remove disruption pod budget from kubemacpool. (#170)

Currently kubemacpool works in a active-backup approach, where the waiting pod is in not-ready state. This can cause conflict when performing node drain, since we have kubemapool disruption pod budget set to minimum 1 ready pods. In order to fix this, we need to remove the disruption pod budget, and return it only after we implement active-active approach in the kubemacpool pods Signed-off-by: Ram Lavi <[email protected]> Co-authored-by: Ram Lavi <[email protected]>
k8snetworkplumbingwg · Jun 3, 2020 · ebd6d3d · ebd6d3d
1 parent 0e2ca51
commit ebd6d3d
Show file tree

Hide file tree

Showing 3 changed files with 0 additions and 32 deletions.
diff --git a/config/default/manager/manager.yaml b/config/default/manager/manager.yaml
@@ -118,16 +118,6 @@ spec:
             secretName: service
 
 ---
-apiVersion: policy/v1beta1
-kind: PodDisruptionBudget
-metadata:
-  name: mac-controller-manager
-spec:
-  minAvailable: 1
-  selector:
-    matchLabels:
-      control-plane: mac-controller-manager
----
 apiVersion: v1
 kind: Service
 metadata:

diff --git a/config/release/kubemacpool.yaml b/config/release/kubemacpool.yaml
@@ -320,14 +320,3 @@ spec:
       - name: tls-key-pair
         secret:
           secretName: kubemacpool-service
----
-apiVersion: policy/v1beta1
-kind: PodDisruptionBudget
-metadata:
-  name: kubemacpool-mac-controller-manager
-  namespace: kubemacpool-system
-spec:
-  minAvailable: 1
-  selector:
-    matchLabels:
-      control-plane: mac-controller-manager
diff --git a/config/test/kubemacpool.yaml b/config/test/kubemacpool.yaml
@@ -320,14 +320,3 @@ spec:
       - name: tls-key-pair
         secret:
           secretName: kubemacpool-service
----
-apiVersion: policy/v1beta1
-kind: PodDisruptionBudget
-metadata:
-  name: kubemacpool-mac-controller-manager
-  namespace: kubemacpool-system
-spec:
-  minAvailable: 1
-  selector:
-    matchLabels:
-      control-plane: mac-controller-manager