Run scripts on Deephaven Enterprise workers
The run_local_script command runs a local Groovy or Python script on a remote Deephaven Enterprise worker. The tool sends the script body to the worker for evaluation. The tool creates a temporary Persistent Query for the worker and closes it when the script completes.
Usage
Authentication
Exactly one authentication flag is required:
| Flag | Description |
|---|---|
-u / --user <username> | Authenticate with a username. Prompts for a password. |
-k / --key-file <path> | Authenticate with a private key file. |
-a / --saml | Authenticate using SAML. |
Options
| Flag | Description | Default |
|---|---|---|
-s / --script <path> | Path to the local script to run. Required. | — |
-c / --connection-json <url> | URL to the server's connection.json file. | Local server configuration |
-q / --query-server <name> | Name of the target query server. | Controller selection |
-m / --heap-size <GB> | JVM heap size in GB for the worker. | 4 |
-e / --worker-kind <kind> | Worker engine type. Use DeephavenCommunity for Core+ workers, or a custom WorkerKind configured by your system administrator. | DeephavenCommunity |
-j / --jvm-arg <arg> | Additional JVM argument for the worker. May be repeated. | — |
-t / --timeout <seconds> | Worker creation timeout in seconds. | 60 |
-l / --log-file <path> | File to write log messages from the worker. | stdout |
Script language
The script language is inferred from the file extension. Files ending in .groovy run as Groovy; all other files run as Python.
Python environment
run_local_script requires the Core+ Python client. When the current Python environment does not include the client, the command creates a temporary virtual environment, installs the client wheel from /usr/illumon/coreplus/latest/py/wheel, and deletes the environment when the script completes.
To reuse one virtual environment across runs, set DEEPHAVEN_CLIENT_VENV_DIR to a directory that the user running the command can write to. The first run creates the virtual environment in that directory, and later runs reuse it instead of rebuilding it:
Examples
Run a Python script
Run a Groovy script on a specific query server
Connect to a remote server
Use --connection-json to specify a remote server:
Increase heap size and pass JVM arguments
Exit codes
The command exits with 0 on success and 1 if the script throws an exception or fails to execute.
Related documentation
- Merging data — Using
run_local_scriptto run merge scripts - Public and private keys