Installation plugin error

[tianshiqingge]

Self Checks

• I have read the Contributing Guide and Language Policy.
• This is only for bug report, if you would like to ask a question, please head to Discussions.
• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report, otherwise it will be closed.
• 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：）
• Please do not modify this template :) and fill in all the required fields.

Dify version

1.13.0

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

Installation plugin error

failed to launch plugin: failed to install dependencies: failed to install dependencies: signal: killed, output: DEBUG uv 0.9.26 TRACE Checking lock for /root/.cache/uv at /root/.cache/uv/.lock DEBUG Acquired shared lock for /root/.cache/uv DEBUG Searching for default Python interpreter in virtual environments TRACE Querying interpreter executable at /app/storage/cwd/langgenius/tongyi-0.1.18@a214145a5793380db3c576f5460b50a9086cdf5a84089e1897f8c8196b20d816/.venv/bin/python3 DEBUG Found `cpython-3.12.3-... available=2097152 TRACE transition_after; stream=StreamId(3); state=HalfClosedLocal(Streaming); is_closed=false; pending_send_empty=true; buffered_send_data=0; num_recv=0; num_send=1 TRACE poll TRACE poll_complete TRACE schedule_pending_open TRACE flushing buffer TRACE release_capacity; size=2515 TRACE release_connection_capacity; size=2515, connection in_flight_data=2515 TRACE connection.state=Open TRACE poll TRACE poll_complete TRACE schedule_pending_open TRACE flushing buffer failed to init environment

✔️ Expected Behavior

Installing model plugins or other plugins mostly results in errors. Is this related to the plugin version? Installing the latest version or previous versions will result in errors. Is it related to the system? We are using the CentOS Stream 9 system. How can I solve this problem

❌ Actual Behavior

No response

[dosubot]

The signal: killed error during plugin installation is typically caused by two issues:

1. Timeout Issue - The Python environment initialization times out (default is 120 seconds). Increase the PYTHON_ENV_INIT_TIMEOUT in your docker-compose.yaml under the plugin_daemon service:

PYTHON_ENV_INIT_TIMEOUT: 320  # or higher, up to 1200 for severe cases

2. Memory/Resource Constraints - The "killed" signal often indicates the process was terminated by the OS due to insufficient memory. Check your Docker container's CPU and memory limits and increase them if needed.

Additional fixes to try:

• PyPI Mirror: If you're in China or a restricted network, ensure PIP_MIRROR_URL is set in your .env file (e.g., https://pypi.tuna.tsinghua.edu.cn/simple), then regenerate docker-compose with python3 generate_docker_compose

• Clear stuck tasks: If installations keep failing, you may need to clear all plugin installation tasks via the plugin daemon's internal API

After making changes, restart the plugin_daemon service and try installing the plugin again.

Related issues: #14617, #29506, #31349

To reply, just mention @dosu.

---

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

   

[fatelei]

set PIP_MIRROR_URL to aliyun

[tianshiqingge]

@dosu According to your method, the time for PYTHON_ENV_NIT_TIMEOUT has been increased, and the address for PIP-MIROR-URL has been set to https://mirror.aliyuncs.com I have checked the resource limit of the container and it has not been configured. It can occupy all system resources and now the system space is sufficient

[tianshiqingge]

@fatelei The address for PIP-MIRROR_LURL has been set to https://mirror.aliyuncs.com Installing the plugin still produces errors

[dosubot]

A few things to check:

1. PIP_MIRROR_URL format - The URL must end with /simple. Your current URL https://mirror.aliyuncs.com is incorrect. Use:

PIP_MIRROR_URL=https://mirrors.aliyun.com/pypi/simple

2. Add UV_HTTP_TIMEOUT - Add this environment variable for the plugin_daemon to allow longer dependency downloads:

UV_HTTP_TIMEOUT=120s

3. Clear stuck installation tasks - Previous failed installations may be blocking. Clear them via API:

POST http://localhost/console/api/workspaces/current/plugin/tasks/delete_all

(Use browser F12 to get X-Csrf-Token and Cookie headers)

4. Check plugin_daemon logs for more details:

docker logs docker-plugin_daemon-1

After fixing the PIP_MIRROR_URL format and adding UV_HTTP_TIMEOUT, restart the plugin_daemon container and try installing again.

References: #20387, #29589

To reply, just mention @dosu.

---

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

   

[fatelei]

@fatelei The address for PIP-MIRROR_LURL has been set to https://mirror.aliyuncs.com Installing the plugin still produces errors

increase PYTHON_ENV_INIT_TIMEOUT

[fatelei]

The process monitors UV output activity and kills the process if there's no output for 120 seconds (2 minutes):

• Slow network/PyPI downloads
• Compiling packages from source
• Resource constraints

[tianshiqingge]

@fatelei PYTHON_ENV_INIT_TIMEOUT=320 That's how I set it up

[tianshiqingge]

@dosu Modified PIP-MIRROR_LURL= https://mirrors.aliyun.com/pypi/simple After configuring, it can be installed normally, but the error message from installing the plugin before cannot be cleared. Therefore, the plugin installation cache should be cleared. There are no more installation tasks being executed in the system, so it should be in the cache. I don't know how to clear it?

[tianshiqingge]

@fatelei It has been resolved, thank you

[dosubot]

Glad to hear the PIP_MIRROR_URL fix worked! To clear those stuck/failed installation task records:

Step 1: Get your tenant_id from the database:

docker compose exec db psql -U postgres -d dify
select id from tenants;

Step 2: Call the API from inside the API container:

docker exec -it docker-api-1 bash

curl -X POST "http://plugin_daemon:5002/plugin/{tenant_id}/management/install/tasks/delete_all" \
     -H "X-Api-Key: {PLUGIN_DAEMON_KEY}"

Replace {tenant_id} with the UUID from step 1, and {PLUGIN_DAEMON_KEY} with the value from your .env file.

Alternative - If you have access via browser, you can also try:

POST http://localhost/console/api/workspaces/current/plugin/tasks/delete_all

Using X-Csrf-Token and Cookie headers from browser DevTools (F12).

After clearing the tasks, the error messages should disappear from the UI.

References: #30825, #30900

To reply, just mention @dosu.

---

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