Bump k8s version for test cluster and add rollback demo target (#115)

* Bump k8s version for test cluster and add rollback demo target

* Wait only if env var set

Makefile

@@ -2,7 +2,7 @@
 
 NAME ?= adobe/k8s-shredder
 K8S_SHREDDER_VERSION ?= "dev"
-KINDNODE_VERSION ?= "v1.25.11"
+KINDNODE_VERSION ?= "v1.28.0"
 COMMIT ?= $(shell git rev-parse --short HEAD)
 TEST_CLUSTERNAME ?= "k8s-shredder-test-cluster"
 
@@ -78,13 +78,19 @@ e2e-tests:  ## Run e2e tests for k8s-shredder deployed in a local kind cluster
 	@echo "Run e2e tests for k8s-shredder..."
 	@KUBECONFIG=${PWD}/kubeconfig go test internal/testing/e2e_test.go -v
 
+# DEMO targets
+# -----------
+.PHONY: demo.prep demo.run demo.rollback
 demo.prep: build ## Setup demo cluster
 	echo "Setup demo cluster..."
 	./internal/testing/local_env_prep.sh "${K8S_SHREDDER_VERSION}" "${KINDNODE_VERSION}" "${TEST_CLUSTERNAME}"
 
 demo.run: ## Run demo
 	./internal/testing/cluster_upgrade.sh "${TEST_CLUSTERNAME}"
 
+demo.rollback: ## Rollback demo
+	./internal/testing/rollback_cluster_upgrade.sh "${TEST_CLUSTERNAME}"
+
 
 ci: local-test e2e-tests clean ## Run CI
 

internal/testing/cluster_upgrade.sh

@@ -19,11 +19,16 @@ fi
 
 kubectl label node "${K8S_CLUSTER_NAME}-worker" --kubeconfig=kubeconfig --overwrite shredder.ethos.adobe.net/parked-node-expires-on="${EXPIRES_ON}"
 
-# For moving node back as active, useful during debug process
-#export K8S_CLUSTER_NAME=k8s-shredder-test-cluster
-#kubectl uncordon "${K8S_CLUSTER_NAME}-worker" --kubeconfig=kubeconfig
-#kubectl label node "${K8S_CLUSTER_NAME}-worker" --kubeconfig=kubeconfig shredder.ethos.adobe.net/upgrade-status-
-#kubectl label node "${K8S_CLUSTER_NAME}-worker" --kubeconfig=kubeconfig --overwrite shredder.ethos.adobe.net/parked-node-expires-on-
-#kubectl delete -n ns-k8s-shredder-test $(kubectl get pods -n ns-k8s-shredder-test -oname) --force --wait=0 --timeout=0
-#kubectl delete -n ns-team-k8s-shredder-test $(kubectl get pods -n ns-team-k8s-shredder-test -oname) --force --wait=0 --timeout=0
-#kubectl get po -A --field-selector=spec.nodeName=k8s-shredder-test-cluster-worker
+if [[ ${WAIT_FOR_PODS:-false} == "true" ]]
+then
+  while [[ $pod_status != "No resources found" ]]
+  do
+    echo "Info: Waiting for all pods to be evicted from the node..."
+    sleep 10
+    pod_status=$(kubectl get pods -A --field-selector metadata.namespace!=kube-system,metadata.namespace!=local-path-storage,spec.nodeName=k8s-shredder-test-cluster-worker 2>&1 >/dev/null)
+  done
+
+  # This is to simulate the upgrade process. We are going to wait for 1 minute and then uncordon the node.
+  kubectl label node "${K8S_CLUSTER_NAME}-worker" --kubeconfig=kubeconfig shredder.ethos.adobe.net/upgrade-status-
+  kubectl label node "${K8S_CLUSTER_NAME}-worker" --kubeconfig=kubeconfig --overwrite shredder.ethos.adobe.net/parked-node-expires-on-
+fi

internal/testing/rollback_cluster_upgrade.sh

@@ -0,0 +1,13 @@
+#!/usr/bin/env bash
+set -e
+
+K8S_CLUSTER_NAME=$1
+
+# For moving node back as active, useful during debug process
+export K8S_CLUSTER_NAME=k8s-shredder-test-cluster
+kubectl uncordon "${K8S_CLUSTER_NAME}-worker" --kubeconfig=kubeconfig
+kubectl label node "${K8S_CLUSTER_NAME}-worker" --kubeconfig=kubeconfig shredder.ethos.adobe.net/upgrade-status-
+kubectl label node "${K8S_CLUSTER_NAME}-worker" --kubeconfig=kubeconfig --overwrite shredder.ethos.adobe.net/parked-node-expires-on-
+kubectl delete -n ns-k8s-shredder-test $(kubectl get pods -n ns-k8s-shredder-test -oname) --force --wait=0 --timeout=0
+kubectl delete -n ns-team-k8s-shredder-test $(kubectl get pods -n ns-team-k8s-shredder-test -oname) --force --wait=0 --timeout=0
+kubectl get po -A --field-selector=spec.nodeName=k8s-shredder-test-cluster-worker