Skip to content

Latest commit

 

History

History
75 lines (47 loc) · 2.49 KB

DEVELOPERS.md

File metadata and controls

75 lines (47 loc) · 2.49 KB
Raw

Contributing and developing

We welcome any contribution or feedback to vpc-router. Please use the issue tracker to file bugs or request features.

Installation of required dependencies

When developing or contributing to vpc-router it is helpful to install some useful packages for running local unit tests. You can do so with this command (after you have performed the install outlined in the README):

$ pip install -r requirements/develop.txt

Running out of the source tree

Use the vpcrouter-runner.py script to run the program without having to actually install (setup.py) the application.

$ ./vpcrouter-runner.py ...

Testing

To run all unit tests, use the run_tests.sh script, like so:

$ ./run_tests.sh

This also creates code coverage reports in HTML, which are by default stored in /tmp/vpc-router-coverage.

To just run specific tests, use (for example):

$ python -m unittest vpcrouter.tests.test_utils

Use the style_test.sh script to check some common pep8 style guides and code complexity.

$ ./style_test.sh

Architecture

The architecture of vpc-router is simple:

  • A health-monitor thread detects if there are any failed hosts (vpcrouter.monitor.plugins.*).
  • A configuration-watcher thread detects if there are any updates to the routing configuration (vpcrouter.watcher.plugins.*)
  • A main loop receives notifications from both those threads via queues (vpcrouter.watcher._event_monitor_loop).
  • If an update is received on either queue (failed hosts or new config) the 'route-spec' is processed (vpcrouter.vpc.handle_spec): Check the current routes in VPC against the spec, see if all requested routes are present and if the current routers for each route are still healthy. If this is not the case the route is updated or removed or a new route is added.
  • If a new route configuration is received, the main event loop updates the health-monitor thread with the new combined list of all hosts, via a third queue.
  • The configuration-watcher thread utilizes plugins, which are dynamically loaded, and which are used depending on the selected mode (-m option on the command line).
  • The health monitor thread also uses plugins to support different instance health monitoring options (--health option on the command line). Those plugins are conceptually very similar to the watcher plugins.

Developing plugins

Please read the plugin developer documentation.