CockroachDB
A server for direct interaction with CockroachDB databases.
CockroachDB MCP Server
This is a CockroachDB MCP server for Cursor, implemented based on the Model Context Protocol (MCP) specification, allowing you to interact directly with CockroachDB database in Cursor.
Features
- Connect to CockroachDB database
- Get all tables from the database
- Get table structure information
- Execute SQL queries
- Provide database status resources
- Provide SQL query templates
- Automatic reconnection mechanism to ensure connection stability
- Connection keep-alive mechanism to prevent connection timeout
- Graceful process exit handling
- Detailed logging for troubleshooting
- Support manual disconnection
Installation
Installing via Smithery
To install CockroachDB MCP Server for Claude Desktop automatically via Smithery:
npx -y @smithery/cli install @Swayingleaves/cockroachdb-mcp-server --client claude
- Clone the repository and enter the project directory
- Install dependencies:
pip install -r requirements.txt
- Install uv
curl -LsSf https://astral.sh/uv/install.sh | sh
Using in Cursor
{
"mcpServers": {
"cockroachdb-mcp": {
"command": "uv",
"args": [
"--directory",
"/Users/local/cockroachdb-mcp",
"run",
"server.py"
],
"jdbc_url": "jdbc:postgresql://localhost:26257/defaultdb",
"username": "root",
"password": "root"
}
}
}
MCP Function Description
Tools
connect_database
Connect to CockroachDB database.
Parameters:
jdbc_url: JDBC connection URL (e.g., jdbc:postgresql://localhost:26257/defaultdb)username: Database usernamepassword: Database password
initialize_connection
Initialize database connection, can be called immediately after connecting to the MCP server to establish a database connection.
Parameters:
jdbc_url: JDBC connection URL (e.g., jdbc:postgresql://localhost:26257/defaultdb)username: Database usernamepassword: Database password
disconnect_database
Manually disconnect from the database.
No parameters.
get_tables
Get all tables from the database.
No parameters.
get_table_schema
Get structure information of a specified table.
Parameters:
table_name: Table name
execute_query
Execute SQL query.
Parameters:
query: SQL query statement
Resources
db://status
Get database connection status.
Returns:
- When not connected:
"Not connected" - When connected:
"Connected - [Database version]" - When connection error:
"Connection error - [Error message]"
Prompts
sql_query_template
SQL query template to help users write SQL queries.
Logs
Server logs are saved in logs/cockroachdb_mcp.log file. You can check this file to understand the server's running status and detailed logs.
The log file uses a rotating log mechanism, with each log file maximum size of 10MB and keeping up to 5 backup files to prevent excessive disk space usage.
Special Character Handling
This server uses psycopg2 to connect directly to CockroachDB database, which automatically handles special characters in usernames and passwords without additional URL encoding. This ensures correct database connection even when passwords contain special characters (such as @, %, &, etc.).
TCP Keep-alive Settings
The server is configured with TCP keep-alive mechanism by default to prevent connections from being closed due to long periods of inactivity:
keepalives=1: Enable TCP keepalivekeepalives_idle=30: Send keepalive after 30 seconds of idle timekeepalives_interval=10: Send keepalive every 10 secondskeepalives_count=5: Give up after 5 attempts
Troubleshooting
If you encounter problems, please check the log file logs/cockroachdb_mcp.log, which will help you understand the server's running status and potential issues.
Common Issues
- Connection Refused: Ensure the CockroachDB server is running and accessible from your machine.
- Authentication Failed: Check if the username and password are correct.
- Connection Timeout: Check if the network connection is stable, especially when connecting to a remote database.
- Database Server Issues: Check if the CockroachDB server is running properly.
Похожие серверы
Federal Reserve Economic Data
Access financial datasets from the Federal Reserve Economic Data (FRED) API.
Amazon Neptune
Query Amazon Neptune databases using openCypher, Gremlin, and SPARQL. Supports both Neptune Database and Neptune Analytics.
CData CSV Files
A read-only MCP server for CSV files from CData, requiring an external JDBC driver for connection.
Supabase Memory Service
A memory service using Supabase PostgreSQL with pgvector for semantic search and knowledge graph storage.
KuzuMem-MCP
A distributed memory bank MCP tool that stores memories in a KùzuDB graph database, with repository and branch filtering capabilities.
InfluxDB
A server for integrating with InfluxDB, a time-series database.
Redshift MCP Server
An MCP server for Amazon Redshift, allowing AI assistants to interact with Redshift databases.
Opera Omnia
Access a rich collection of JSON datasets for games, storytelling, and bot development from the Opera Omnia project.
Gralio SaaS Database
Access a comprehensive database of over 30,000 SaaS products, including reviews, pricing, alternatives, and growth metrics.
MCP Persistence
MCP Persistence: your AI Agent now creates and manages databases on its own