pytest-netdut 0.3.0

"Automated software testing for switches using pytest"

Netdut: automated software testing for switches using pytest

Netdut is a pytest plugin which provides infrastructure (e.g. pytest fixtures) which make it easy to write automated tests for network switches.

Features

• Brings the power, maturity and ecosystem of pytest to testing on network switches.
• The dut fixture (Device Under Test) providing serial, ssh or EAPI connectivity for running CLI commands on a network switch.
• Command-line configuration of a hostname and console name.
• Markers for skipping tests based on the device's type or software configuration.
• Compatibility with both Arista's EOS operating system, and the Metamako MOS operating system.
• A pythonic interfaces for writing EOS CLI commands.

Requirements

• Requires python3 >= 3.6, pexpect, pyeapi
• Network devices under test (DUTs) must be accessible via SSH.

Installation

You can install pytest-netdut via pip from PyPI:

$ pip install pytest-netdut

The pytest-netdut plugin is not auto-loaded upon installation. To enable pytest-netdut for your tests, add the following to a conftest.py file:

pytest_plugins = ["pytest_netdut"]

or explicitly add it when you run pytest:

pytest -p pytest_netdut

Usage

Basics

Several example tests using pytest-netdut are contained in examples/. Note that there is a conftest.py in that directory which enables pytest-netdut and increases the verbosity of the CLI output.

The DUT must be accessible via SSH to use the SSH or EAPI fixtures (the latter enables and disables the EAPI using SSH). This can usually be achieved using this EOS config:

    enable
        configure
            username admin privilege 15 nopassword
            aaa authorization exec default local

test_showver.py demonstrates some basic examples using the dut fixture. The tests can be be run using the command below.

pytest --device=<YOUR_DEVICE_IP_OR_DNS> examples/test_showver.py

test_showver runs show version via EAPI:

def test_showver(dut):
    info = dut.eapi.sendcmd("show version")
    logging.critical(f"DUT model was: {info['modelName']}")

Running this using a device with a hostname of dmb224 gives some simple output:

Assertions

Because this is pytest, we get great feedback when tests fail. Tests fail upon assertions. Here we use the test below as an example.

def test_check_version(dut):
    info = dut.eapi.sendcmd("show version")
    assert info["version"] == "1.0.0"

Netdut will start EAPI and use it to run the show version command, returning the results as a python dictionary. The assertion then fails, assuming the DUT is not running version 1.0.0 of EOS (in this case it is running an old internal build):

Running multiple commands

It's rare that a test will execute a single command on the DUT. Netdut allows for multi-command tests. This can be seen in examples/test_daemon.py where the daemon is configured:

dut.eapi.sendcmds("""
        enable
            configure
                daemon sleeper
                    exec /usr/bin/sleep 10
                    no shutdown
    """)

In this case the result that is returned is list of the results of each command. It's often useful to just return the last one:

def daemon_has_started():
    daemon_info = dut.eapi.sendcmds(["enable", f"show daemon {sleeper_daemon}"])[-1]
    return daemon_info["daemons"][sleeper_daemon]["starttime"] != 0.0

Waiting for results

It's often useful to provide some configuration to the switch, and then wait for the switch to act (since it's running asynchronously). pytest-netdut provides the wait_for fixture for this purpose. Again, in examples/test_daemon.py we use wait_for to wait for the daemon to start and stop.

# daemon_has_started is defined as a local function
# wait_for calls it repeatedly until it's true
wait_for(daemon_has_started, timeout=10.0)

Skipping based on DUT type

Perhaps some tests won't run on every type of device -- pytest-netdut provides some skipping mechanisms for this purpose. To skip a test on a particular on a DUT type use the provided decorator:

@pytest.mark.skip_device_type("DCS-7130.*", reason="Demo")
def test_that_skips_7130(dut):
    logging.info("Must not be 7130!")

Similar decorators exist for OS type (i.e. EOS-only or MOS-only tests), and can be stacked:

@pytest.mark.eos
@pytest.mark.skip_device_type("DCS-7130.*")
def test_that_only_runs_on_eos_on_7130(dut):
    logging.info("Must be EOS on 7130!")

Building test harnesses

Pytest's fixture mechanism is very powerful in this scenario, allowing us to set up and tear down test configurations with low overhead.

@pytest.fixture
def my_test_harness(dut):
    dut.eapi.sendcmds(["enable",
                          "configure",
                              "banner motd My test is running, hands off!"])
    yield
    dut.eapi.sendcmds(["enable", "configure", "no banner"])

def test_mytest(my_test_harness):
    logging.info("Running my test here")

Translations

In order to reduce test code verbosity and complexity a translator class provides a way to standardize tests to a particular OS. Both CLI commands and results are standardized to EOS by default.

A particular test can be written with EOS CLI and be run on MOS as well. Consider the l1 source command which differs between EOS and MOS:

@pytest.fixture
def l1_connect(dut):
    if dut.ssh.cli_flavor == "eos":
        dut.eapi.sendcmds(["enable",
                              "configure",
                                  "interface Ethernet10",
                                      "l1 source interface Ethernet12"])
    elif dut.ssh.cli_flavor == "mos":
        dut.eapi.sendcmds(["enable",
                              "configure",
                                  "interface Ethernet10",
                                      "source Ethernet12"])

which can be reduced to:

@pytest.fixture
def l1_connect(dut):
    dut.eapi.sendcmds(["enable",
                        "configure",
                            "interface Ethernet10",
                                "l1 source interface Ethernet12"])

The translator will try to find a match for each command in the predefined list of regex configuration patterns.

The translator will also process return values; MOS EAPI result keys are camelCased and the translator will convert all keys to snake_case.

The translator has a predefined set of translations which can be extended by subclassing the Translator class and overriding config_patterns. Return values are processed by the translate_key function which must be defined in the subclass.

Set the new translator class via eapi.set_translator(<class instance>).

Contributing

This project is under active use and development. We would appreciate help to improve it, via pull request. Tests can be run with make ci or tox.

Docstrings are according to Google's docstring conventions (examples).

License

Distributed under the terms of the BSD-3 license, pytest-netdut is free and open source software

Issues

If you encounter any problems, please file an issue along with a detailed description.

Copyright (c) 2021, David Snowdon All rights reserved.

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

• Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

• Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.

• Neither the name of pytest-netdut nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.