1 什么是分布式测试？

• 在进行本文之前，先了解些基础知识，什么是分布式测试？
• 分布式测试：是指通过局域网和Internet，把分布于不同地点、独立完成特定功能的测试计算机连接起来，以达到测试资源共享、分散操作、集中管理、协同工作、负载均衡、测试过程监控等目的的计算机网络测试。
• 通俗的讲：分布式测试 就是活太多，一个人干费时间，那就让多个人一起干，节省了资源和时间。

2 为什么要进行分布式测试？

2.1 场景1：自动化测试场景

• 自动化测试时，我们有很多用例，比如2000条用例，按照顺序执行，每条用例执行1分钟，那需要2000分钟；
• 什么概念？2000分钟就30多个小时，如果是冒烟测试，估计还没人工跑的快；
• 还有，如果是线上发布，跑完2000条用例就太浪费时间了；
• 那如果我们让我们让用例分布式执行，是不是可以节省很多时间？

2.2 场景2：性能测试场景

• 如果数据量很大，我们使用1台压测机，可能并发压力过大；
• 那就需要选择使用多台压测机（比如Jmeter的 Agent/负载机）；
• 这样也是一种分布式压测或者分布式性能测试场景。

所以总结来说，其实就是为了提升效率和质量。

3 分布式测试有什么特点？

4 分布式测试关键技术是什么？

5 分布式执行用例的前置条件是什么？

• 用例之间是独立且没有依赖关系，完全独立运行；
• 用例执行没有顺序，随机顺序都能正常执行；
• 每个用例都能重复运行，运行结果不会影响其他用例。

6 pytest-xdist安装

• pytest-xdist让自动化测试用例分布式执行，节省测试时间，属于进程级别的并发；
• 使用以下方法安装：

pip3 install pytest-xdist

C:\Users\Administrator>pip3 install pytest-xdist
Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple
Requirement already satisfied: pytest-xdist in d:\python37\lib\site-packages (1.31.0)
Requirement already satisfied: six in d:\python37\lib\site-packages (from pytest-xdist) (1.15.0)
Requirement already satisfied: execnet>=1.1 in d:\python37\lib\site-packages (from pytest-xdist) (1.8.0)
Requirement already satisfied: pytest>=4.4.0 in d:\python37\lib\site-packages (from pytest-xdist) (6.2.4)
Requirement already satisfied: pytest-forked in d:\python37\lib\site-packages (from pytest-xdist) (1.1.3)
Requirement already satisfied: apipkg>=1.4 in d:\python37\lib\site-packages (from execnet>=1.1->pytest-xdist) (1.5)
Requirement already satisfied: toml in d:\python37\lib\site-packages (from pytest>=4.4.0->pytest-xdist) (0.10.2)
Requirement already satisfied: attrs>=19.2.0 in d:\python37\lib\site-packages (from pytest>=4.4.0->pytest-xdist) (20.3.0)
Requirement already satisfied: colorama in d:\python37\lib\site-packages (from pytest>=4.4.0->pytest-xdist) (0.4.4)
Requirement already satisfied: atomicwrites>=1.0 in d:\python37\lib\site-packages (from pytest>=4.4.0->pytest-xdist) (1.4.0)
Requirement already satisfied: pluggy<1.0.0a1,>=0.12 in d:\python37\lib\site-packages (from pytest>=4.4.0->pytest-xdist) (0.13.1)
Requirement already satisfied: py>=1.8.2 in d:\python37\lib\site-packages (from pytest>=4.4.0->pytest-xdist) (1.10.0)
Requirement already satisfied: importlib-metadata>=0.12 in d:\python37\lib\site-packages (from pytest>=4.4.0->pytest-xdist) (2.1.1)
Requirement already satisfied: packaging in d:\python37\lib\site-packages (from pytest>=4.4.0->pytest-xdist) (20.8)
Requirement already satisfied: iniconfig in d:\python37\lib\site-packages (from pytest>=4.4.0->pytest-xdist) (1.1.1)
Requirement already satisfied: zipp>=0.5 in d:\python37\lib\site-packages (from importlib-metadata>=0.12->pytest>=4.4.0->pytest-xdist) (1.2.0)
Requirement already satisfied: pyparsing>=2.0.2 in d:\python37\lib\site-packages (from packaging->pytest>=4.4.0->pytest-xdist) (2.4.7)

7 pytest-xdist的优势

• 测试运行并行化；
• 在子进程中重复运行测试；
• 可指定不同的Python解释程序或不同的平台，并行运行测试。

8 pytest-xdist的使用

8.1 普通执行

# -*- coding:utf-8 -*-
# 作者：虫无涯
# 日期：2023/3/16 
# 文件名称：test_xdist.py
# 作用：pytest-xdist分布式测试
# 联系：VX(NoamaNelson)
# 博客：https://blog.csdn.net/NoamaNelson

import pytest
import time

class TestCase01():
    def test_case_01(self):
        time.sleep(1)
        print("case01$$$$$$$$$$$$$$$$$$$$$")

    def test_case_02(self):
        time.sleep(1)
        print("case02$$$$$$$$$$$$$$$$$$$$$")

    def test_case_03(self):
        time.sleep(1)
        print("case03$$$$$$$$$$$$$$$$$$$$$")

    def test_case_04(self):
        time.sleep(1)
        print("case04$$$$$$$$$$$$$$$$$$$$$")

    def test_case_05(self):
        time.sleep(1)
        print("case05$$$$$$$$$$$$$$$$$$$$$")

    def test_case_06(self):
        time.sleep(1)
        print("case06$$$$$$$$$$$$$$$$$$$$$")

class TestCase02():
    def test_case_07(self):
        time.sleep(1)
        print("case07$$$$$$$$$$$$$$$$$$$$$")

    def test_case_08(self):
        time.sleep(1)
        print("case08$$$$$$$$$$$$$$$$$$$$$")

    def test_case_09(self):
        time.sleep(1)
        print("case08$$$$$$$$$$$$$$$$$$$$$")


if __name__ == '__main__':
    pytest.main(["-s", "test_xdist.py"])

