Step 3 - Node Configuration
- 26 Jan 2024
- 1 Minute to read
- Print
- DarkLight
- PDF
Step 3 - Node Configuration
- Updated on 26 Jan 2024
- 1 Minute to read
- Print
- DarkLight
- PDF
Article summary
Did you find this summary helpful?
Thank you for your feedback
Update
26/01/2024: Document initialization (Ops)
Procedure for Configuring Node Distribution for Embedding Feature in Paradigm
Correct node configuration is essential for deploying the embedding feature of Paradigm, such as ChromaDB integration. Follow these instructions to configure your nodes.
Configuration for Single Node Deployment
- If you plan to use the same machine for all features, check the Embedding Single Node Deployment box.
- This will allow for quicker setup by applying the same node name to all components of the embedding feature, including ChromaDB.
Configuration for Multi-Node Deployment
- If you wish to deploy on multiple nodes, proceed as follows:
For Embedding Feature:
- Node Name for Embedding (Integration): Enter the name of your integration node. The Kubernetes node name should correspond to the
kubernetes.io/hostname
attribute.
For ChromaDB:
- Node Name for ChromaDB: Provide the name of your ChromaDB node following the same format.
Important Tips and Remarks
- The assignment of node names should follow Kubernetes conventions to ensure correct matching and efficient pod distribution.
- The configuration determines pod affinity, influencing how they are distributed across nodes.
Saving and Applying Settings
- After setting the node names, confirm that the information is accurate.
- Save the settings to establish the node distribution for the embedding features of your Paradigm deployment.
Final Note
The decision to use a single node or multiple nodes should be based on the expected workload and available resources. Flexibility and performance can be affected by how nodes are configured.
Was this article helpful?