Run an import job (UI)
| You can validate your specification file locally via the Local Dataflow Runner. The tool allows you to locally verify that your job file conforms to the template specification and execute it against a temporary Neo4j container (it relies on Apache Beam’s Direct Runner; there may still be runtime errors in Dataflow). This form of validation is much quicker than submitting a Dataflow job. |
Once the job specification is ready, upload its JSON file to your Cloud Storage bucket and go to Dataflow → Create job from template.
To create a job, specify the following fields:
-
Job name — A human-friendly name for the job.
-
Regional endpoint — Must match the region of the Google Cloud Storage bucket containing configuration and source files. If running the full example provided in docs, set the region to one of the
usones. -
Dataflow template — Select
Google Cloud to Neo4j. -
Path to job configuration file — The JSON job specification file (from a Google Cloud Storage bucket).
-
Optional Parameters > Options JSON — Values for variables used in the job specification file.
The connection metadata is specified either as a secret or a plain-text JSON resource. Exactly one of these options must be specified!
-
Optional Parameters > Path to the Neo4j connection metadata — The JSON connection information file (from a Google Cloud Storage bucket).
-
Optional Parameters > Secret ID for the Neo4j connection metadata — The ID of the secret containing the JSON connection information (from Google Secret Manager).
|
If you are using GCP Private Service Connect (PSC): To prevent driver routing and discovery errors (e.g., Even with PSC Global Access enabled, cross-region Bolt connections frequently drop during routing table refreshes. When launching your job, ensure you align your regional settings:
|