• 执行结果如下，使用了9.14s：

test_xdist.py::TestCase01::test_case_01 
test_xdist.py::TestCase01::test_case_02 
test_xdist.py::TestCase01::test_case_03 
test_xdist.py::TestCase01::test_case_04 
test_xdist.py::TestCase01::test_case_05 
test_xdist.py::TestCase01::test_case_06 
test_xdist.py::TestCase02::test_case_07 PASSED                           [ 11%]case01$$$$$$$$$$$$$$$$$$$$$
PASSED                           [ 22%]case02$$$$$$$$$$$$$$$$$$$$$
PASSED                           [ 33%]case03$$$$$$$$$$$$$$$$$$$$$
PASSED                           [ 44%]case04$$$$$$$$$$$$$$$$$$$$$
PASSED                           [ 55%]case05$$$$$$$$$$$$$$$$$$$$$
PASSED                           [ 66%]case06$$$$$$$$$$$$$$$$$$$$$
PASSED                           [ 77%]case07$$$$$$$$$$$$$$$$$$$$$

test_xdist.py::TestCase02::test_case_08 PASSED                           [ 88%]case08$$$$$$$$$$$$$$$$$$$$$

test_xdist.py::TestCase02::test_case_09 PASSED                           [100%]case08$$$$$$$$$$$$$$$$$$$$$


============================== 9 passed in 9.14s ==============================

8.2 上述代码分布式执行：

• 执行命令：

pytest -s -n auto test_xdist.py

• 结果如下，用时4.51s，可见分布式执行后大大缩短了测试时间：

(venv) F:\pytest_study\test_case\test_j>pytest -s -n auto test_xdist.py
============================================ test session starts =============================================
platform win32 -- Python 3.7.0, pytest-6.2.4, py-1.10.0, pluggy-0.13.1
rootdir: F:\pytest_study, configfile: pytest.ini
plugins: allure-pytest-2.8.12, assume-2.4.3, cov-2.8.1, forked-1.1.3, html-2.0.1, metadata-1.8.0, ordering-0.6,
 repeat-0.9.1, rerunfailures-10.3, xdist-1.31.0
gw0 [9] / gw1 [9] / gw2 [9] / gw3 [9] / gw4 [9] / gw5 [9] / gw6 [9] / gw7 [9]
.........
============================================= 9 passed in 4.51s ==============================================

8.3 指定CPU运行数量

• -n auto：可以自动检测到系统的CPU核数；
• 使用auto利用了所有CPU来跑用例；
• 也可以指定使用几个CPU来跑用例：

# x为cpu个数
pytest -s -n x

• 如下可以看到使用两个CPU来跑用例时长为6.27s：

(venv) F:\pytest_study\test_case\test_j>pytest -s -n 2 test_xdist.py
============================================ test session starts =============================================
platform win32 -- Python 3.7.0, pytest-6.2.4, py-1.10.0, pluggy-0.13.1
rootdir: F:\pytest_study, configfile: pytest.ini
plugins: allure-pytest-2.8.12, assume-2.4.3, cov-2.8.1, forked-1.1.3, html-2.0.1, metadata-1.8.0, ordering-0.6,
 repeat-0.9.1, rerunfailures-10.3, xdist-1.31.0
gw0 [9] / gw1 [9]
.........
============================================= 9 passed in 6.27s ==============================================

8.4 与pytest-html一起使用

• 命令如下：

pytest -s -n auto --html=report.html --self-contained-html

• 运行结果：

pytest -s -n auto test_xdist.py --html=report.thml --self-contained-htm
l

gw0 [9] / gw1 [9] / gw2 [9] / gw3 [9] / gw4 [9] / gw5 [9] / gw6 [9] / gw7 [9]
.........
------------------ generated html file: file://F:\pytest_study\test_case\test_j\report.thml ------------------
============================================= 9 passed in 4.68s ==============================================

8.5 让pytest-xdist按照指定顺序执行

• pytest-xdist执行默认是无须的；
• 可通过 --dist 参数来控制顺序；

8.6 pytest-xdist如何保持session执行一次

• pytest-xdist没有内置的支持来确保会话范围的夹具仅执行一次；
• 可使用FileLock方法仅仅产生一次fixture数据：

import pytest
from filelock import FileLock


@pytest.fixture(scope="session")
def login():
    print("====登录===")
    with FileLock("session.lock"):
        name = "zhang"
        password= "123456"
        # web ui自动化
        # 声明一个driver，再返回

        # 接口自动化
        # 发起一个登录请求，将token返回都可以这样写

    yield name, password
    print("====退出====")