operator: add tolerations for rook-ceph operator deployment

This addresses rook#12660. As discussed in the issue when zone failure happens and if the operator running in that zone, then it takes the default toleration of 300 seconds. We want to override this default toleration seconds to 5 seconds so that operator gets redeployed immediately Signed-off-by: Mahesh Shetty <[email protected]>
travisn · Oct 5, 2023 · 5185371 · 5185371
1 parent 04c9f98
commit 5185371
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 0 deletions.
diff --git a/deploy/charts/rook-ceph/templates/deployment.yaml b/deploy/charts/rook-ceph/templates/deployment.yaml
@@ -26,6 +26,11 @@ spec:
 {{- if .Values.priorityClassName }}
       priorityClassName: {{ .Values.priorityClassName }}
 {{- end }}
+      tolerations:
+        - effect: NoExecute
+          key: node.kubernetes.io/unreachable
+          operator: Exists
+          tolerationSeconds: 5
       containers:
       - name: rook-ceph-operator
         image: "{{ .Values.image.repository }}:{{ .Values.image.tag }}"

diff --git a/deploy/examples/operator-openshift.yaml b/deploy/examples/operator-openshift.yaml
@@ -669,6 +669,11 @@ spec:
       labels:
         app: rook-ceph-operator
     spec:
+      tolerations:
+        - effect: NoExecute
+          key: node.kubernetes.io/unreachable
+          operator: Exists
+          tolerationSeconds: 5
       serviceAccountName: rook-ceph-system
       containers:
         - name: rook-ceph-operator

diff --git a/deploy/examples/operator.yaml b/deploy/examples/operator.yaml
@@ -595,6 +595,11 @@ spec:
       labels:
         app: rook-ceph-operator
     spec:
+      tolerations:
+        - effect: NoExecute
+          key: node.kubernetes.io/unreachable
+          operator: Exists
+          tolerationSeconds: 5
       serviceAccountName: rook-ceph-system
       containers:
         - name: rook-ceph-operator