Is a timeout necessary to prevent `requests.get()` from blocking, if the network interface goes down temporarily?

Question

I've been developing an application, where I need to handle temporarily disconnects on the client (network interface goes down).

I initially thought the below approach would work, but sometimes if restart the network interface, the s.get(url) call would hang indefinitely:

s = requests.Session()
s.mount('http://stackoverflow.com', HTTPAdapter(max_retries=Retry(total=10, connect=10, read=10)))
s.get(url)

By adding the timeout=10 keyword argument to s.get(url) , the code is now able to handle this blocking behavior:

s = requests.Session()
s.mount('http://stackoverflow.com', HTTPAdapter(max_retries=Retry(total=10, connect=10, read=10)))
s.get(url, timeout=10)

Why is a timeout necessary to handle the cases, where a network interface resets or goes down temporarily? Why is max_retries=Retry(total=10, connect=10, read=10) not able to handle this? In particular, why is s.get() not informed that the network interface went offline, so that it could retry the connection instead of hanging?

Answer 1

Try: https://urllib3.readthedocs.io/en/latest/reference/urllib3.util.html#urllib3.util.retry.Retry

from requests.adapters import HTTPAdapter

s = requests.Session()
s.mount('http://stackoverflow.com', HTTPAdapter(max_retries=5))

Or:

retries = Retry(connect=5, read=2, redirect=5)
http = PoolManager(retries=retries)
response = http.request('GET', 'http://stackoverflow.com')

Or:

response = http.request('GET', 'http://stackoverflow.com', retries=Retry(10))