Refgenieserver deploy guide

We're running 4 server instances on an AWS ECS cluster. These are available at:

Primary server

service-name: rgs-service-primary
url: http://refgenomes.databio.org
repository: http://github.com/refgenie/refgenomes.databio.org
update: The primary server is deployed with change to the config/refgenie_config_archive.yaml on master branch. It uses the docker.io/databio/refgenieserver:latest image as base.

Data staging

service-name: rgs-service-software-staging
url: http://rg.databio.org:81
repository: http://github.com/refgenie/data_staging
update: The data-staging server is deployed push to master branch of the data_staging repository
description: Uses the release software to show how some new test data would render.

Software staging

service-name: rgs-service-data-staging
url: http://rg.databio.org:82
repository: http://github.com/refgenie/refgenieserver
update: The software-staging server is deployed on push or PR to the staging branch. It deploys to docker.io/databio/refgenieserver:staging.
description: Uses the released data to show how new software would render.

Basic demo

service-name: rgs-service-demo
url: http://rg.databio.org:83
repository: http://github.com/refgenie/server_deploy_demo
update: Deployed with change to the config archive file in /config.
description: This is a basic self-contained demo that shows the entire process, from data to AWS container service.

Two-stage container strategy

We use a 2-stage container building strategy:

we build the refgenieserver application with no data, which lives at docker.io/databio/refgenieserver.
for deploying a server, we start from the docker.io/databio/refgenieserver image and add a config file, building a new image that we can then deploy.