argo-cd/test/e2e/testdata/syncwaves-prune-order
Siddhesh Ghadi e5c88c914b
feat: Prune resources in reverse order of syncwave during sync (#15074) (#16748)
* Add e2e & doc for prune order during sync

Signed-off-by: Siddhesh Ghadi <sghadi1203@gmail.com>

* Point gitops-engine to fork with reverse prune changes

Signed-off-by: Siddhesh Ghadi <sghadi1203@gmail.com>

* Fix ci linting failures

Signed-off-by: Siddhesh Ghadi <sghadi1203@gmail.com>

* Update gitops-engine commit ref

Signed-off-by: Siddhesh Ghadi <sghadi1203@gmail.com>

---------

Signed-off-by: Siddhesh Ghadi <sghadi1203@gmail.com>
2024-01-30 10:08:59 -05:00
..
pod.yaml feat: Prune resources in reverse order of syncwave during sync (#15074) (#16748) 2024-01-30 10:08:59 -05:00
rbac.yaml feat: Prune resources in reverse order of syncwave during sync (#15074) (#16748) 2024-01-30 10:08:59 -05:00
README.md feat: Prune resources in reverse order of syncwave during sync (#15074) (#16748) 2024-01-30 10:08:59 -05:00

Test Scenario

This test example is for testing the reverse pruning of resources with syncwaves during sync operation.

Resource creation happens in below order

  • wave 0: sa & role
  • wave 1: rolebinding
  • wave 2: pod

They are setup in such a way that the resources will be cleaned up properly only if they are deleted in the reverse order of creation i.e

  • wave 0: pod
  • wave 1: rolebinding
  • wave 2: sa & role

If above delete order is not followed the pod gets stuck in terminating state due to a finalizer which is supposed to be removed by k8s container lifecycle hook on delete if delete order is correct.