Iceberg

Provides read-only access to Iceberg tables via Impala for schema inspection and read-only queries.

python

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "cloudera-iceberg-mcp-server": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/cloudera/iceberg-mcp-server@main",
        "run-server"
      ],
      "env": {
        "IMPALA_HOST": "coordinator-default-impala.example.com",
        "IMPALA_PORT": "443",
        "IMPALA_USER": "username",
        "IMPALA_DATABASE": "default",
        "IMPALA_PASSWORD": "password"
      }
    }
  }
}

You can access Cloudera Iceberg data through a read-only MCP server that interfaces with Apache Impala. This server lets you inspect Iceberg schemas and run read-only SQL queries, enabling tools like LLMs to understand data structures and extract results safely.

How to use

Use the MCP server with your MCP client to read Iceberg schemas and execute read-only queries against Impala. You have two official stdio-based configurations you can run locally or from a remote source. Choose the one that fits your deployment needs, then point your MCP client to the corresponding server entry.

How to install

Prerequisites: you need a runtime environment capable of executing MCP servers (Node.js or Python runtime as used by the examples), and access to an Impala coordinator. The server configurations below describe how to start the MCP server in stdio mode, either by pulling the server directly from GitHub or by running it from a local clone.

Configuration and transport

The MCP transport can be configured with the MCP_TRANSPORT environment variable. Supported values are stdio (default), http, and SSE. For local, command-line driven usage, stdio is recommended as it integrates smoothly with tools that communicate via standard input and output.

Tools and endpoints

This MCP server exposes two read-only capabilities that you can invoke from your MCP client: 1) execute_query(query: string) — Run a SQL query on Impala and return results as JSON. 2) get_schema() — List all tables in the current database.

Examples of usage patterns

Use the stdio configurations below to start the MCP server and then connect your MCP client to the running process. The client can perform read-only queries and inspect schemas to assist data understanding and query generation.

Security considerations

Since this server provides read-only access to data, ensure your Impala credentials are stored securely and that access to the MCP client is restricted to trusted processes. Use strong passwords for Impala and prefer network isolation or encryption for transport if you expose the server beyond a local machine.

Available tools

execute_query

Executes a SQL query against Impala and returns the results as JSON.

get_schema

Retrieves the list of tables in the current Impala database.

Built by

VeilStrat

AI signals for GTM teams