openblockchain

Architecture

The project is split into several services:

Cassandra persists the data: blocks, transactions, and visualisations (analysed data).
Bitcoin is used to connect to the Bitcoin blockchain. It's a simple Bitcoin Core node whose role is to index all the blocks and transactions and make them consumable through a HTTP JSON RPC interface.
Scanner connects to the bitcoin service through its APIs and stores all the blocks and transactions in the Cassandra database.
Spark analyses the Bitcoin blockchain data stored in Cassandra.
API is a REST interface that allows clients to consume the data stored in Cassandra.
Frontend is the web app used to explore the blockchain and visualise analysed data.

Each service contains 1 or more containers and can be scaled independently from each other.

All docker images are hosted: https://hub.docker.com/u/openblockchaininfo/

Introduction to the following instructions

The following instructions will walk you through installing the entire openblockchain application including all system dependencies and requirements allowing you to run openblockchain from most base systems (Linux, OSX, Windows).

The instructions will start by guiding you through installing Ubuntu on VirtualBox.

System requirements

Due to the computing power and disk space requirements, the data storage and processing for the project was designed to be run on several machines, preferably in a cloud with local management from a host machine.

Prerequisites

The following instructions guide you through setting up your host machine with Ubuntu and docker-engine, docker-machine and docker-compose.

Setting up the host machine

First you need to install VirtualBox by following these instructions

You then need to download Ubuntu 16.04.1 LTS - 64 bit by following these instructions

Creating an Ubuntu Virtual Machine

You need to create a new virtual machine using VirtualBox. The virtual machine (hereafter referred to as 'the host machine') will be used to deploy the application. The host machine requires:

Ubuntu 64 bit
RAM: 2GB
CPU: Any
Disk: 20GB

When you start the host machine for the first time select Ubuntu 16.04.1 LTS - 64 bit (which you downloaded in the previous step) as the start-up disk (it is possible to run the project on other versions of Ubuntu and even other operating systems such as Windows and OS X, but instructions for those have not been included here).

The host machine will only be used to send commands to the cluster.

The cluster machines (where the services will be deployed) require (each):

RAM: 4GB+
CPU: 4+ cores
Disk: 150GB+

Once you set up the host machine (see the following "Prerequisites" section), the "Cluster Setup" instructions will guide you through the steps required to create 3 machines on Scaleway.

Installing the prerequisites on the host machine

These must be installed on the host machine only. Make sure that your host machine has Ubuntu installed (which you will have if you have followed the previous steps).

1. Install Docker Engine

Docker Engine includes the docker client. docker will be used to communicate with the docker daemons which run on every machine in the cluster.

Update package information. Ensure that APT works with the https method, and that CA certificates are installed:

$ sudo apt-get update
$ sudo apt-get install apt-transport-https ca-certificates

Add the official GPG key and APT repository:

$ sudo apt-key adv --keyserver hkp://p80.pool.sks-keyservers.net:80 --recv-keys 58118E89F3A912897C070ADBF76221572C52609D
$ sudo echo "deb https://apt.dockerproject.org/repo ubuntu-wily main" > /etc/apt/sources.list.d/docker.list

Update the APT package index and install Docker:

$ sudo apt-get update
$ sudo apt-get install docker-engine

Start the docker daemon and verify it's installed correctly:

$ sudo service docker start
$ sudo docker run hello-world

Always up-to-date instructions are available at docker.com.

2. Install Docker Machine

docker-machine is used to provision machines for the cluster.

Download the Docker Machine 0.8 binary and extract it to your PATH:

$ sudo curl -L https://github.com/docker/machine/releases/download/v0.8.0/docker-machine-`uname -s`-`uname -m` > /usr/local/bin/docker-machine
$ sudo chmod +x /usr/local/bin/docker-machine

Check the installation by displaying the Machine version:

$ docker-machine version

Always up-to-date instructions are available at docker.com.

3. Install Docker Compose

Docker Compose is used to manage the services.

$ sudo curl -L https://github.com/docker/compose/releases/download/1.8.0/docker-compose-`uname -s`-`uname -m` > /usr/local/bin/docker-compose
$ sudo chmod +x /usr/local/bin/docker-compose

Check the installation by displaying the Compose version:

$ docker-compose --version

Always up-to-date instructions are available at docker.com.

Congratulations!... and a recap

You have now configured your host machine. You have:

installed VirtualBox
installed Ubuntu 16.04.1 LTS 64-bit on a virtual machine (the host machine)
installed Docker Engine
installed Docker Machine
installed Docker Compose

This will allow you to manage all the microservices which constitue the openblockchain application from within the host machine.

Cluster Setup

The following instructions guide you through deploying the cluster in the cloud, using Scaleway. The cluster will contain the following services:

3 instances of Cassandra
1 Bitcoin node
1 Spark master
2 Spark workers
1 Spark job submitter
1 scanner container
1 api container
1 frontend container

The microservices are managed via docker-* on your host machine.

