This repository contains Apache Brooklyn blueprints for a Hyperledger Fabric v0.6.1 cluster deployment using the official Docker images.

Note: If you select the Vagrant installation, replace the following line in the guide below:

git clone https://github.com/cloudsoft/amp-vagrant.git

with:

git clone -b hyperledger-fabric https://github.com/cloudsoft/amp-vagrant.git

Use this guide to install AMP.

For the simplest path to a Hyperledger Fabric deployment, select the Vagrant installation option and create 5 BYON nodes.

Note: You can skip this step if you selected the Vagrant installation option.

• Download the latest release JAR from this repository and place it in the deploy folder inside the AMP install directory.
• Browse to the AMP user interface (address listed in the installation guide)
• Click the button with nine small squares on the top right
• Click "Blueprint importer"
• Copy and paste the contents of hyperledger.bom into the editor
• Click "Import" on the bottom right

Note: You can skip this step if you selected the Vagrant installation option and started the 5 BYON nodes, in which case you will already have a location named byon-cluster configured and ready to be used. If you would like to deploy to a different location, proceed with this step.

Use this guide to create a deployment location.

• Browse to the AMP user interface (address listed in the installation guide)
• Scroll through the "Quick launch" tile and click "Hyperledger Fabric Cluster"
• Optionally provide a name
• Provide the location that you created (or byon-cluster by default, if using Vagrant)
• Increase the number of peers if desired
• Click the "Deploy" button

Note: If you plan on deploying to Vagrant VMs running locally, the maximum supported value of hyperledger.peers.per.location is 4. However, the size of the cluster can be increased if you add more VMs to servers.yaml in your amp-vagrant repository.

This deployment is capable of creating multiple clusters of validating peer nodes across multiple locations; all of the nodes are part of the same Hyperledger Fabric.

• Download the Brooklyn CLI using this guide
• Login to your AMP server with the CLI using this guide
• Copy and paste the contents of the multi-cluster blueprint into an editor, add the names of your configured location(s), and save the modified file to disk
• Deploy the blueprint with the CLI using this guide -- webapp-policy.yaml should be replaced with the path to your blueprint file on disk

Note: This deployment requires at least 5 locations listed (can contain duplicates).

Once your cluster has successfully deployed, perform the following steps to deploy the asset management demonstration app. This app repeatedly assigns an asset "Picasso" from one owner to another. For more information about this app as well as its source code, see the Fabric repository.

• Browse to the AMP user interface (address listed in the installation guide)
• Click the "App Inspector" tile
• Hover over the down arrow next to your Hyperledger Fabric cluster and press the listed key combination to "expand all children"
• Click "CLI Node"
• Click the "Effectors" tab
• Click "Invoke" next to "Run Demo Application"

• Browse to the AMP user interface (address listed in the installation guide)
• Click the "App Inspector" tile
• Hover over the down arrow next to your Hyperledger Fabric cluster and press the listed key combination to "expand all children"
• Click "Expand All"
• Click "CLI Node"
• Click the "Sensors" tab
• Copy the host.sshAddress value
• Open up your terminal and run command: ssh <ssh-address-here>

Note 1: You may need to supply an SSH key or a username / password depending on your deployment location's configuration.

Note 2: If host.sshAddress ends with a port (e.g. :22), remove the colon and the port from the SSH command.

Note 3: If you deployed to local Vagrant VMs, you can SSH into any of these VMs by running this command instead: vagrant ssh byon<number here>. The name of the VM is based on the last digit of the IP address. For example, if the CLI node's IP is 10.10.10.102 then the command would be: vagrant ssh byon2.

From the same terminal window from the previous step, execute the following commands:

sudo docker exec -it cli bash
cd $APP_HOME
go build
./app

This enters the CLI container, builds the app, and executes the app. When the app runs, the output should clearly indicate the transfer of ownership of "Picasso" ultimately ending with "Dave" as the owner.

If deploying to AWS You should use an up-to-date CentOS AMI. The "CentOS 7 (x86_64) with Updates HVM" images (link) are a good choice. The following is a sample catalogue blueprint for the us-east-1 version of this AMI:

brooklyn.catalog:
  id:       aws-virginia-centos7
  name:     AWS Virginia CentOS 7
  itemType: location
  item:
    type: jclouds:aws-ec2
    brooklyn.config:
      identity:   <YOUR IDENTITY>
      credential: <YOUR CREDENTIAL>
      region:     us-east-1
      imageId:    us-east-1/ami-6d1c2007
      minRam:     2000
      loginUser:  centos
      installDevUrandom: true
      allocatePTY: true