容器内的进程无法创建新的线程 #493
Description
生产报错信息：
/opt/node/v22.2.0/bin/node[78]: std::unique_ptr node::WorkerThreadsTaskRunner::DelayedTaskScheduler::Start() at ../src/node_platform.cc
Assertion failed: (0) == (uv_thread_create(t.get(), start_thread, this))
Native stack trace:
1: 0xf66937 node::Assert(node::AssertionInfo const&) [/opt/node/v22.2.0/bin/node]
2: 0xf6f3e node::WorkerThreadsTaskRunner::WorkerThreadsTaskRunner(int) [/opt/node/v22.2.0/bin/node]
3: 0xfbc7b node::NodePlatform::NodePlatform(int, v8::TracingController*, v8::PageAllocator*) [/opt/node/v22.2.0/bin/node]
4: 0xf1d21c node::Start(int, char**) [/opt/node/v22.2.0/bin/node]
5: 0x7f300064f1ca __libc_start_main [/lib/x86_64-linux-gnu/libc.so.6]
6: 0x7f300064f28b _start [/lib/x86_64-linux-gnu/libc.so.6]
7: 0xe67fee _start [/opt/node/v22.2.0/bin/node]
[I 2026-03-19 02:14:53.959 ServerApp] jupyter_lsp | extension was successfully linked.
[I 2026-03-19 02:14:53.968 ServerApp] jupyter_server_terminals | extension was successfully linked.
[W 2026-03-19 02:14:53.977 ServerApp] 'token' has moved from NotebookApp to ServerApp. This config will be passed to ServerApp. Be sure to update your config before
[I 2026-03-19 02:14:53.977 ServerApp] jupyter_server_terminals | extension was successfully linked.
[I 2026-03-19 02:14:53.984 ServerApp] notebook | extension was successfully linked.
[I 2026-03-19 02:14:53.986 ServerApp] notebook | extension was successfully loaded.
[I 2026-03-19 02:14:54.590 ServerApp] Writing Jupyter server cookie secret to /root/.local/share/jupyter/runtime/jupyter_cookie_secret
[I 2026-03-19 02:14:54.590 ServerApp] notebook_shim | extension was successfully linked.
[I 2026-03-19 02:14:54.523 ServerApp] jupyter_lsp | extension was successfully loaded.
[I 2026-03-19 02:14:54.525 ServerApp] jupyter_server_terminals | extension was successfully loaded.
[I 2026-03-19 02:14:54.530 LabApp] JupyterLab extension loaded from /opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/site-packages/jupyterlab
[I 2026-03-19 02:14:54.530 LabApp] JupyterLab application directory is /opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/share/jupyter/lab
[I 2026-03-19 02:14:54.531 LabApp] Extension Manager is 'pippi'.
[I 2026-03-19 02:14:54.597 ServerApp] jupyterlab | extension was successfully loaded.
[I 2026-03-19 02:14:54.607 ServerApp] notebook | extension was successfully loaded.
[I 2026-03-19 02:14:54.608 ServerApp] Serving notebooks from local directory: /workspace
[I 2026-03-19 02:14:54.608 ServerApp] Jupyter Server 2.17.0 is running at:
[I 2026-03-19 02:14:54.608 ServerApp] http://127.0.0.1:44771/tree?token=...
[I 2026-03-19 02:14:54.608 ServerApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
[E 2026-03-19 02:14:54.617 ServerApp] Exception in callback functools.partial(<bound method IOLoop._discard_future_result of <tornado.platform.asyncio.AsyncIOMainLoop
<Task finished name='Task-2' coro=<initialize() done, defined at /opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/site-packages/jupyter_lsp/server
TimeoutError("can't start new thread")>
Traceback (most recent call last):
File "/opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/site-packages/tornado/ioloop.py", line 758, in _run_callback
ret = callback()
File "/opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/site-packages/tornado/ioloop.py", line 782, in _discard_future_result
future.result()
File "/opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/site-packages/jupyter_lsp/serverextension.py", line 24, in initialize
await nbapp.io_loop.run_in_executor(pool, manager.initialize)
File "/opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/site-packages/tornado/platform/asyncio.py", line 272, in run_in_executor
return self.asyncio_loop.run_in_executor(executor, func, *args)
File "/opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/asyncio/base_events.py", line 867, in run_in_executor
executor.submit(func, *args), loop=self)
File "/opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/concurrent/futures/thread.py", line 180, in submit
self._adjust_thread_count()
File "/opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/concurrent/futures/thread.py", line 203, in _adjust_thread_count
t.start()
File "/opt/python/versions/cpython-3.12.12-linux-x86_64-gnu/lib/python3.12/threading.py", line 994, in start
_start_new_thread(self._bootstrap, ())
RuntimeError: can't start new thread
使用ai工具分析得到的结论如下：
从你提供的 Docker 日志来看，核心问题非常明确：容器内的进程无法创建新的线程。
错误分析
日志中重复出现了以下关键错误信息：
- Go 语言运行时崩溃：
runtime/cgo: pthread_create failed: Operation not permitted
这表明 Go 程序试图调用底层系统函数pthread_create来创建线程时，被操作系统（Linux）拒绝，返回了“操作不允许”的错误。 - Node.js 崩溃：
Assertion failed: (0) == (uv_thread_create(t.get(), start_thread, this))
Node.js 基于 libuv，同样在尝试创建工作线程时失败，导致断言失败并退出。 - Python/Jupyter 异常：
RuntimeError: can't start new thread
Python 的threading模块或异步执行器在尝试启动新线程时抛出异常。
结论：这不是代码逻辑错误，而是Docker 容器的安全限制或资源限制导致的。你的容器环境缺少创建线程所需的权限（Capabilities），或者受到了严格的 Cgroups 限制。
- 我使用的是
opensandbox-server==0.1.2
code-interpreter 是根据v1.0.1 然后加入了一点业务需要使用的包自己构建的。
部署方式目前没上k8s。这个还在研究，相当于是使用单docker方式运行在云主机。后续还是会走k8s的模式部署生产。
其他配置均为默认的。几乎没有什么改动。
目前在开发环境没有出现过这个问题。测试环境偶尔会出现。生产环境因为调用量是最大的，频繁出现