Vineyard Data Sharing Workflow
Workflow capability for sharing distributed in-memory objects across computation engines using Vineyard. Supports data scientists and ML engineers managing zero-copy data sharing, object lifecycle, and Kubernetes cluster operations for big data analytics pipelines.
What You Can Do
MCP Tools
connect-to-vineyard
Connect to a vineyard in-memory data manager server via IPC or TCP
store-object
Store a Python object in vineyard shared memory for zero-copy sharing
retrieve-object
Retrieve a distributed in-memory object by its ObjectID
inspect-object-metadata
Inspect metadata for a vineyard object including type, size, and cluster location
remove-object
Delete a vineyard object from shared memory
persist-object-globally
Make a vineyard object visible across all cluster instances
register-object-name
Associate a human-readable name with a vineyard ObjectID for discovery
lookup-object-by-name
Resolve a named vineyard object to its ObjectID
allocate-blob
Allocate a raw memory blob in vineyard for low-level data operations
retrieve-blob
Retrieve a raw memory blob from vineyard storage