1. Create the discovery service

Using Scaleway

In order to use Scaleway, you'll first need to install the Scaleway driver for Docker Machine:

$ wget https://github.com/scaleway/docker-machine-driver-scaleway/releases/download/v1.2.1/docker-machine-driver-scaleway_1.2.1_linux_amd64.tar.gz
$ tar -xvf docker-machine-driver-scaleway_1.2.1_linux_amd64.tar.gz
$ sudo cp docker-machine-driver-scaleway_1.2.1_linux_amd64/docker-machine-driver-scaleway /usr/local/bin/
$ sudo chmod +x /usr/local/bin/docker-machine-driver-scaleway

Create a machine to host a Consul discovery service:

$ docker-machine create -d scaleway \
  --scaleway-commercial-type=C2S --scaleway-name=obc-consul \
  --scaleway-organization=<SCALEWAY-ACCESS-KEY> --scaleway-token=<SCALEWAY-SECRET-KEY> \
  obc-consul

Point your docker client to the new machine:

$ eval $(docker-machine env obc-consul)

Create and start the service:

$ docker run --name consul --restart=always -p 8400:8400 -p 8500:8500 \
  -p 53:53/udp -d progrium/consul -server -bootstrap-expect 1 -ui-dir /ui

2. Launch the Docker Swarm (cluster)

Using Scaleway

You can provision the machines from Scaleway, which provides affordable high-end servers that are perfect for this project.

Get your Scaleway credentials (access key and secret key) from this page, then deploy 3 machines:

$ docker-machine create -d scaleway \
  --scaleway-commercial-type=C2L --scaleway-name=obc \
  --scaleway-organization=<SCALEWAY-ACCESS-KEY> --scaleway-token=<SCALEWAY-SECRET-KEY> \
  --swarm --swarm-master \
  --swarm-discovery consul://`docker-machine ip obc-consul`:8500 \
  --engine-opt cluster-store=consul://`docker-machine ip obc-consul`:8500 \
  --engine-opt cluster-advertise=eth0:2376 \
  obc

$ docker-machine create -d scaleway \
  --scaleway-commercial-type=C2L --scaleway-name=obc-01 \
  --scaleway-organization=<SCALEWAY-ACCESS-KEY> --scaleway-token=<SCALEWAY-SECRET-KEY> \
  --swarm \
  --swarm-discovery consul://`docker-machine ip obc-consul`:8500 \
  --engine-opt cluster-store=consul://`docker-machine ip obc-consul`:8500 \
  --engine-opt cluster-advertise=eth0:2376 \
  obc-01

$ docker-machine create -d scaleway \
  --scaleway-commercial-type=C2L --scaleway-name=obc-02 \
  --scaleway-organization=<SCALEWAY-ACCESS-KEY> --scaleway-token=<SCALEWAY-SECRET-KEY> \
  --swarm \
  --swarm-discovery consul://`docker-machine ip obc-consul`:8500 \
  --engine-opt cluster-store=consul://`docker-machine ip obc-consul`:8500 \
  --engine-opt cluster-advertise=eth0:2376 \
  obc-02

By default the Scaleway machines have a 50GB disk. However, Scaleway also attaches a 250GB SSD, which needs to be mounted manually:

$ printf "obc\nobc-01\nobc-02" | \
  xargs -n 1 -I CONT_NAME docker-machine ssh CONT_NAME \
  "echo ; echo ; echo CONT_NAME ;"`
  `"echo 'Formatting...' ;"`
  `"yes | mkfs -t ext4 /dev/sda ;"`
  `"echo 'Mounting...' ;"`
  `"mkdir -p /openblockchain ;"`
  `"mount /dev/sda /openblockchain ;"`
  `"echo 'Adding to fstab...' ;"`
  `"echo '/dev/sda /openblockchain auto defaults,nobootwait,errors=remount-ro 0 2' >> /etc/fstab ;"`
  `"echo 'Done' ;"

This additional storage will not be visible within the UI of your Scaleway account. You can manually check to verify the additional storage has been mounted (in this case checking obc):

$ docker-machine ssh obc
$ cat /etc/fstab
$ ls -al /openblockchain

3. Verify the swarm

Point your Docker environment to the machine running the swarm master:

$ eval $(docker-machine env --swarm obc)

Print cluster info:

$ docker info

You can also list the IPs that Consul has "discovered":

$ docker run --rm swarm list consul://`docker-machine ip obc-consul`:8500

Launch the services

1. Create the default network

Create an overlay network that allows containers to "see" each other no matter what node they are on:

$ docker network create --driver overlay --subnet 10.0.9.0/24 obcnet

List all networks, obcnet should be shown as overlay:

$ docker network ls

Getting openblockchain code

openblockchain has been designed with multiple microservices which are linked in the main Docker engine as git submodules.

