MCP Server Cancels All Tasks When Client Sends notifications/cancelled — MCP error -32000 Connection Closed

What this error means

MCP error -32000: Connection closed is a Claude Code / MCP SDK failure pattern reported for developers trying to 开发者在 mcp 客户端/服务器交互中取消请求后，服务端整个接收循环崩溃，后续所有请求都报 connection closed，导致 claude code 或自定义 mcp 客户端完全不可用. Based on the imported evidence, treat this as a tool-specific troubleshooting page rather than a generic API error.

Why this happens

GitHub Issue #2610 on modelcontextprotocol/python-sdk (updated 2026-05-18). RequestResponder.exit leaks CancelledError killing stdio receive loop. Affects mcp 1.26.0 and 1.27.1 confirmed with FastMCP 3.1.x and Claude Code as host. This is a critical reliability issue for anyone using Claude Code or custom MCP integrations in production workflows. Exact root cause and fix identified. High commercial impact because MCP/AI coding tool failure directly blocks development.

Common causes

• GitHub Issue #2610 on modelcontextprotocol/python-sdk (updated 2026-05-18). RequestResponder.exit leaks CancelledError killing stdio receive loop. Affects mcp 1.26.0 and 1.27.1 confirmed with FastMCP 3.1.x and Claude Code as host. This is a critical reliability issue for anyone using Claude Code or custom MCP integrations in production workflows. Exact root cause and fix identified. High commercial impact because MCP/AI coding tool failure directly blocks development.

Quick fixes

1. Confirm the exact error signature matches MCP error -32000: Connection closed.
2. Check the Claude Code / MCP SDK account, local tool state, and provider configuration involved in the failing workflow.
3. Compare the failing environment with a known working setup, then change one configuration value at a time.

Platform/tool-specific checks

• Verify the command, editor, extension, or API client that produced the error.
• Compare local settings with CI, deployment, or editor-level settings when the error appears in only one environment.
• Avoid deleting credentials, local model data, or project settings until the failing scope is clear.

Step-by-step troubleshooting

1. Capture the exact error message and the command, editor action, or request that triggered it.
2. Check whether the failure is account/auth, quota/rate, model/provider, local runtime, or deployment configuration.
3. Review the source evidence below and compare it with your environment.
4. Apply one change at a time and rerun the smallest failing action.
5. Keep the working fix documented for the team or deployment environment.

How to prevent it

• Keep provider/tool configuration documented.
• Record non-secret diagnostics such as tool version, provider name, model name, and command path.
• Add a lightweight check before CI or production workflows depend on the tool.

Sources checked

• https://github.com/modelcontextprotocol/python-sdk/issues/2610

Evidence note: GitHub Issue #2610 on modelcontextprotocol/python-sdk (updated 2026-05-18). RequestResponder.exit leaks CancelledError killing stdio receive loop. Affects mcp 1.26.0 and 1.27.1 confirmed with FastMCP 3.1.x and Claude Code as host. This is a critical reliability issue for anyone using Claude Code or custom MCP integrations in production workflows. Exact root cause and fix identified. High commercial impact because MCP/AI coding tool failure directly blocks development.

Related errors

• AI Coding Tools

FAQ

What should I check first?

Start with the exact MCP error -32000: Connection closed text and the smallest action that reproduces it.

Can I ignore this error?

No. Treat it as a failed Claude Code / MCP SDK workflow until the root cause is understood.

Is this guaranteed to have one fix?

No. The imported evidence supports the troubleshooting path above, but tool behavior can vary by account, plan, version, provider, and local configuration.

How do I know the fix worked?

Rerun the same command, editor action, or request. The fix is working when that action completes without MCP error -32000: Connection closed.