CI/TST: Call join on server process test

[mroeschke]

Troubleshooting this current CI failure

    @pytest.fixture
    def responder(request):
        """
        Fixture that starts a local http server in a separate process on localhost
        and returns the port.
    
        Running in a separate process instead of a thread to allow termination/killing
        of http server upon cleanup.
        """
        # Find an available port
        with socket.socket() as sock:
            sock.bind(("localhost", 0))
            port = sock.getsockname()[1]
    
        server_process = multiprocessing.Process(
            target=process_server, args=(request.param, port)
        )
        server_process.start()
        yield port
        server_process.terminate()
        kill_time = 5
        wait_time = 0
        while server_process.is_alive():
            if wait_time > kill_time:
                server_process.kill()
                break
            else:
                wait_time += 0.1
                time.sleep(0.1)
>       server_process.close()

pandas/tests/io/test_user_agent.py:231: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <Process name='Process-10' pid=13556 parent=8289 started>

    def close(self):
        '''
        Close the Process object.
    
        This method releases resources held by the Process object.  It is
        an error to call this method if the child process is still running.
        '''
        if self._popen is not None:
            if self._popen.poll() is None:
>               raise ValueError("Cannot close a process while it is still running. "
                                 "You should first call join() or terminate().")
E               ValueError: Cannot close a process while it is still running. You should first call join() or terminate().

[jbrockmendel]

IIRC there was something about using pytest.mark.single for some of these. did that pan out?

At this point I'd be OK with just skipping this entire file.

[mroeschke]

IIRC there was something about using pytest.mark.single for some of these. did that pan out?

At this point I'd be OK with just skipping this entire file.

Ya I'm trying to thin out the number of builds before adding a single build because GHA can supposedly run ~20 builds concurrently across all PRs/main.

If this doesn't work I'll skip this file until we have that single process build up.

[mroeschke]

This appears like a victory.

Failure due to the runners

2022-01-26T03:47:04.4612135Z Found online and idle hosted runner in the current repository's organization account that matches the required labels: 'ubuntu-latest'
2022-01-26T03:47:04.5087280Z Waiting for a Hosted runner in the 'organization' to pick this job...
2022-01-26T03:47:05.3360343Z Job is waiting for a hosted runner to come online.
2022-01-26T03:47:10.9423229Z Job is about to start running on the hosted runner: GitHub Actions 5 (hosted)

[mroeschke]

@meeseeksdev backport 1.4.x

[lumberbot-app]

Something went wrong ... Please have a look at my logs.