Troubleshooting#

Frequently Asked Questions#

Q: What file formats are supported?#

A: PDF, docx and TXT (plain-text) files are supported.

Q: How large can my document be?#

A: The maximum file size may vary depending on the summarization model being used.

Q: Can I use my own summarization model?#

A: Yes, update the model files and configuration in the OpenVINO™ model server service.

Q: Where are summaries stored?#

A: Summaries are not stored by default; they are returned in the response.

Q: How do I update environment variables?#

A: Edit the .env file in the project root and restart the services.

Q: How do I configure the chunk size for document processing?#

A: The chunk size determines how large documents are split for processing. You can configure it by setting the CHUNK_SIZE environment variable in the .env file located in app/.env.

Default value: 4096 characters per chunk

Recommended values based on file size:

  • Small documents (<100 KB): 1024-2048 characters

  • Medium documents (100-200 KB): 4096-6144 characters

  • Large documents (>200 KB): 8192-16384 characters

Note: These values are suggestive guidelines. Smaller chunk sizes create more chunks, which increases processing time and may cause timeouts for large files. Larger chunk sizes reduce the number of chunks but require more memory per chunk. Experiment with different values to find the optimal setting for your specific use case, file types, and system resources.

Q: What is the ideal time for services or pods to become ready when deployed via Helm?#

A: The typical initialization time for services and pods deployed using Helm is approximately 6 to 8 minutes, depending on the system resources and network conditions.

Q: What should be done if the OVMS service deployed via Helm remains stuck in the init phase for more than 10 minutes?#

A: If the OVMS pod does not proceed beyond the init phase within 10 minutes, consider the following debugging steps:

  • Check Init Script Logs:

    Identify the exact command where the init container is stuck by inspecting its logs:

    kubectl logs -n <your-namespace> <pod-name> -c init-script

  • Clean Up Persistent Volume Claims (PVCs):

    When Values.global.keeppvc in values.yaml is set to false, the PVC is expected to be deleted automatically when the Helm release is uninstalled. However, in some cases — especially when a deployment gets stuck or fails — the PVC might not be removed. Stale PVCs can cause issues during future deployments.

    Follow the steps below to clean up PVCs:

    • Uninstall the Helm release to allow Helm to delete any associated PVCs:

      helm uninstall <release-name> -n <your-namespace>

    • Then delete any leftover PVCs (if not auto-deleted):

      kubectl get pvc -n <your-namespace>

      kubectl delete pvc <pvc-name> -n <your-namespace>

    After PVC cleanup, try redeploying the service.

  • Verify Network Stability:

    Ensure there are no underlying hardware issues such as a faulty network cable or unstable network connectivity, which might cause timeouts during initialization.