Refactor poddisruptionbudget reconciler #118
Conversation
eguzki
changed the title
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## main #118 +/- ##
==========================================
- Coverage 68.62% 67.50% -1.12%
==========================================
Files 16 16
Lines 1157 1148 -9
==========================================
- Hits 794 775 -19
- Misses 324 329 +5
- Partials 39 44 +5
Flags with carried forward coverage won't be shown. Click here to find out more.
|
| if limitadorObj.Spec.PodDisruptionBudget == nil { | ||
| helpers.TagObjectToDelete(pdb) | ||
| return pdb | ||
| } |
There was a problem hiding this comment.
I have a question that raise a concern with me on this. If I understand PR will mean that we always create the PDB CR even if it is not being used. When it is not being used we mark it to be delete. Is it possible for us to get into an infinite loop here?
If we create a resource we should be reconciling any changes the user might make to that resource. By this logic if the resource is delete from the cluster we should check to see if that is the correct state, and with this change we would recreate the resource marking it for deletion causing the loop to start over. If we are not watching the resource for change causing use to not be affected by the infinite loop, then the question becomes how do we guarantee the resource is in the condition stated in the limitador CR?
There was a problem hiding this comment.
when we "tag" object to be deleted, the reconciler will make sure it is deleted and if it exists, it will delete it. https://github.com/Kuadrant/kuadrant-operator/blob/main/pkg/reconcilers/base_reconciler.go#L110-L127
Objects tagged as deleted will never be created
IMO the implementation is simpler doing exactly the same.