Atlassian-MCP Connection Fails

Symptom

The Atlassian-MCP connection fails and the following error message is displayed:

MCP error -32000: Connection closed Server stderr: (node:36788) Warning: Setting the NODE_TLS_REJECT_UNAUTHORIZED environment variable to '0' makes TLS connections and HTTPS requests insecure by disabling certificate verification. (Use `node --trace-warnings ...` to show where the warning was created) [9860] Using existing client port: 39570 [9860] Discovering OAuth server configuration... (node:9860) Warning: Setting the NODE_TLS_REJECT_UNAUTHORIZED environment variable to '0' makes TLS connections and HTTPS requests insecure by disabling certificate verification. (Use `node --trace-warnings ...` to show where the warning was created) [9860] Discovered authorization server: https://auth.atlassian.com/VCeDsk8ZHncYF1g234fKtc4lNipbBhu3 [9860] [9860] Connecting to remote server: https://mcp.atlassian.com/v1/mcp/authv2 [9860] Using transport strategy: http-first [9860] Please authorize this client by visiting: https://auth.atlassian.com/authorize?response_type=code&client_id=wfiikAOsPzJNcW94j9xZQCxyDRF5yqOt&code_challenge=9xv1nceaTjD0iE_MCCJiZSujUXDCE090V91EKlVP-C8&code_challenge_method=S256&redirect_uri=http%3A%2F%2Flocalhost%3A39570%2Foauth%2Fcallback&state=5a86d09f-43f3-48de-9ad3-ec915094f6c1&scope=read%3Ame+read%3Aaccount+offline_access+email+read%3Ajira-work+write%3Ajira-work+search%3Aconfluence+read%3Aconfluence-user+read%3Apage%3Aconfluence+write%3Apage%3Aconfluence+read%3Acomment%3Aconfluence+write%3Acomment%3Aconfluence+read%3Aspace%3Aconfluence+read%3Ahierarchical-content%3Aconfluence+write%3Acomponent%3Acompass+read%3Acomponent%3Acompass+read%3Ascorecard%3Acompass+write%3Ascorecard%3Acompass+read%3Aevent%3Acompass+read%3Ametric%3Acompass+read%3Aall%3Atwg+write%3Aall%3Atwg&prompt=consent&resource=https%3A%2F%2Fmcp.atlassian.com%2Fv1%2Fmcp%2Fauthv2 [9860] Browser opened automatically. [9860] Authentication required. Initializing auth... [9860] Initializing auth coordination on-demand node:events:486 throw er; // Unhandled 'error' event ^ Error: listen EADDRINUSE: address already in use 127.0.0.1:39570 at Server.setupListenHandle [as _listen2] (node:net:1948:16) at listenInCluster (node:net:2005:12) at node:net:2214:7 at process.processTicksAndRejections (node:internal/process/task_queues:90:21) Emitted 'error' event on Server instance at: at emitErrorNT (node:net:1984:8) at process.processTicksAndRejections (node:internal/process/task_queues:90:21) { code: 'EADDRINUSE', errno: -4091, syscall: 'listen', address: '127.0.0.1', port: 39570 } Node.js v24.14.1

Cause Analysis

Your Atlassian-MCP connection may fail due to any of the following reasons:

• The official remote MCP access address of Atlassian is https://mcp.atlassian.com/v1/mcp/authv2, which uses OAuth for web login authorization. You may not have confirmed the authorization on the displayed web page in a timely manner.
• You may have logged in to another Atlassian account before, so your server has already authorized that account.

Solution

1. Close the mcp_settings.json file.
2. Run the following command to delete the local OAuth authorization cache of mcp-remote:
   • Windows (PowerShell administrator or common terminal)

     Remove-Item -Recurse -Force "$env:USERPROFILE\.mcp-auth"

   • Mac/Linux

     rm -rf ~/.mcp-auth

3. You can use either of the following methods to prepare the browser environment:
   • (Recommended) Use the incognito or private browsing mode of the browser to access the authorization page to prevent the browser cookies from automatically logging in to other accounts.
   • Manually access the Atlassian unified login page. If you are logged in to another account, log out. Open the Jira unified account login page using a regular browser.

4. On the Installed tab, click in the row where Atlassian-MCP is located. If the MCP service status is displayed in green, the service is enabled successfully.

   If the issue persists, submit a service ticket to get help from Huawei Cloud customer service.