fail early without needing to reduce file_read_timeout

[madhur-ob]

context:

consider the flow at f.py

from metaflow import FlowSpec, kubernetes, step, current, project, Flow
import os, time

class HelloFlow(FlowSpec):

    @step
    def start(self):
        x.foo()
        self.next(self.end)

    @step
    def end(self):
        print('value is', self.model)

if __name__ == '__main__':
    HelloFlow()

and consider using it with:

from metaflow import Runner

with Runner('./f.py').run() as running:
    print('done')

the runner gets stuck after merging of #2169

if we try the following, it exits when the timeout passes...

from metaflow import Runner

with Runner('./f.py', file_read_timeout=10).run() as running:
    print('done')

since the default is 3600 secs, we should exit early..

[romain-intel]

LGTM