From 2bf72779171adba3e1d2322962d35a2af96565d6 Mon Sep 17 00:00:00 2001
From: Nikholas Pcenicni <82239765+nikpcenicni@users.noreply.github.com>
Date: Wed, 13 May 2026 23:59:30 -0400
Subject: [PATCH] Enhance csi-snapshot-controller README with troubleshooting
 guidance for CrashLoopBackOff issues and update kustomization.yaml to include
 deployment patch. This improves user experience and deployment reliability.

---
 .../csi-snapshot-controller/README.md         | 14 ++++++++++++++
 .../controller/deployment-patch.yaml          | 19 +++++++++++++++++++
 .../controller/kustomization.yaml             |  3 +++
 3 files changed, 36 insertions(+)
 create mode 100644 clusters/noble/bootstrap/csi-snapshot-controller/controller/deployment-patch.yaml

diff --git a/clusters/noble/bootstrap/csi-snapshot-controller/README.md b/clusters/noble/bootstrap/csi-snapshot-controller/README.md
index 81a1b1c..60fd6cb 100644
--- a/clusters/noble/bootstrap/csi-snapshot-controller/README.md
+++ b/clusters/noble/bootstrap/csi-snapshot-controller/README.md
@@ -14,3 +14,17 @@ kubectl -n kube-system rollout status deploy/snapshot-controller --timeout=120s
 ```
 
 After this, create or label a **VolumeSnapshotClass** for Longhorn (`velero.io/csi-volumesnapshot-class: "true"`) per `clusters/noble/bootstrap/velero/README.md`.
+
+## Troubleshooting
+
+If **`snapshot-controller`** is **CrashLoopBackOff** with exit **255** / **1**, check previous logs:
+
+```bash
+kubectl -n kube-system logs deploy/snapshot-controller --previous --tail=80
+```
+
+Typical causes:
+
+1. **Volume Snapshot CRDs missing or wrong version** — re-apply: `kubectl apply -k clusters/noble/bootstrap/csi-snapshot-controller/crd`
+2. **Volume group snapshot APIs** — this repo installs **GA** `VolumeSnapshot*` CRDs only. The controller overlay sets **`--feature-gates=CSIVolumeGroupSnapshot=false`** so the binary does not expect **VolumeGroupSnapshot*** CRDs.
+3. **RBAC** — re-apply the controller kustomize (includes `rbac-snapshot-controller.yaml`).
diff --git a/clusters/noble/bootstrap/csi-snapshot-controller/controller/deployment-patch.yaml b/clusters/noble/bootstrap/csi-snapshot-controller/controller/deployment-patch.yaml
new file mode 100644
index 0000000..570c2e0
--- /dev/null
+++ b/clusters/noble/bootstrap/csi-snapshot-controller/controller/deployment-patch.yaml
@@ -0,0 +1,19 @@
+# Overlay on upstream setup-snapshot-controller.yaml (external-snapshotter v8.5.0).
+# - We only install GA VolumeSnapshot CRDs (no VolumeGroupSnapshot*). Explicitly disable the
+#   group-snapshot feature so the controller does not wait on / watch APIs we did not install.
+# - Longer CRD wait helps apiserver/etcd under load during bootstrap (avoids startup timeout).
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: snapshot-controller
+  namespace: kube-system
+spec:
+  template:
+    spec:
+      containers:
+        - name: snapshot-controller
+          args:
+            - "--v=5"
+            - "--leader-election=true"
+            - "--feature-gates=CSIVolumeGroupSnapshot=false"
+            - "--retry-crd-interval-max=5m"
diff --git a/clusters/noble/bootstrap/csi-snapshot-controller/controller/kustomization.yaml b/clusters/noble/bootstrap/csi-snapshot-controller/controller/kustomization.yaml
index 230d1b8..85dc98f 100644
--- a/clusters/noble/bootstrap/csi-snapshot-controller/controller/kustomization.yaml
+++ b/clusters/noble/bootstrap/csi-snapshot-controller/controller/kustomization.yaml
@@ -6,3 +6,6 @@ namespace: kube-system
 resources:
   - https://raw.githubusercontent.com/kubernetes-csi/external-snapshotter/v8.5.0/deploy/kubernetes/snapshot-controller/rbac-snapshot-controller.yaml
   - https://raw.githubusercontent.com/kubernetes-csi/external-snapshotter/v8.5.0/deploy/kubernetes/snapshot-controller/setup-snapshot-controller.yaml
+
+patches:
+  - path: deployment-patch.yaml