Dify Database Plugin Installation Failed: Continuous Retry + Clear Button Unresponsive – Need Help!

[henrybjsoft]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:)
• Please do not modify this template :) and fill in all the required fields.

Content

I'm encountering an issue while using Dify. I think I might have clicked to install a database plugin once, and the installation probably failed. Since then, the system has been continuously attempting to install it but failing every time—there are numerous errors visible in the plugin installation log box. Moreover, clicking the "Clear" button doesn't trigger any response.

Could someone help me take a look? Thanks so much!

[dosubot[bot]]

Your issue is caused by stuck plugin installation tasks in the backend, which block new installs and make the "Clear" button unresponsive. This usually happens after a timeout or dependency install failure in Docker, and the UI can't clear the error state without backend/API intervention.

How to clear all stuck plugin installation tasks:

1. Get your plugin daemon API URL and key:

   • The default API endpoint is usually http://plugin_daemon:5002 (or host.docker.internal:5002 if running from your host).
   • Get the API key by running this inside your plugin_daemon container:

     docker exec -it <plugin_daemon_container_name> printenv PLUGIN_DAEMON_KEY
     

2. Run this command (replace placeholders):

   curl -X POST "http://plugin_daemon:5002/plugin/1/management/install/tasks/delete_all" \
     -H "X-Api-Key: <your PLUGIN_DAEMON_KEY>"

   • In the open-source version, tenant_id is always 1.
   • If you get a 401 error, double-check the key and that you're using the plugin_daemon's key, not the backend's.

3. After clearing, refresh the UI and retry plugin installation.

Why the "Clear" button is unresponsive:
The button calls a backend API to delete install tasks, but if the backend task state is stuck or the API call fails, nothing happens in the UI. Only the backend API can clear these stuck tasks in this situation. details

How to prevent future install timeouts:

• Increase these environment variables in your docker-compose.yaml or .env:

  PYTHON_ENV_INIT_TIMEOUT=320
  UV_HTTP_TIMEOUT=120s
  PLUGIN_DAEMON_TIMEOUT=1200
  

• Use a fast PyPI mirror (especially if in China):

  PIP_MIRROR_URL=https://pypi.tuna.tsinghua.edu.cn/simple
  

• Make sure Docker containers have enough CPU/memory and can access the internet (port 443 open). details

Manual workaround for dependency install issues:
If dependency installation is the bottleneck, you can enter the plugin_daemon container, find the plugin directory, and run:

pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple

But this does not clear stuck install tasks—use the API above for that.

If you need more help, check the plugin_daemon logs (docker logs <plugin_daemon_container_name>) for details on why installs are failing.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[henrybjsoft]

Thank you for your reply! I tried following the steps you provided. The first step worked, and I was able to get the Key, but the second step doesn't seem to work.

I installed via Docker. You mentioned the endpoint http://plugin_daemon:5002
, but I can't access this endpoint from the host machine. I tried using docker exec to enter the plugin_daemon container—there does seem to be this service endpoint inside—but when I ran the curl -X POSTcommand, it returned a 404 error.

Could you please help me check what the issue might be? Thanks.

[dosubot[bot]]

The correct API endpoint for clearing all plugin installation tasks is:

POST /plugin/1/management/install/tasks/delete_all

You must include the header:

X-Api-Key: <your PLUGIN_DAEMON_KEY>

A 404 error usually means the path is wrong, the key is missing/incorrect, or the plugin_daemon service isn't running or accessible. Here’s a checklist to troubleshoot:

1. Double-check the endpoint path (it should be exactly as above, with tenant_id=1).
2. Make sure you’re using the PLUGIN_DAEMON_KEY from the plugin_daemon container, not the backend.
3. Confirm you’re running the curl command inside the plugin_daemon container.
4. Ensure both the backend and plugin_daemon containers are on the same Docker network and the plugin_daemon service is running.
5. If you still get a 404, try restarting the plugin_daemon service—this often clears stuck tasks if the API call fails source.

If you’re comfortable with SQL and as a last resort, you can clear the install_tasks table directly in the dify_plugin database (not recommended unless you know your setup supports it):

docker exec -it <db_container_name> bash
psql -U postgres -d dify_plugin
delete from install_tasks;
\q

But usually, restarting the plugin_daemon service is safer and effective.

Let me know if you need help with any of these steps!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}