MCP sanity checks

This is salt-based set of tests for basic verification of MCP deployments

How to start

Target: test engineers

1. Clone repo to any node (node must have an access via http to salt master):

   # root@cfg-01:~/# git clone https://github.com/Mirantis/cvp-sanity-checks
   # cd cvp-sanity-checks

Use git config --global http.proxy http://proxyuser:proxypwd@proxy.server.com:8080 if needed.

2. Install virtualenv

   # curl -O https://pypi.python.org/packages/source/v/virtualenv/virtualenv-X.X.tar.gz
   # tar xvfz virtualenv-X.X.tar.gz
   # cd virtualenv-X.X
   # sudo python setup.py install

or

   # apt-get install python-virtualenv

3. Create virtualenv and install requirements and package:

   # virtualenv --system-site-packages .venv
   # source .venv/bin/activate
   # pip install --proxy http://$PROXY:8678 -r requirements.txt
   # python setup.py install
   # python setup.py develop

4. Configure:

   # vim cvp-sanity/global_config.yaml

SALT credentials are mandatory for tests.

Other settings are optional (please keep uncommented with default values)

Alternatively, you can specify these settings via env variables:

export SALT_URL=http://10.0.0.1:6969

For array-type settings please do:

export skipped_nodes=ctl01.example.com,ctl02.example.com

5. Start tests:

   # pytest --tb=short -sv cvp-sanity/tests/

or

   # pytest -sv cvp-sanity/tests/ --ignore cvp-sanity/tests/test_mtu.py

CVP-sanity-checks supports tags (labels/marks/sets). As of now we have smoke, full and sl_dup sets. Smoke will run essential tests only, full set runs all tests and sl_dup is a special set that collects the tests that already exist in Stacklight. Please do not forget to mark your test when you add it. Example (run smoke tests only):

   # pytest -v -m smoke

Logging

Target: test developers

To make a logging in the cvp-sanity-tests module more consistent you should follow next recommendations:

• Do not use print methods

• Each string that needs to be in the stdout should be called with logging.warning level and higher

import logging

def test_t1():
    logging.warning("Alert: Houston we have a problem!")
    logging.error("We've lost a booster!")
    logging.critical("Everything is broken!")

So these messages will be displayed in the pytest output after failed tests

• Use logging.info and logging.debug to log some trivial string. E.g saving requests/responses, saving states os some variables

import logging

def test_t2():
    logging.info("Skip {node} in this test")
    logging.debug("Some function is called")

These messages will be logged in the log file only. You can forcely enable logging in the sdtout by calling pytest with next parameter

pytest --log-level=DEBUG

Logging to file

By default full.log is located in the working directory. And you can define folder for logs to store by the PYTEST_REPORT_DIR environment variable

List of available environment variables

todo