bug: event leases are dropped on migration #4407

etcd --data-dir .etcd

./bin/sharded-test-server --number-of-shards=3 \
  --shard-feature-gates=LogicalClusterMigration=true \
  --external-etcd-servers=http://localhost:2379

export KUBEOCNFIG=.kcp/admin.kubeconfig
kubectl apply -f - <<EOF
apiVersion: apis.kcp.io/v1alpha2
kind: APIBinding
metadata:
  name: migration
spec:
  reference:
    export:
      path: root
      name: migration.kcp.io
EOF

kubectl create-workspace event-migration --location-selector name=shard-1 --enter
kubectl apply -f - <<EOF
apiVersion: v1
kind: Event
metadata:
  name: demo-event
  namespace: default
involvedObject:
  apiVersion: v1
  kind: ConfigMap
  name: kube-root-ca.crt
  namespace: default
reason: Demo
message: lease check
type: Normal
EOF

kubectl ws :root
lc="$(kubectl get workspace event-migration -o jsonpath='{.spec.cluster}')"
etcdctl get "/shard/shard-1/core/events/$lc/default/demo-event" -w json | grep -o '"lease":[0-9]*'

kubectl apply -f - <<EOF
apiVersion: migration.kcp.io/v1alpha1
kind: LogicalClusterMigration
metadata:
  name: event-migration
spec:
  logicalCluster: $lc
  destinationShard: shard-2
EOF
kubectl wait logicalclustermigration event-migration --for=jsonpath='{.status.phase}'=Completed

etcdctl get "/shard/shard-1/core/events/$lc/default/demo-event" -w json