#520 Abort tasks on non-success phase Pull request opened by hamersaw TL;DR FlytePropeller fast-fails when aborting tasks in a terminal phase. The result is that

Pods

marked as a retryable failure are never cleaned up. This PR only fast-fails when aborting tasks in a success phase, to ensure that all resources are correctly cleaned up at the risk of unnecessary API calls (ex. object does not exist, etc). Type ☑︎ Bug Fix ☐ Feature ☐ Plugin Are all requirements met? ☑︎ Code completed ☑︎ Smoke tested ☐ Unit tests added ☐ Code documentation added ☐ Any pending items have an associated Issue Complete description Currently, the

TaskHandler

fast-fails when aborting resources in a terminal phase. This causes issues where resources fail to cleanup when the task is reported as failed (ie. retryable failure or permanent failure) but the resource remains running. One such scenario is handling

Pods

in the "ImagePullBackoff" state. We mark these as failed, but because of the fast-fail during abort these k8s resources are never deleted. This may occur within other plugins as well. To be safe, we may want to attempt to abort all task phases other than

Success

. Tracking Issue fixes flyteorg/flyte#3239 Follow-up issue NA flyteorg/flytepropeller GitHub Actions: Build & Push Flytepropeller Image GitHub Actions: Goreleaser GitHub Actions: Bump Version ✅ 11 other checks have passed 11/14 successful checks