Set up a database on a server

Make sure you have docker setup and ready to go. Follow these instructions if necessary.

Next, clone this repository on the server of interest. cd into it and run

docker-compose pull
docker-compose up -d

Check if the containers are running:

docker ps

CONTAINER ID IMAGE                 ... CREATED STATUS PORTS          ...
097f58433a2b hasura/graphql-engine ... 1m ago  Up 1m  8080->8080/tcp ...
b0b1aac0508d postgres              ... 1m ago  Up 1m  5432/tcp       ...

Access the Hasura Console

To access the Hasura Console, you will first need to install the Hasura CLI following these instructions.

Next, clone this repository on a machine with access to a browser and a monitor. Add a file named config.yaml to the root of this repository:

version: 3
endpoint: http://your-endpoint:1200
metadata_directory: metadata
actions:
  kind: synchronous
  handler_webhook_baseurl: http://localhost:3000

Replace your-endpoint with the url for your server machine, e.g. http://something.eecs.umich.edu:1200. Note that port 1200 is hard-coded in docker-compose.yaml. If there are issues with that port, you can change it under graphql-engine:, under ports:.

Now (from inside the root directory of this repository), run

hasura console

If hasura indicates that there is a new version available for the CLI, feel free to update. This command should open the hasura console in your browser. Currently there is no data or metadata in your database so there is not much to see here yet.

Import Hasura setup

First, we run the migrations which load the correct schema into our postgres database:

hasura migrate apply --all-databases

To confirm that this worked, run hasura console, look on the DATA tab, and click on public on the left (represented with a small folder icon). This should open a page that show a handful of untracked tables, views, and functions.

In order to set up Hasura to use these postgres entities, we apply our metadata to the endpoint:

hasura metadata apply

You should now see a list of tables under public (DATA tab):

chart
parameter_choices
run
run_blob
run_log
sweep

You will also see a long list of GraphQL Queries under the "Explorer" when you click on the API tab.

Backup data

Even if you do not yet have data in your database, you will want to set up regular data backups. The command to dump the contents of your postres data base is:

docker exec hasura_postgres_1 pg_dump -U postgres -d postgres > [output-file]

This command makes a few assumptions which are true by default:

• Your Postgres container is named hasura_postgres_1 is the the container running Postgres. You can confirm this by running docker ps.
• Both your Postgres username and database name is postgres (true per the default docker-compose.yml)

One way to perform regular backups is to run this command periodically using a cron job. On your server (not on localhost), run

crontab -e

Edit the file that comes up so that it looks like this:

  # m h  dom mon dow   command
  0 *  *   *   *     /path/to/hasura/directory/backup.sh /path/to/backup

Replacing /path/to/hasura/directory and /path/to/backup with the appropriate file paths. Your cron file may have other jobs running on it, and that is ok.

This will execute the backup.sh script every day and create a new file named X.sql where X is the day of the month. As a result, your data will overwrite itself every month, but you will have access to the version of your database that existed on each of the days of the past month.

Restore data

Run

docker exec -i hasura_postgres_1 psql -U postgres -d postgres < [backup-file]

This will load the actual data into the postgres database backing hasura. Finally run

hasura metadata apply