Data Server

Deploy an ASGI server and API to receive and explore test results. When deployed, you can access the server's current API at the /docs route.

Setup

Create a local directory to store data posted to the server's host's file system.

mkdir "$HOME/data_server/results" -p

Local Development

Requirements

• minimum 600 MiB RAM
• Python ^3.7.1 (translating caret requirements)
• poetry
• PostgreSQL service
• Current working directory is project root

poetry install --dev

Environment Variable Configuration

Define a .env with these environment variables.

---

Data Server Configuration

• DATA_SERVER_LOG_LVL

  Default: info

  Data server log level. Current Uvicorn server options: 'critical', 'error', 'warning', 'info', 'debug', 'trace'.

• DATA_SERVER_PORT

  Default: 7070

  This application's service port.

• DATA_SERVER_PUBLIC_HOST

  Default: localhost

  URL to public host of data server.

• DATA_SERVER_SECRET

  Secret to encode passwords in database.

---

Postgres Service Configuration

• POSTGRES_PORT

  Docker: Empty string (i.e. POSTGRES_PORT=)

  Podman: 5432

  Postgres service port.

• POSTGRES_SERVER

  Docker: name of the database network alias, see docker-compose.sh.

  Podman: localhost

  Domain name, or IP address, hosting the postgres service.

• POSTGRES_PASSWORD

  Postgresql database super user password.

---

Initial User Configuration

• FIRST_SUPERUSER

  Username for the first super user, in the format of an e-mail.

• FIRST_SUPERUSER_PASSWORD

  Password for the first super user.

---

Pruner Configuration

ENABLE_PRUNER

Var to enable or disable pruner

Configure the pruner_config.toml configuration file to add folders on which the pruner should be active. In the toml file you also need to set the days_to_live parameter, which will determine the number of days after which the files/folders in the path are pruned.

The pruner and server logs can be viewed by running the command podman logs snap_web on the machine/instance on which snappy server is deployed.

Steps to create and run a new container image of snappy server

Make your changes in the snappy-data-server directory.

This will create a local podman image name snappy

podman build --tag snappy -f Dockerfile

In the same directory as pod-compose.sh, create a .env configuration file.

Full example

POSTGRES_PASSWORD=barfoo
DATA_SERVER_SECRET=secretfoobar
DATA_SERVER_LOG_LVL=info
DATA_SERVER_PUBLIC_HOST=localhost
DATA_SERVER_PORT=7070
FIRST_SUPERUSER=foo@bar.com
FIRST_SUPERUSER_PASSWORD=foobar

Start snappy server pod with the image you built locally.

./pod-compose.sh localhost/snappy

Containerized Application Usage

Overview

1. Configure .env for PostgreSQL service and Snappy server application.

2. Create a means of communication between the containers.

   • Podman containers can easily communicate with localhost when they're in the same pod.
   • Docker containers can easily communicate over the default bridge network when each service is given a network-alias.

3. Start a PostgreSQL database service.

4. Add usernames and passwords.

   • Assign username and password to the FIRST_SUPERUSER environment variables. Usernames must be a valid email regular expression, but do not have to be capable of receiving email.
   • Run the application's inital_data.py to seed the database with user accounts.

5. Execute the deployment's start.sh script to get the server running.

Notes

• The bind mount volume is labeled as a shared directory with the host, :z, to persist results and facilitate integration with other tools on the host.

Next step

Using the data server with the data server CLI