ConcurrentMultiSpanProcessor: cannot schedule new futures after shutdown

Describe your environment

OS: macOS 15.2
Python version: 3.13.0

Dependencies from pip install opentelemetry-sdk:

deprecated==1.2.18
importlib-metadata==8.5.0
opentelemetry-api==1.30.0
opentelemetry-sdk==1.30.0
opentelemetry-semantic-conventions==0.51b0
typing-extensions==4.12.2
wrapt==1.17.2
zipp==3.21.0

What happened?

I tried to use the opentelemetry.sdk.trace.ConcurrentMultiSpanProcessor for a library but get an error on shutdown.

Steps to Reproduce

Based on the basic_trace.py example:

from opentelemetry import trace
from opentelemetry.sdk.trace import (
    ConcurrentMultiSpanProcessor,
    TracerProvider,
)
from opentelemetry.sdk.trace.export import (
    BatchSpanProcessor,
    ConsoleSpanExporter,
)

span_proc = ConcurrentMultiSpanProcessor()
span_proc.add_span_processor(BatchSpanProcessor(ConsoleSpanExporter()))

trace.set_tracer_provider(TracerProvider())
trace.get_tracer_provider().add_span_processor(span_proc)

tracer = trace.get_tracer(__name__)
with tracer.start_as_current_span("foo"):
    print("Hello world!")

Running with:

python basic_trace.py

Expected Result

Hello world!
{
    "name": "foo",
    "context": {
        "trace_id": "0xcfd5173b1021291f64aaf996aa5a653c",
        "span_id": "0x4b7f1558bcce9066",
        "trace_state": "[]"
    },
    "kind": "SpanKind.INTERNAL",
    "parent_id": null,
    "start_time": "2025-03-05T13:33:42.622950Z",
    "end_time": "2025-03-05T13:33:42.623034Z",
    "status": {
        "status_code": "UNSET"
    },
    "attributes": {},
    "events": [],
    "links": [],
    "resource": {
        "attributes": {
            "telemetry.sdk.language": "python",
            "telemetry.sdk.name": "opentelemetry",
            "telemetry.sdk.version": "1.30.0",
            "service.name": "unknown_service"
        },
        "schema_url": ""
    }
}

Actual Result

Hello world!
Exception ignored in atexit callback <bound method TracerProvider.shutdown of <opentelemetry.sdk.trace.TracerProvider object at 0x100ae74d0>>:
Traceback (most recent call last):
  File "/Users/helder/Projects/opentelemetry-python/docs/examples/basic_tracer/.venv/lib/python3.13/site-packages/opentelemetry/sdk/trace/__init__.py", line 1282, in shutdown
    self._active_span_processor.shutdown()
  File "/Users/helder/Projects/opentelemetry-python/docs/examples/basic_tracer/.venv/lib/python3.13/site-packages/opentelemetry/sdk/trace/__init__.py", line 180, in shutdown
    sp.shutdown()
  File "/Users/helder/Projects/opentelemetry-python/docs/examples/basic_tracer/.venv/lib/python3.13/site-packages/opentelemetry/sdk/trace/__init__.py", line 262, in shutdown
    self._submit_and_await(lambda sp: sp.shutdown)
  File "/Users/helder/Projects/opentelemetry-python/docs/examples/basic_tracer/.venv/lib/python3.13/site-packages/opentelemetry/sdk/trace/__init__.py", line 243, in _submit_and_await
    future = self._executor.submit(func(sp), *args, **kwargs)
  File "/Users/helder/Library/Application Support/uv/python/cpython-3.13.0-macos-aarch64-none/lib/python3.13/concurrent/futures/thread.py", line 170, in submit
    raise RuntimeError('cannot schedule new futures after shutdown')
RuntimeError: cannot schedule new futures after shutdown

Additional context

I can avoid the error by manually shutting down in the end of the script:

trace.get_tracer_provider().shutdown()

But since the tracer provider already calls shutdown on exit, can't this be handled automatically without the manual shutdown?

Would you like to implement a fix?

None

Start with opentelemetry/sdk/trace/init.py, especially ConcurrentMultiSpanProcessor._submit_and_await and TracerProvider.shutdown; compare them with the basic_trace.py reproduction. Verify the behavior when shutdown is invoked by the atexit callback, and consider the issue resolved when that example exits without the RuntimeError while still shutting down processors.