解决 Python RabbitMQ/Pika 报错:pop from an empty deque

使用 python 的 pika 包连接rabbitmq,代码如下:

python 复制代码
import pika
import threading
import time


def on_message(channel, method_frame, header_frame, body):
    print(f'on_message thread id: {threading.get_ident()}')
    delivery_tag = method_frame.delivery_tag

    print(body, "start")
    for i in range(10):
        print(i)
        time.sleep(20)
    print(body, "end")
    
    channel.basic_ack(delivery_tag)


credentials = pika.PlainCredentials('username', 'password')
parameters = pika.ConnectionParameters('test.webapi.username.com', credentials=credentials, heartbeat=5)
connection = pika.BlockingConnection(parameters)

channel = connection.channel()
channel.queue_declare(queue="standard", durable=True)
channel.basic_qos(prefetch_count=1)
channel.basic_consume('standard', on_message)

print(f'main thread id: {threading.get_ident()}')
try:
    channel.start_consuming()
except KeyboardInterrupt:
    channel.stop_consuming()
connection.close()

执行结果:

bash 复制代码
(Test) [user@user test]$ python mq1.py 
main thread id: 140682766104384
on_message thread id: 140682766104384
b'1' start
0
1
2
3
4
5
6
7
8
9
b'1' end
Traceback (most recent call last):
  File "mq1.py", line 38, in <module>
    channel.start_consuming()
  File "/home/user/conda/envs/Test/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 1865, in start_consuming
    self._process_data_events(time_limit=None)
  File "/home/user/conda/envs/Test/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 2026, in _process_data_events
    self.connection.process_data_events(time_limit=time_limit)
  File "/home/user/conda/envs/Test/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 833, in process_data_events
    self._dispatch_channel_events()
  File "/home/user/conda/envs/Test/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 567, in _dispatch_channel_events
    impl_channel._get_cookie()._dispatch_events()
  File "/home/user/conda/envs/Test/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 1492, in _dispatch_events
    consumer_info.on_message_callback(self, evt.method,
  File "mq1.py", line 24, in on_message
    channel.basic_ack(delivery_tag)
  File "/home/user/conda/envs/Test/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 2112, in basic_ack
    self._flush_output()
  File "/home/user/conda/envs/Test/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 1335, in _flush_output
    self._connection._flush_output(lambda: self.is_closed, *waiters)
  File "/home/user/conda/envs/Test/lib/python3.8/site-packages/pika/adapters/blocking_connection.py", line 523, in _flush_output
    raise self._closed_result.value.error
pika.exceptions.StreamLostError: Stream connection lost: ConnectionResetError(104, 'Connection reset by peer')

从结果来看,异常发生在一次长时间的消费过程(200s)完成后报错,具体为调用channel.basic_ack(delivery_tag)发生报错;推测是此时与MQ Server的连接已经被重置ConnectionResetError(104, 'Connection reset by peer'),此时再主动确认就发生报错。

正确解决方案如下:

python 复制代码
"""
@author: Zhigang Jiang
@date: 2022/1/16
@description:
"""
import functools
import pika
import threading
import time


def ack_message(channel, delivery_tag):
    print(f'ack_message thread id: {threading.get_ident()}')
    if channel.is_open:
        channel.basic_ack(delivery_tag)
    else:
        # Channel is already closed, so we can't ACK this message;
        # log and/or do something that makes sense for your app in this case.
        pass


def do_work(channel, delivery_tag, body):
    print(f'do_work thread id: {threading.get_ident()}')
    print(body, "start")
    for i in range(10):
        print(i)
        time.sleep(20)
    print(body, "end")

    cb = functools.partial(ack_message, channel, delivery_tag)
    channel.connection.add_callback_threadsafe(cb)


def on_message(channel, method_frame, header_frame, body):
    print(f'on_message thread id: {threading.get_ident()}')
    delivery_tag = method_frame.delivery_tag
    t = threading.Thread(target=do_work, args=(channel, delivery_tag, body))
    t.start()


credentials = pika.PlainCredentials('username', 'password')
parameters = pika.ConnectionParameters('test.webapi.username.com', credentials=credentials, heartbeat=5)
connection = pika.BlockingConnection(parameters)
channel = connection.channel()
channel.queue_declare(queue="standard", durable=True)
channel.basic_qos(prefetch_count=1)
channel.basic_consume('standard', on_message)

print(f'main thread id: {threading.get_ident()}')
try:
    channel.start_consuming()
except KeyboardInterrupt:
    channel.stop_consuming()
connection.close()

思路是pika是线程不安全的,所以在接收消息和ACK响应消息时需要另外线程。

相关推荐
CCblandy20 分钟前
前缀缓存命中率不靠引擎靠排列:本机仿真 3 轮,易变字段放开头 98.0%→0.0%
人工智能·python·算法·自然语言处理·性能优化
Sakura_j_h_1 小时前
给 FastAPI 接大模型:超时、重试和 token 成本该怎么管
python·大模型·fastapi·超时重试·token计费
forest_091 小时前
用Python复刻ClaudeDuet:主动式Agent四件套,定时晨报、后台巡检、危险动作审批(附完整可运行代码)
人工智能·python·架构·prompt·agi
小小张说故事1 小时前
Python 的函数调用到底贵在哪?拆开一次 call:51 纳秒栈帧与 *args 的 40 纳秒罚单
python
程序猿乐锅2 小时前
【从0-1学习Python | 第七篇】Python 面向对象基础
开发语言·笔记·后端·python·学习·1024程序员节
lzqrzpt2 小时前
太阳能控制器选型要点:电路工艺与工程适配技术解析
python·1024程序员节
databook3 小时前
时间序列数据中,如何找出缺失值
python·数据分析
小小龙学IT4 小时前
Python Polars 高性能数据处理引擎深度解析
开发语言·python·rust·1024程序员节
csdn_aspnet4 小时前
JavaScript 未排序数组中第 k 个最小/最大元素 (预期线性时间)
python
quantdash_cc4 小时前
如何利用标的信息为研究股票池补充名称和交易所信息?
开发语言·python·pandas·量化交易·1024程序员节·股票数据·quantdash