Sync with Longhorn manager

commit 8e060dce288c4dfa054a4b0a188559e624aeb3c8
Author: James Oliver <joliver@rancher.com>
Date:   Fri Aug 31 16:09:06 2018 -0700

    Add script to tear down Longhorn system

examples/deployment.yaml

@@ -30,6 +30,8 @@ metadata:
   labels:
     app: mysql
 spec:
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -47,6 +49,9 @@ spec:
         volumeMounts:
         - name: mysql-volume
           mountPath: /var/lib/mysql
+        env:
+        - name: MYSQL_ROOT_PASSWORD
+          value: "rancher"
       volumes:
       - name: mysql-volume
         persistentVolumeClaim:

scripts/cleanup.sh

@@ -0,0 +1,65 @@
+#!/bin/bash
+
+NAMESPACE=${NAMESPACE:-longhorn-system}
+
+remove_and_wait() {
+  local crd=$1
+  out=`kubectl -n ${NAMESPACE} delete $crd --all 2>&1`
+  if [ $? -ne 0 ]; then
+    echo $out
+    break
+  fi
+  while true; do
+    out=`kubectl -n ${NAMESPACE} get $crd -o yaml | grep 'items: \[\]'`
+    if [ $? -eq 0 ]; then
+      break
+    fi
+    echo $out
+    sleep 1
+  done
+  echo all $crd instances deleted
+}
+
+remove_crd_instances() {
+  remove_and_wait volumes.longhorn.rancher.io
+  # TODO: remove engines and replicas once we fix https://github.com/rancher/longhorn/issues/273
+  remove_and_wait engines.longhorn.rancher.io
+  remove_and_wait replicas.longhorn.rancher.io
+  remove_and_wait engineimages.longhorn.rancher.io
+  remove_and_wait settings.longhorn.rancher.io
+  # do this one last; manager crashes
+  remove_and_wait nodes.longhorn.rancher.io
+}
+
+# Delete driver related workloads in specific order
+remove_driver() {
+  kubectl -n ${NAMESPACE} delete daemonset.apps/longhorn-csi-plugin
+  kubectl -n ${NAMESPACE} delete statefulset.apps/csi-attacher
+  kubectl -n ${NAMESPACE} delete service/csi-attacher
+  kubectl -n ${NAMESPACE} delete statefulset.apps/csi-provisioner
+  kubectl -n ${NAMESPACE} delete service/csi-provisioner
+  kubectl -n ${NAMESPACE} delete daemonset.apps/longhorn-flexvolume-driver
+  kubectl -n ${NAMESPACE} delete deployment.apps/longhorn-driver-deployer
+}
+
+# Delete all workloads in the namespace
+remove_workloads() {
+  kubectl -n ${NAMESPACE} get daemonset.apps -o yaml | kubectl delete -f -
+  kubectl -n ${NAMESPACE} get deployment.apps -o yaml | kubectl delete -f -
+  kubectl -n ${NAMESPACE} get replicaset.apps -o yaml | kubectl delete -f -
+  kubectl -n ${NAMESPACE} get statefulset.apps -o yaml | kubectl delete -f -
+  kubectl -n ${NAMESPACE} get pods -o yaml | kubectl delete -f -
+  kubectl -n ${NAMESPACE} get service -o yaml | kubectl delete -f -
+}
+
+# Delete CRD definitions with longhorn.rancher.io in the name
+remove_crds() {
+  for crd in $(kubectl get crd -o jsonpath={.items[*].metadata.name} | tr ' ' '\n' | grep longhorn.rancher.io); do
+    kubectl delete crd/$crd
+  done
+}
+
+remove_crd_instances
+remove_driver
+remove_workloads
+remove_crds