Skip to content

Errors might be good!

Compare
Choose a tag to compare
@groundnuty groundnuty released this 31 Oct 01:51
· 59 commits to master since this release

During development of a test framework based on k8s jobs/pods the need arose to generalize the waiting condition to allow for pods and jobs to fail. For that, I added new options pod-we (read: pods with errors) and job-we (read: job with errors). By using those, your wait will end when eg. all the pods that are members of a job will enter into Completed or Failed states.

The complete changelog:

  • refactored go-template code to use define block - it looks much cleaner now
  • added pod-we and job-we that treat erroneous jobs and pods as completed
  • upgrade kubectl to version v1.11.0 thanks to pull request by @kostegit
  • fixed a bug where state for a single pod was reported wrongly
  • fixed argument expansion in get_job_state function
  • improved Makefile to infer repo name