BSIS

This page describes how to deploy a BSIS infrastructure

Configuration

First we need to create an configuration (the easiest way is to create your own configuration by adapting an existing one)

For BSIS, there are two main sections in the configuration

  • shiny: the R-Shiny specific setups

  • user: user sepecfic information

  • aws: AWS specific setups

Configuration (Shiny):

For shiny, we need to define where the Shiny application comes from. How to create a Shiny on S3 can be obtained here.

An simple example is given below:

shiny:
   names:
      hello_world: hello_world/link1
   url: https://github.com/jzanetti/shiny_aws_examples.git
   branch: main
   cred: null
The above provides some basic shiny information:

  • the Shiny server takes the main branch from https://github.com/jzanetti/shiny_aws_examples.git.

  • hello_world is the only application we want to host, and it will use the link name hello_world/link1 (e.g., the IP for this app will be <IP Address>/hello_world/link1)

  • there is no credentials needed

For private repository we can apply a private token to access the repository. For example, we can have the credentials to access the repository as below:

shiny:
   names:
      hello_world: hello_world/link1
   url: https://github.com/jzanetti/shiny_aws_examples.git
   branch: main
   cred:
      user: <user-name>
      token: <token>

Configuration (AWS):

For aws, we need to provide usual AWS configurations such as VPC, subnets and the ami to be used. An example is given below:

aws:
   ImageId: ami-xxxxxxxxxxxxxxx
   KeyName: xxxxxxxxx-key
   SecurityGroupIds:
      - sg-xxxxxxxxx
   SubnetId: subnet-xxxxxxxxx
   InstanceType: t2.medium
   Placement:
      AvailabilityZone: ap-southeast-2a
   IamInstanceProfile:
      Arn: arn:aws:iam::xxxxxxxxx:instance-profile/ShinyApp_role
   BlockDeviceMappings:
      DeviceName: /dev/sda1
      Ebs:
         VolumeType: gp2
         DeleteOnTermination: true
         VolumeSize: 15

Configuration (user):

We will have to provide a few user defined parameters:

user:
   elastic_ip: eipalloc-xxxxxxxxx
   authentication: true
   spot_price: 0.1

Where:

  • elastic_ip: this can be set to null if we don’t want to use AWS Elastic IP.

  • authentication: we can set up authentication (using Nginx) for our server (setup for nginx can be found Authentication.).

  • spot_price: Spot price we are willing to pay in the EC2 spot market.

An full example for the BSIS configuration can be found at etc/cfg/bsis.yml

Deployment

The BSIS can be deployed under the environment shiny_aws:

conda activate shiny_aws
start_bsis --cfg /tmp/bsis.yml --workdir /tmp --lifespan 60

Where --cfg is the configuration for BSIS and --workdir is the working directory holds all the intermediate files. lifespan indicates how long the instance will be active (in minutes). By default, lifespan is 60 minutes, while in production we can set it to unlimited (e.g., being defined by UNLIMITED_LIFESPAN_FLAG). Please set lifespan carefully since AWS charges us every seconds we use the server.

After the deployment, we should be able to find an EC2 instance in the console with the shiny application being installed.

Using HTTPS

By default, BSIS in SHINY_AWS will provide a link with http, however, for accessing the link securely, https is usually recommendded. Details for setting up https with BSIS can be found here