute Engine
- Compute Engine Autoscaler
Add and Obtain Credentials:
- Navigate to Credentials
- New credentials > Service account key
- Select… > New service account
- Provide a Name and click Create
- chmod 400 the file that downloads
Add GCP credentials in $HOME/.spinnaker/spinnaker-local.yml:
- Set project ID for provider.google.primaryCredentials.project
- Set full absolute path of downloaded file for providers.google.primaryCredentials.jsonPath

Start Spinnaker Services

( If you’re running on Mac OS, please make sure that your redis server and cassandra server is up and running. Make sure that cqlsh is accessible in your path and you can connect to cassandra by calling cqlsh in your terminal ).

cd $SPINNAKER_HOME/build
../spinnaker/dev/run_dev.sh [service]

If a service is provided, then just that one service will be started. If no service is provided, then all the services will be started (including redis and cassandra unless they are specified with a remote host). If a service is already running (even if not yet available) then it will not be restarted.

Note run_dev.sh might get stuck waiting on a service to start. Hitting CTRL-C just stops the waiting on service it doesn’t terminate the services. If it seems stuck stop and restart run_dev.sh.

Stop Spinnaker Services

cd $SPINNAKER_HOME/build
../spinnaker/dev/stop_dev.sh [service]

If a service is provided, then just that one service will be stopped. If no service is provided then all the spinnaker services will be stopped. Cassandra and redis are not affected by stop_dev.sh

Xeth-Waxman / spinnaker