[Core][runtime_env] Avoid zombie subprocesses in check_output_cmd

[d3c3mber]

When PipPlugin.delete_uri cancels the outer task, the pending CancelledError caused 'await proc.wait()' to return immediately without reaping the child, leaving a process that got re-parented to init.

Wrap the wait in asyncio.shield + a bounded timeout, and fall back to transport.close() so the asyncio child watcher can reap the child in the background.

Thank you for contributing to Ray! 🚀
Please review the Ray Contribution Guide before opening a pull request.

⚠️ Remove these instructions before submitting your PR.

💡 Tip: Mark as draft if you want early feedback, or ready for review when it's complete.

Description

Briefly describe what this PR accomplishes and why it's needed.

Related issues

Link related issues: "Fixes #1234", "Closes #1234", or "Related to #1234".

Additional information

Optional: Add implementation details, API changes, usage examples, screenshots, etc.

[gemini-code-assist]

Code Review

This pull request addresses a zombie process leak in check_output_cmd by wrapping the process wait call in asyncio.shield and asyncio.wait_for during cleanup, and adds corresponding unit tests. The review feedback suggests two important improvements: first, ensuring that any caught asyncio.CancelledError is re-raised to prevent swallowing task cancellation; second, checking the specific spawned process PID in the zombie test instead of using a wildcard os.waitpid(-1, ...) to avoid flakiness in concurrent test environments.