bregman-arie / ci-runner Goto Github PK

Right now you can run jobs on localhost or remote host, but it would be nice to integrate it into a platform like OpenShift where we can quickly set up Pods and run the jobs on them.

Right now exceptions are LOG.error(...) + sys.exit(2) which is terrible.
Work has to be done to convert such code into proper custom exceptions.

For a given job, identify which type of node is used and pull all the relevant data. This will be used for provisioning a similar node in the user's own environment.

The data we should pull regarding the node is mainly based on the node type. For example, for Pods, it's critical we'll pull the Dockerfile so we can create a container/pod using that Dockerfile

Cover every function and method with proper unit tests