graphql-engine/frontend/docker

Hasura Frontend Docker Setup

The docker-compose.yml file in this directory is intended to create several local dbs, and relevant services for local Hasura frontend development. In addition, the aim for this will be to automatically populate db's with demo data as part of the initialization process.

There are a few basic steps:

1. Run docker compose up -d from this directory.
2. Enter the BigQuery service account private key for the BigQuery source. Property is, configuration.service_account.private_key in the default_metadata.json file. Alternatively, you can remove this section of the json file.
3. Go to Settings > Metadata Actions -- or just click here (you will need to enter your admin secret: moocow)
4. Select Import Metadata and select default_metadata.json

Note: By default, no tables are tracked. This is intentional as the Chinook database is being used for most of these and things can get a bit confusing if they are all tracked at once. I would recommend either using namespaces, or just track as needed.

BigQuery

If you would like to have a working BigQuery database first obtain the BigQuery service account key JSON data (if you don't know, then ask). Then, enter the private key in your default_metadata.json.

You may also need to flip the feature flags here to see everything in your local instance correctly.

MSSQL / SQL Server

In order to initialize and populate a MSSQL demo database automatically, we making use of the mssql-tools image. This image runs a few commands against the mssql image to create the demo data.

This scripting is done in docker/DataSources/mssql/run-initialization.sh.

After run-initialization.sh attempts to execute docker/DataSources/mssql/mssql-setup.sql, it creates the file docker/DataSources/mssql/REMOVE_ME_TO_RERUN_SETUP.txt. The presence of this .txt file will prevent the mssql-setup.sql script from executing again if the container is started again.

To force it to re-run, just delete /DataSources/mssql/REMOVE_ME_TO_RERUN_SETUP.txt and restart your container. Because /DataSources/mssql is mounted as a volume, it uses the local filesystem.

MongoDB

The Docker compose file starts a Mongo database with a single collection called mycollection. This collection is defined with a json schema file, since for the moment the Mongo agent can only track collections with a schema.

Database

The docker-compose file will start a Mongo database. You can connect to it with the following command:

mongosh "mongodb://host.docker.internal:27017"

For now the agent only works with databases without authentication, so keep that in mind if you're planning to use a different Mongo database than the one provided in docker-compose.yml

Agent

For the moment we don't have a docker image for the Mongo agent, but you can start it locally with the following steps:

• Install cabal (Haskell build tools)
• Go to the pro/dc-agents/mongodb directory
• cabal build
• cabal update
• cabal run
• Add an agent. Set URL to http://host.docker.internal:8888
• Add a Mongo database. Set URL to host.docker.internal:27017

Using Different HGE Images

You may need to replace the image: in the compose file with a different image. You can do this by downloading an image, and running docker load -i <the new image>. Once it's loaded copy the image name into the compose. Here's an example using a CI built image:

image: hasura/graphql-engine:dev-e3bc433-main.ubuntu.arm64