ZEDEDA WiKi

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Deleting Longhorn ======

===== Step 1 — Identify Longhorn Controllers =====

<code bash>
# See exactly which controllers own those pods
kubectl get pod -n default longhorn-csi-plugin-95cws -o wide
kubectl get pod -n default longhorn-csi-plugin-9drl8 -o wide
kubectl get pod -n default longhorn-csi-plugin-w4rvw -o wide

# List DaemonSets / Deployments to find Longhorn controllers
kubectl get ds,deploy -n default | grep -i longhorn
</code>

===== Step 2 — Delete Controllers =====

<code bash>
# If they are DaemonSets
kubectl delete ds longhorn-csi-plugin -n default --ignore-not-found

# If you see any of these deployments, delete them
kubectl delete deploy longhorn-manager longhorn-ui longhorn-driver-deployer \
  -n default --ignore-not-found

# Remove one-shot jobs
kubectl delete job longhorn-pre-upgrade longhorn-post-upgrade -n default --ignore-not-found
</code>

===== Step 3 — Confirm No Longhorn Pods Remain =====

<code bash>
kubectl get pods -A | grep -i longhorn || echo "No Longhorn pods"
</code>

===== Step 4 — Delete All CRD Instances =====

<code bash>
for crd in $(kubectl get crd -o jsonpath={.items[*].metadata.name} | tr ' ' '\n' | grep longhorn.io); do
  echo "Cleaning CRD: $crd"
  kubectl delete "$crd" --all --wait=false || true
  kubectl delete crd "$crd" --wait=false || true
done
</code>

Confirm CRDs are gone:

<code bash>
kubectl get crd | grep longhorn.io || echo "No Longhorn CRDs"
</code>

===== Step 5 — Remove Longhorn StorageClass =====

<code bash>
kubectl get storageclass | grep -i longhorn || echo "No Longhorn StorageClass"

# If something shows up, delete it:
kubectl delete storageclass longhorn --ignore-not-found
</code>

===== Step 6 — Remove Remaining Longhorn CRD Objects =====

<code bash>
# Check what remains
kubectl get backuptargets.longhorn.io --all-namespaces || echo "no backuptargets"
kubectl get engineimages.longhorn.io --all-namespaces  || echo "no engineimages"
kubectl get nodes.longhorn.io --all-namespaces         || echo "no nodes"

# Delete CSI deployments
kubectl delete deploy csi-attacher    -n default --ignore-not-found
kubectl delete deploy csi-provisioner -n default --ignore-not-found
kubectl delete deploy csi-resizer     -n default --ignore-not-found
kubectl delete deploy csi-snapshotter -n default --ignore-not-found

# Delete engine image DaemonSet
kubectl delete ds engine-image-ei-ff1cedad -n default --ignore-not-found

# Delete remaining CRD objects
kubectl delete backuptargets.longhorn.io --all -n default
kubectl delete engineimages.longhorn.io --all -n default
kubectl delete nodes.longhorn.io --all -n default
</code>

===== Step 7 — Remove Webhooks (if stuck) =====

<code bash>
kubectl delete ValidatingWebhookConfiguration longhorn-webhook-validator --ignore-not-found
kubectl delete MutatingWebhookConfiguration   longhorn-webhook-mutator   --ignore-not-found
</code>

===== Final Verification =====

<code bash>
kubectl get pods -A       | grep -i longhorn   || echo "No Longhorn pods"
kubectl get crd           | grep longhorn.io   || echo "No Longhorn CRDs"
kubectl get storageclass  | grep -i longhorn   || echo "No Longhorn StorageClass"
</code>