Title: Python Node Consistently Returns "-500 no space left on device" Error with Simple Data Processing

[chihtsenhuang-cpu]

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

Description

I'm experiencing a persistent error in Dify's cloud version where Python nodes fail with -500 no space left on device during simple data processing operations. The error occurs reliably even with minimal input data and straightforward code logic.

Error Message

Got error code: -500. Got error msg: open /var/sandbox/sandbox-python/tmp/0ca1bbb0_a680_4aa0_a8ef_cf171dc072b6.py: no space left on device

Input Data Sample

{
  "arr1": [
    {
      "metadata": {
        "_source": "knowledge",
        "dataset_id": "e53c7839-bb13-4da0-a5f0-dfd414bf9cb5",
        "dataset_name": "勞動法規與施行細則",
        "document_id": "3ec38e0c-fd2d-4b6c-9f30-e4e95085ee4c",
        "document_name": "勞工保險條例_formatted.txt",
        "data_source_type": "upload_file",
        "segment_word_count": 383
      },
      "title": "勞工保險條例_formatted.txt",
      "content": "【文件】勞工保險條例\n【條號】第6條...[383 characters total]"
    }
  ],
  "arr2": [],
  "arr3": []
}

Python Node Code

def main(arr1: list = None, arr2: list = None, arr3: list = None) -> dict:
    tagged = []
    
    def process_data(data_list, tag):
        if isinstance(data_list, list):
            for item in data_list:
                if isinstance(item, dict):
                    item["source_type"] = tag
                    tagged.append(item)
                elif isinstance(item, str):
                    tagged.append({"content": item, "source_type": tag})
    
    process_data(arr1, "【法規】")
    process_data(arr2, "【解釋令】")
    process_data(arr3, "【職災人因】")
    
    return {"result": tagged}

Steps to Reproduce

1. Create a workflow with RAG nodes retrieving from multiple knowledge bases
2. Add a Python node with the above code
3. Pass arrays of RAG results (similar to the sample above)
4. Execute the workflow

Observations

• Code is simple: Only basic list iteration and dictionary append operations
• Input is small: ~383 characters of text data; arr2 and arr3 are empty
• No infinite loops or recursion: Straightforward single-pass processing
• No large intermediate structures: Only appending items to a list
• Issue persists across multiple workflows: Reproduced with copied/new workflows

Environment

• Deployment: Dify Official Cloud (dify.ai)
• Python Node: Standard Dify Python runtime

Expected Behavior

The Python node should process the input arrays and return a dictionary with tagged results.

Actual Behavior

Consistently fails with -500 no space left on device error before completing execution.

Additional Context

This appears to be a sandbox container issue (/var/sandbox/sandbox-python/tmp) rather than a code logic problem, as:

• The error occurs with minimal data and simple operations
• Multiple isolated workflows exhibit the same error
• The code follows best practices (type hints, guard clauses)

Would appreciate assistance in diagnosing whether this is a temporary resource issue or a configuration problem with the Python sandbox environment.

[dosubot[bot]]

Hi @chihtsenhuang-cpu! I'm Dosu and I’m helping the dify team.

This is a known sandbox infrastructure issue — the -500 error occurs when /var/sandbox/sandbox-python/tmp/ runs out of disk space, and it's not caused by your code. The sandbox writes a temporary .py file for each execution but has no automatic cleanup mechanism, so these files accumulate until the disk is full [1].

Your analysis is correct: with ~383 characters of input and straightforward list operations, this is purely a disk exhaustion problem on the sandbox container. There are actually two other issues open today reporting the same thing on Dify Cloud: #34346 and #34347.

Since you're on Dify Cloud, your options are unfortunately limited — you can't access the sandbox container to clear temp files yourself. This needs to be addressed by the Dify infrastructure team. Reducing workflow execution frequency may help as a temporary measure, but it won't fully resolve the issue.

For anyone running self-hosted Dify hitting the same error, there are a few workarounds:

1. Clear temp files manually:

   docker exec -it <sandbox_container> sh -c "rm -f /var/sandbox/sandbox-python/tmp/*.py"

2. Mount a tmpfs in docker-compose.yaml to auto-clear on restart and cap usage [2]:

   sandbox:
     tmpfs:
       - /var/sandbox/sandbox-python/tmp:size=512m

3. Set up periodic cleanup:

   find /var/sandbox/sandbox-python/tmp -name "*.py" -mmin +10 -delete

It's worth noting that the Dify codebase currently has no built-in automatic cleanup for sandbox temp files — this is an unresolved gap.

To reply, just mention @dosu.

---

Docs are dead. Just use Dosu.