python 超时重试方法

在应用中,有时候会 依赖第三方模块执行方法,比如调用某模块的上传下载,数据库查询等操作的时候,如果出现网络问题或其他问题,可能有超时重新请求的情况;

目前的解决方案有

1. 信号量,但不支持window;

2.多线程,但是 如果是大量的数据重复操作尝试,会出现线程管理混乱,开启上万个线程的问题;

3.结合采用 eventlet 和 retrying模块 (eventlet 原理尚需深入研究)

下面的方法实现:超过指定时间重新尝试某个方法

# -*- coding: utf-8 -*-
import random
import time

import eventlet
from retrying import retry

eventlet.monkey_patch()


class RetryTimeOutException(Exception):
    def __init__(self, *args, **kwargs):
        pass


def retry_if_timeout(exception):
    """Return True if we should retry (in this case when it's an IOError), False otherwise"""
    return isinstance(exception, RetryTimeOutException)


def retry_fun(retries=3, timeout_second=2):
    """
    will retry ${retries} times when process time beyond ${timeout_second} ;
    :param retries: The retry times
    :param timeout_second: The max process time
    """

    def retry_decor(func):
        @retry(stop_max_attempt_number=retries, retry_on_exception=retry_if_timeout)
        def decor(*args, **kwargs):
            print("In retry method..")
            pass_flag = False
            with eventlet.Timeout(timeout_second, False):
                r = func(*args, **kwargs)
                pass_flag = True
                print("Success after method.")
            if not pass_flag:
                raise RetryTimeOutException("Time out..")
            print("Exit from retry.")
            return r

        return decor

    return retry_decor


def do_request():
    print("begin request...")
    sleep_time = random.randint(1, 4)
    print("request sleep time: %s." % sleep_time)
    time.sleep(sleep_time)
    print("end request...")
    return True


@retry_fun(retries=3)
def retry_request():
    r = do_request()
    print(r)


if __name__ == '__main__':
    retry_request()

  参考:

安装依赖模块:pip install retrying eventlet -i https://pypi.tuna.tsinghua.edu.cn/simple/

装饰器用法:https://blog.csdn.net/u013205877/article/details/78872278

retry: https://blog.csdn.net/lxy210781/article/details/95253026

超时:https://blog.csdn.net/yuanpython/article/details/90522567

其他方法:https://www.cnblogs.com/lyxdw/p/10033118.html

 

posted @ 2019-09-28 13:56  百变小超  阅读(1790)  评论(1编辑  收藏  举报