Skip to content
Nexa Platform Docs

S3 & UC Volumes

A file-system connection (target system File System, labelled Storage in the UI) points Nexa at object storage so a connector can publish files there. Configure it under Settings > Connection > Storage. Unlike database and message-queue connections, a Storage connection has no secret — access is governed by the data platform’s grants, so there are no cloud access keys to manage.

Supported storage types

Section titled “Supported storage types”

S3, ADLS (Azure Data Lake Storage), and GCS (Google Cloud Storage).

Path configuration

Section titled “Path configuration”

On Databricks, object storage is reached through Unity Catalog volumes rather than raw bucket URLs. You provide the volume path and grant the connector’s service principal access — no AWS/Azure/GCP keys are stored in Nexa.

Field Required Example
Volume Path Yes /Volumes/<catalog>/<schema>/<volume_name>

Grant the service principal that runs the connector READ_VOLUME and WRITE_VOLUME on the volume:

GRANT READ_VOLUME, WRITE_VOLUME
  ON VOLUME <catalog>.<schema>.<volume>
  TO `service-principal-name`;

Because the underlying S3/ADLS/GCS credentials are held by the Unity Catalog external location, the connection itself carries no secret.

Create a Storage connection

Section titled “Create a Storage connection”

  1. Go to Settings > Connection > Storage and click Add.

  2. Enter a Connection Name and choose the storage type (S3, ADLS, or GCS).

  3. Enter the Volume Path — the full path to your governed volume/stage, for example /Volumes/catalog/schema/volume_name.

  4. Confirm the connector’s service principal or role has read/write access to that location (see the grant above).

  5. Click Submit to save. The connection is marked Active.

Next step

Section titled “Next step”