CLI: Running jobs

To run Graphile Worker, simply run the graphile-worker binary passing your PostgreSQL connection string via the -c entry. Worker manages its own database schema (graphile_worker) into which jobs are stored; when you start Graphile Worker it will automatically create or update this schema if necessary.

npx graphile-worker -c "postgres:///my_db"

tip

npx looks for the graphile-worker binary locally; it's often better to use the "scripts" entry in package.json instead.

info

Graphile Worker expects the Postgres role used at runtime to be the same as the role used while running the migrations. If you need to run your migrations as a different role, one solution is to explicitly change the owner of the graphile_worker.* tables to be the same role as is used at runtime.

CLI options

The following CLI options are available:

Options:
      --help                    Show help                              [boolean]
      --version                 Show version number                    [boolean]
  -c, --connection              Database connection string, defaults to the
                                'DATABASE_URL' envvar                   [string]
  -s, --schema                  The database schema in which Graphile Worker is
                                (to be) located
                                           [string] [default: "graphile_worker"]
      --schema-only             Just install (or update) the database schema,
                                then exit             [boolean] [default: false]
      --once                    Run until there are no runnable jobs left, then
                                exit                  [boolean] [default: false]
      --crontab                 override path to crontab file           [string]
  -j, --jobs                    number of jobs to run concurrently
                                                           [number] [default: 1]
  -m, --max-pool-size           maximum size of the PostgreSQL pool
                                                          [number] [default: 10]
      --poll-interval           how long to wait between polling for jobs in
                                milliseconds (for jobs scheduled in the
                                future/retries)         [number] [default: 2000]
      --no-prepared-statements  set this flag if you want to disable prepared
                                statements, e.g. for compatibility with
                                pgBouncer             [boolean] [default: false]

Connection string

A PostgreSQL connection string looks like this:

postgres://[user]:[pass]@[host]:[port]/[databaseName]?[parameter]=[value]

Where each of the [...] placeholders are optional. Here are some examples:

• postgres:///my_db — connect to database my_db on the default host (localhost) and default port (5432)
• postgres://127.0.0.1/my_db
• postgres://127.0.0.1:5432/my_db
• postgres://postgres:[email protected]:5432/my_db
• postgres://postgres:[email protected]:5432/my_db?ssl=1