Ingestor Server Volume Mounting for NVIDIA RAG Blueprint#

You can mount a host directory to access extraction results from NeMo Retriever Library directly from the filesystem when you use the NVIDIA RAG Blueprint. Designed for advanced developers who need programmatic access to raw extraction results for custom processing pipelines or external vector database integration.

Configuration#

Environment Variables#

Variable	Default	Description
`INGESTOR_SERVER_DATA_DIR`	`/data/`	Container internal path. Mapped to the `rag-vol-ingestor` Docker named volume.
`APP_NVINGEST_SAVETODISK`	`False`	Enable disk persistence

Setup#

Enable disk persistence:
```
# Enable disk persistence
export APP_NVINGEST_SAVETODISK=True

# (Optional) Override container internal path
export INGESTOR_SERVER_DATA_DIR=/data/
```
The ingestor-server compose file already mounts rag-vol-ingestor at INGESTOR_SERVER_DATA_DIR; nothing else needs to be configured to persist results.

Accessing the Results from the Host#

Docker named volumes are owned by root on the host, so use one of the following patterns to read the files:

# Copy a single result file out of the volume:
docker run --rm -v rag-vol-ingestor:/src:ro -v "$PWD":/dst alpine \
  cp /src/nv-ingest-results/<collection>/<file>.results.jsonl /dst/

# List the directory tree inside the volume:
docker run --rm -v rag-vol-ingestor:/src:ro alpine ls -la /src/nv-ingest-results

# Or copy directly from the running ingestor-server container:
docker cp ingestor-server:/data/nv-ingest-results ./nv-ingest-results

See Manage Persistent Data Volumes for backup, reset, and migration commands.

Result Structure#

Results are saved as .jsonl files with naming convention: {original_filename}.results.jsonl

rag-vol-ingestor:/
└── nv-ingest-results/
    ├── collection_name1/
    │   ├── document1.pdf.results.jsonl
    │   ├── presentation.pptx.results.jsonl
    │   └── spreadsheet.xlsx.results.jsonl
    └── collection_name2/
        ├── report.pdf.results.jsonl
        ├── analysis.docx.results.jsonl
        └── data.xlsx.results.jsonl

Each .jsonl file contains structured extraction metadata including text segments, document structure, images, tables, and chunk boundaries.

Advanced Usage: These .jsonl files can be used for storing data in vector databases or performing custom processing workflows as desired. This functionality is intended for advanced developers who need direct access to the structured extraction results.

Note

This is an advanced feature for custom processing workflows. Standard RAG functionality stores results directly in the vector database.

Helm (Kubernetes)#

Overview#

The Helm chart supports persisting ingestor-server data to a PersistentVolumeClaim (PVC). When enabled, the chart mounts a PVC at the same path used by INGESTOR_SERVER_DATA_DIR (default /data/). Set APP_NVINGEST_SAVETODISK=True to write extraction results to disk.

Values#

Edit values.yaml and set:

ingestor-server:
  envVars:
    # Ensure results are written to disk inside the pod
    APP_NVINGEST_SAVETODISK: "True"
    # Directory inside the container where results will be written
    INGESTOR_SERVER_DATA_DIR: "/data/"

  # PVC configuration (created automatically unless existingClaim is set)
  persistence:
    enabled: true
    existingClaim: ""         # set to use an existing PVC; leave empty to create one
    storageClass: ""          # set if your cluster requires a specific class (e.g., "standard")
    accessModes:
      - ReadWriteOnce
    size: 50Gi
    # Optional: explicitly set the mount path (defaults to INGESTOR_SERVER_DATA_DIR)
    mountPath: "/data/"
    # Optional: mount a subPath within the PVC
    subPath: ""

Notes:

If existingClaim is empty, the chart will create a PVC named <appName>-data. With the default appName of ingestor-server, the PVC name will be ingestor-server-data.
The container writes results under /data/ by default. Structure matches the compose example: /data/nv-ingest-results/<collection>/file.results.jsonl.

Deploy the Changes#

After modifying values.yaml, apply the changes as described in Change a Deployment.

For detailed HELM deployment instructions, see Helm Deployment Guide.

List and Access Files#

List results inside the ingestor-server pod (default mount path /data/):

kubectl -n rag exec -it <ingestor-pod> -- ls -l /data/

Copy data from the pod to your local computer:

kubectl -n rag cp <ingestor-pod>:/data/ ./ingestor-data