Skip to content
This repository has been archived by the owner on Mar 3, 2023. It is now read-only.

Improving Kubernetes scheduler logic #3653

Merged
merged 4 commits into from
Jan 4, 2021
Merged

Conversation

nicknezis
Copy link
Contributor

@nicknezis nicknezis commented Dec 30, 2020

This fixes #3652 . This Pull Request adds extra logic to handle Kubernetes delete requests for Topology StatefulSet and Service that do not exist in Kubernetes. This code should now gracefully handle the HTTP_NOT_FOUND response code.

@nicknezis nicknezis self-assigned this Dec 31, 2020
@nicknezis nicknezis marked this pull request as ready for review December 31, 2020 00:40
@nicknezis
Copy link
Contributor Author

I updated the logic to use a try-with-resource. I was noticing warnings about unclosed connections when it was throwing the ApiException. Hopefully this change will also help with that.

@joshfischer1108
Copy link
Member

Are there no unit tests that cover this code?

@nicknezis nicknezis merged commit 9af45c7 into master Jan 4, 2021
@nicknezis nicknezis deleted the nicknezis/k8s-delete-fix branch January 4, 2021 01:59
@joshfischer1108
Copy link
Member

@nicknezis are these changes covered by unit tests or integration tests?

@nicknezis
Copy link
Contributor Author

nicknezis commented Jan 4, 2021

The Kubernetes scheduler doesn't seem to have many unit tests. Do the other schedulers? Might be worth adding an issue to create some?

@joshfischer1108
Copy link
Member

joshfischer1108 commented Jan 4, 2021

I think creating issues would be a good start. I can add some as well if you let me know which ones I should write them for.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Kubernetes scheduler fails when Heron is asked to kill a non-existent topology
2 participants