Privacy Aware Messaging System (PAMS)

Introduction

This repository contains helm chart and associated services for the creation of the cloud components required by IBM-FL to execute in a pubsub system, as per example1, example2, and example3.

These services can also be used to provide a safer communication infrastructure for other scenarios where parties require to communicate in a private manner using pycloudmessanger.

Requirements

• Docker
• Helm version >= 3
• Minikube version >= 1.26

Setup (on minikube)

1. Install minikube:

• OS X: brew install minikube
• Linux: XXXX

2. Install helm

• OS X: brew install helm
• Linux: XXXX
• Windows: XXXX

3. Start minikube $ minikube start

Install Chart

Install the chart using the following command:

helm install -f myvalues.yaml funny-test ./privacy-aware-messaging-system

where:

• funny-test is the release name, it can be anything that makes the current installation unique within the target system
• myvalues.yaml is a (optional) file that contains the overload of specific default values.

An example of content of myvalues.yaml is as follows. Note that the values specified in the dictionary at address utility.registryCredentials in myvalues.yaml need to be updated according to the user's artifactory credentials, thus replacing the placeholders EMAIL and KEY with the appropriate values as specified in the Artifactory website.

# Docker registry
docker:
  registry:
    name: ""
    username: ""
    password: ""
  timezone: "UTC"
# Database configuration
utility:
  db:
    imageName: "res-drl-docker-local.artifactory.swg-devops.com/privacy-aware-messaging-system-utility"
    imageTag: "postgres"
  registryCredentials:
    registry: res-drl-docker-local.artifactory.swg-devops.com
    username: EMAIL
    password: KEY
    email: EMAIL
service:
  imageName: "res-drl-docker-local.artifactory.swg-devops.com/privacy-aware-messaging-system-services"
db:
  type: "postgres"
  name: "MYDATABASE"
  imageName: "postgres"
  imageTag: "14.4-alpine"
  imagePullPolicy: "IfNotPresent"
  replicaCount: 1
  restartPolicy: "Always"
  host: null
  port: 5432
  protocol: "http"
  schema_name: "MYSCHEMA"
  auth:
    username: "username"
    password: "password"
  persistence:
    size: 2Gi
rabbitmq:
  external: false
  auth:
    username: "username"
    password: "password"
# S3/COS configuration
s3:
  auth:
    username: "username"
    password: "password"

Notes

• When running in minikube on Apple M1 and if the desired DB is DB2, please follow these steps prior to installation:

eval $(minikube docker-env)
docker pull ibmcom/db2:11.5.7.0a --platform amd64

These commands will configure docker to use minikube docker environment, and retrieve the amd64 version of the DB2 image. After that, minikube will not attempt to fetch the image from DockerHub since it already have one with the correct tag name in its local registry.

DEBUG INSTRUCTION

Setup Backend Cluster

• Start your docker client and minikube: minikube start
• Build a local docker utility image (temporary hack while we finish optimizing the setup)
  • Tell your local docker installation to use minikube as docker engine: eval $(minikube docker-env)
  • In <ROOT_DIR>/utils run docker build -t privacy-aware-messaging-system-utility:postgres -f Dockerfile.postgres .
• Now lets build the back end. In <ROOT_DIR>:
  • Install all kubernetes modules: helm install pams ./privacy-aware-messaging-system
  • Check that all minikube modules status are set to Running: kubectl get all

Setup Python Module

• In <ROOT_DIR>/pams-services/ run:

• pipenv install

• pipenv install --dev

• Lets make sure everything's running fine

• pipenv shell

• pytest