You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Because I'm running indexing as the docker container starts it times out.
It seems like if the client does not get an immediate connection it will timeout. So I assume the timeout value is only respected once it is connected?
@Zyles Maybe. On Cline's side the timeout value is passed per the MCP SDK's expectations. Anything beyond that is on the side of either the SDK or the individual MCP server to implement.
I would suggest running this test with the sleep mcp available in the marketplace (or any other existing MCP server) and see if you get the same problem.
#!/usr/bin/env node
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
import { z } from "zod";
import { promises as fs } from 'fs';
// Create an MCP server
const server = new McpServer({
name: "MCP Server",
version: "1.0.0"
});
// Add an addition tool
server.tool("add",
{ a: z.number(), b: z.number() },
async ({ a, b }) => ({
content: [{ type: "text", text: String(a + b) }]
})
);
// Start server, doesn't work:
async function runServer() {
const transport = new StdioServerTransport();
await new Promise(f => setTimeout(f, 30000));
await server.connect(transport);
console.log("MCP Server running on stdio");
}
// Start server, works but MCP is not ready
async function runServer() {
const transport = new StdioServerTransport();
await server.connect(transport);
await new Promise(f => setTimeout(f, 30000));
console.log("MCP Server running on stdio");
}
runServer().catch((error) => {
console.error("Fatal error running server:", error);
process.exit(1);
});
Thanks for reporting this timeout issue! This has been consolidated into the comprehensive MCP system issues tracking issue #4391. The MCP timeout configuration problems you identified are now part of the broader MCP server startup and runtime improvements being tracked there.
Activity
canvrno commentedon Mar 18, 2025
@Zyles what MCP server were you using when this occurred?
Zyles commentedon Mar 18, 2025
I made my own with the Typescript MCP SDK.
Because I'm running indexing as the docker container starts it times out.
It seems like if the client does not get an immediate connection it will timeout. So I assume the timeout value is only respected once it is connected?
celestial-vault commentedon Mar 18, 2025
@Zyles Maybe. On Cline's side the timeout value is passed per the MCP SDK's expectations. Anything beyond that is on the side of either the SDK or the individual MCP server to implement.
I would suggest running this test with the sleep mcp available in the marketplace (or any other existing MCP server) and see if you get the same problem.
celestial-vault commentedon Mar 18, 2025
If you can come up with steps to reproduce on a publicly available MCP, not one that you've privately built, then we can take a look into it.
Zyles commentedon Mar 20, 2025
Here is an example:
celestial-vault commentedon Mar 20, 2025
@Zyles Can you verify that this timeout setting does not work with existing MCP servers on the marketplace?
Zyles commentedon Mar 21, 2025
Like which one?
celestial-vault commentedon Jun 23, 2025
Thanks for reporting this timeout issue! This has been consolidated into the comprehensive MCP system issues tracking issue #4391. The MCP timeout configuration problems you identified are now part of the broader MCP server startup and runtime improvements being tracked there.