opea-project
Added PersistenceVolume and used it in chatqna.yaml
Description
Currently the models data from Hugging Face are downloaded to local shared directory under /mnt/opea-models. However, this is not correct approach in production use, because the models will be downloaded to each kubernetes node separately, which will increase the amount of data required to be downloaded and increase the application startup time.
This change is replacing a local volume mount with Kubernetes Persistent Volume. The mount will still be from the local host path, however it enables the implementor to easily replace it with shared storages available across nodes, like NFS.
Similar change might be added to other example applications.
Issues
n/a
Type of change
List the type of change like below. Please delete options that are not relevant.
- [ ] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds new functionality)
- [ ] Breaking change (fix or feature that would break existing design and interface)
- [x] Others (enhancement, documentation, validation, etc.)
Dependencies
n/a
Tests
Run ChatQnA example use case and confirmed the models are downloaded to local path on the host.
