OverflowError: timeout is too large

[DineshDevaraj]

If I set the log level to ERROR or anything higher I see OverflowError error

Code

import time
import logging
from kafka import KafkaProducer

logging.basicConfig(level=logging.ERROR)
producer = KafkaProducer(bootstrap_servers=["kafka-in-docker:9092"], api_version=(7,3,1))
producer.send("foobar", b"hello world")
time.sleep(3)

Error

ERROR:kafka.producer.sender:Uncaught error in kafka producer I/O thread
Traceback (most recent call last):
  File "/usr/local/lib/gravity/poetry/venvs/hpe-cnx-ctb-b6oAM6xe-py3.10/lib/python3.10/site-packages/kafka/producer/sender.py", line 60, in run
    self.run_once()
  File "/usr/local/lib/gravity/poetry/venvs/hpe-cnx-ctb-b6oAM6xe-py3.10/lib/python3.10/site-packages/kafka/producer/sender.py", line 160, in run_once
    self._client.poll(timeout_ms=poll_timeout_ms)
  File "/usr/local/lib/gravity/poetry/venvs/hpe-cnx-ctb-b6oAM6xe-py3.10/lib/python3.10/site-packages/kafka/client_async.py", line 600, in poll
    self._poll(timeout / 1000)
  File "/usr/local/lib/gravity/poetry/venvs/hpe-cnx-ctb-b6oAM6xe-py3.10/lib/python3.10/site-packages/kafka/client_async.py", line 634, in _poll
    ready = self._selector.select(timeout)
  File "/opt/pyenv/versions/3.10.12/lib/python3.10/selectors.py", line 469, in select
    fd_event_list = self._selector.poll(timeout, max_ev)
OverflowError: timeout is too large

Platform

I am running this code
with kafka-python==2.0.5
on Python 3.10.12
on Ubuntu 22.04.5 LTS (Jammy Jellyfish)
in Docker Desktop 4.38.0 (181591)
on MacOS Sequoia 15.3.1 (24D70)

[Lasall]

I experience the same problem on MSK-brokers since the new kafka-python version 2.0.3+ (sometimes the send fails). I attached following debug output (broker, addresses adjusted to Dinesh's example):

DEBUG kafka.producer.kafka:696 Requesting metadata update for topic foobar
DEBUG kafka.metrics.metrics:156 Added sensor with name bytes-sent
DEBUG kafka.metrics.metrics:156 Added sensor with name bytes-received
DEBUG kafka.metrics.metrics:156 Added sensor with name request-latency
DEBUG kafka.metrics.metrics:156 Added sensor with name node-bootstrap-0.bytes-sent
DEBUG kafka.metrics.metrics:156 Added sensor with name node-bootstrap-0.bytes-received
DEBUG kafka.metrics.metrics:156 Added sensor with name node-bootstrap-0.latency
DEBUG kafka.conn:368 <BrokerConnection client_id=kafka-python-producer-1, node_id=bootstrap-0 host=kafka-in-docker:9092 <disconnected> [unspecified None]>: creating new socket
DEBUG kafka.conn:378 <BrokerConnection client_id=kafka-python-producer-1, node_id=bootstrap-0 host=kafka-in-docker:9092 <disconnected> [IPv4 ('127.0.0.1', 9092)]>: setting socket option (6, 1, 1)
INFO kafka.conn:384 <BrokerConnection client_id=kafka-python-producer-1, node_id=bootstrap-0 host=kafka-in-docker:9092 <connecting> [IPv4 ('127.0.0.1', 9092)]>: connecting to kafka-in-docker:9092 [('127.0.0.1', 9092) IPv4]
DEBUG kafka.client:592 Timeouts: user 9999999990.000000, metadata inf, idle connection inf, request inf
ERROR kafka.producer.sender:62 Uncaught error in kafka producer I/O thread
Traceback (most recent call last):
File "/app/.venv/lib/python3.11/site-packages/kafka/producer/sender.py", line 60, in run
self.run_once()
File "/app/.venv/lib/python3.11/site-packages/kafka/producer/sender.py", line 160, in run_once
self._client.poll(timeout_ms=poll_timeout_ms)
File "/app/.venv/lib/python3.11/site-packages/kafka/client_async.py", line 600, in poll
self._poll(timeout / 1000)
File "/app/.venv/lib/python3.11/site-packages/kafka/client_async.py", line 634, in _poll
ready = self._selector.select(timeout)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/selectors.py", line 468, in select
fd_event_list = self._selector.poll(timeout, max_ev)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
OverflowError: timeout is too large

Potentially with #2480 the reconnect_backoff_ms setting is no longer used, resulting in the 9999999990.000000 to be used as minimum timeout.

[dpkp]

This should be mostly fixed in master. I'll try to release 2.0.6 with the fix today.