Runner is a project that manages starting tasks in a defined order. If tasks fail, the chain can be retried, or halted.
Configuration is done with INI style configuration files.
sleep_time: how long to wait between retries
max_tries: how many times to retry before giving up
halt_task: which task to run to "halt" the process. This could perhaps shut the machine down or terminate the EC2 instance
max_time: maximum amount of time a task can run
Keys and values in this section are passed into tasks as environment variables
Configuration for other tasks or purposes can go into their own sections.
Tasks are loaded from a task dir. Tasks are run as separate processes.
The return code of a task determines what happens next.
Return code 0 means everything went well, and to continue on to the next task.
Return code 2 means to run the "halt" task, and then stop running tasks and exit runner
Other return codes will cause runner to retry tasks from the beginning.
A special environment variable,
"RUNNER_CONFIG_CMD" is always set in the
environment that allows tasks to easily access configuration. e.g.
$RUNNER_CONFIG_CMD -g hg.remote
will return the "remote" configuration variable from the "hg" section
Tests are run via nose
python setup.py nosetests, or nose manually
- implement tasks
- exponential backoff for sleep time
- ccache cleanup (tmp files)
- repo cleanup (old *.git files)
- check slavealloc (changed master?)