The following command clones the project and all required gitmodules.

$ git clone --recursive https://github.com/open-blockchain/openblockchain.git

Once you have cloned the repo make sure to go into the directory.

$ cd openblockchain

2. Up!

From the root folder (which contains docker-compose.yml), run:

$ docker-compose up -d --build

Docker will deploy all the services. By default, it'll launch:

3 instances of Cassandra
1 Bitcoin node
1 Spark master
2 Spark workers
1 Spark job submitter
1 scanner container
1 api container
1 frontend container

3. Check the services

Show all containers and their status (they should all be "Up"):

$ docker-compose ps

Show the Cassandra cluster status (you should see 3 nodes):

$ docker-compose exec cassandra-seed nodetool status

In the mean time, the bitcoin server should've already scanned a few hundred blocks. Let's see it working live:

$ docker-compose logs -f --tail 100 bitcoin

You should see something like this, which means that it downloaded all the blocks from 0 to 187591 (in this example):

bitcoin_1    | 2016-08-25 15:03:16 UpdateTip: new best=00000000000006edbd5920f77c4aeced0b8d84e25bb5123547bc6125bffcc830  height=187589  log2_work=68.352088  tx=4696226  date=2012-07-05 03:45:13 progress=0.015804  cache=48.7MiB(122719tx)
bitcoin_1    | 2016-08-25 15:03:16 UpdateTip: new best=00000000000003f07c1227f986f4687d291af311a346f66247c504b332510931  height=187590  log2_work=68.352117  tx=4696509  date=2012-07-05 03:52:33 progress=0.015805  cache=48.7MiB(122815tx)
bitcoin_1    | 2016-08-25 15:03:16 UpdateTip: new best=00000000000002b34a755d25a5fee4c0ad0c018cf94f4b0afef6aabe823d304a  height=187591  log2_work=68.352146  tx=4696593  date=2012-07-05 04:01:11 progress=0.015805  cache=48.7MiB(122839tx)

Now let's see the Spark dashboards, which should be empty right now. Run the following command to get the dashboard URL of each Spark master and worker:

$ echo "master: "`docker-machine ip obc`":8080"
$ echo "worker 1: "`docker-machine ip obc-01`":8081"
$ echo "worker 2: "`docker-machine ip obc-02`":8081"

Analyse data and test the webapp

1. Scan the blockchain

Show the scanners' logs:

$ docker-compose logs -f --tail 100 scanner-01
$ docker-compose logs -f --tail 100 scanner-02

The scanners were configured (in docker-compose.yml) so that they index a fixed range of blocks. However, in the beginning, when the bitcoin node hasn't yet reached block 330,000, scanner-02 (which is configured to scan blocks from 330,000 to 400,000) will not be able to scan anything. Therefore, it'll wait 10 minutes then try again until those blocks are available.

2. Spark analysis

The spark-submit service runs multiple Spark scripts every hour. These scripts generate visualizations (i.e. data points) that the API will provide to the frontend. You can see its progress by looking at the logs:

$ docker-compose logs -f --tail 100 spark-submit

Let's run a custom script which counts all the blocks and transactions in Cassandra:

$ docker-compose exec spark-submit sbt submit-Counter

When the script finishes, you should see something like this (ignore the logs that begin with a timestamp):

Blocks: 239699
Transactions: 18941698

3. Access the API

Find the name of the machine the api container is running on:

$ docker ps --filter "name=node-api" --format "{{.Names}}"

This will output something like obc-02/api, where the part before the slash is the machine name.

Now get the URL of the api, replacing <machine-name-from-above> accordingly:

$ echo "http://"`docker-machine ip <machine-name-from-above>`":10010"

4. Access the web app frontend

Find the name of the machine the frontend container is running on:

$ docker ps --filter "name=frontend" --format "{{.Names}}"

This will output something like obc-01/frontend, where the part before the slash is the machine name.

Now get the URL of the frontend, replacing <machine-name-from-above> accordingly:

$ echo "http://"`docker-machine ip <machine-name-from-above>`":80"

5. Visualizations

Depending on whether the spark-submit service had time to finish the Spark scripts or not, you might see visualizations in the frontend, at http://<frontend-ip>:80/blocks.

If visualizations are not yet available, check the progress of spark-submit:

$ docker-compose logs -f --tail 100 spark-submit

If it's idling, restart the service (docker-compose restart spark-submit). Otherwise, it's probably processing data from Cassandra. Let it finish the job and refresh the page (go away and make a cup of tea - in 60 mins it will certainly have crunched some numbers).

License

Designed, developed and maintained by Dan Hassan [email protected]

This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU Affero General Public License for more details.

You should have received a copy of the GNU Affero General Public License
along with this program.  If not, see <http://www.gnu.org/licenses/>.