tests/e2e: Add a full lifecycle e2e test and restructure

- new full lifecycle test also tests stop/start functionality
- cleaned up a bit

tests/common.sh

@@ -31,7 +31,7 @@ abort() {
 }
 
 command_exists() {
-    command -v $1 >/dev/null 2>&1
+    command -v "$1" >/dev/null 2>&1
 }
 
 failed() {
@@ -57,23 +57,41 @@ check_url() {
   curl -L --silent -k --output /dev/null --fail "$1"
 }
 
-check_k3d_clusters() {
+# check_clusters verifies that given clusters are reachable
+check_clusters() {
   [ -n "$EXE" ] || abort "EXE is not defined"
   for c in "$@" ; do
-    $EXE --verbose get kubeconfig "$c" --switch
-    #[ -n "$kc" ] || abort "could not obtain a kubeconfig for $c"
+    $EXE get kubeconfig "$c" --switch
     if kubectl cluster-info ; then
       passed "cluster $c is reachable"
     else
       warn "could not obtain cluster info for $c. Kubeconfig:\n$(kubectl config view)"
+      docker ps -a
       return 1
     fi
   done
   return 0
 }
 
+# check_multi_node verifies that a cluster runs with an expected number of nodes
+check_multi_node() {
+  cluster=$1
+  expectedNodeCount=$2
+  $EXE get kubeconfig "$cluster" --switch
+  nodeCount=$(kubectl get nodes -o=custom-columns=NAME:.metadata.name --no-headers | wc -l)
+  if [[ $nodeCount == $expectedNodeCount ]]; then
+    passed "cluster $cluster has $expectedNodeCount nodes, as expected"
+  else
+    warn "cluster $cluster has incorrect number of nodes: $nodeCount != $expectedNodeCount"
+    kubectl get nodes -o=custom-columns=NAME:.metadata.name --no-headers
+    docker ps -a
+    return 1
+  fi
+  return 0
+}
+
 check_registry() {
-  check_url $REGISTRY/v2/_catalog
+  check_url "$REGISTRY/v2/_catalog"
 }
 
 check_volume_exists() {

tests/dind.sh

@@ -15,6 +15,7 @@ k3de2e=$(docker run -d \
           --privileged \
           -e EXE="$K3D_EXE" \
           -e CI="true" \
+          -e LOG_LEVEL="$LOG_LEVEL" \
           --name "k3d-e2e-runner-$TIMESTAMP" \
           k3d:$K3D_IMAGE_TAG)
 

tests/test_basic.sh

@@ -7,15 +7,15 @@ CURR_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
 source "$CURR_DIR/common.sh"
 
 info "Creating two clusters..."
-$EXE --verbose create cluster c1 --wait --timeout 60s --api-port 6443 || failed "could not create cluster c1"
-$EXE --verbose create cluster c2 --wait --timeout 60s --api-port 6444 || failed "could not create cluster c2"
+$EXE create cluster c1 --wait --timeout 60s --api-port 6443 || failed "could not create cluster c1"
+$EXE create cluster c2 --wait --timeout 60s --api-port 6444 || failed "could not create cluster c2"
 
 info "Checking we have access to both clusters..."
-check_k3d_clusters "c1" "c2" || failed "error checking cluster"
+check_clusters "c1" "c2" || failed "error checking cluster"
 
 info "Deleting clusters..."
-$EXE --verbose delete cluster c1 || failed "could not delete the cluster c1"
-$EXE --verbose delete cluster c2 || failed "could not delete the cluster c2"
+$EXE delete cluster c1 || failed "could not delete the cluster c1"
+$EXE delete cluster c2 || failed "could not delete the cluster c2"
 
 exit 0
 

tests/test_full_lifecycle.sh

@@ -0,0 +1,51 @@
+#!/bin/bash
+
+CURR_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
+[ -d "$CURR_DIR" ] || { echo "FATAL: no current dir (maybe running in zsh?)";  exit 1; }
+
+# shellcheck source=./common.sh
+source "$CURR_DIR/common.sh"
+
+clustername="lifecycletest"
+
+info "Creating cluster $clustername..."
+$EXE create cluster "$clustername" --workers 1 --api-port 6443 --wait --timeout 360s || failed "could not create cluster $clustername"
+
+info "Sleeping for 5 seconds to give the cluster enough time to get ready..."
+sleep 5
+
+# 1. check initial access to the cluster
+info "Checking that we have access to the cluster..."
+check_clusters "$clustername" || failed "error checking cluster"
+
+info "Checking that we have 2 nodes online..."
+check_multi_node "$clustername" 2 || failed "failed to verify number of nodes"
+
+# 2. stop the cluster
+info "Stopping cluster..."
+$EXE stop cluster "$clustername"
+
+info "Checking that cluster was stopped"
+check_clusters "$clustername" && failed "cluster was not stopped, since we still have access"
+
+# 3. start the cluster
+info "Starting cluster..."
+$EXE start cluster "$clustername"
+
+info "Sleeping for 5 seconds to give the cluster enough time to get ready..."
+sleep 5
+
+info "Checking that we have access to the cluster..."
+check_clusters "$clustername" || failed "error checking cluster"
+
+info "Checking that we have 2 nodes online..."
+check_multi_node "$clustername" 2 || failed "failed to verify number of nodes"
+
+# Cleanup
+
+info "Deleting cluster $clustername..."
+$EXE delete cluster "$clustername" || failed "could not delete the cluster $clustername"
+
+exit 0
+
+

tests/test_multi_master.sh

@@ -7,30 +7,16 @@ CURR_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
 source "$CURR_DIR/common.sh"
 
 info "Creating cluster multimaster..."
-$EXE --verbose create cluster "multimaster" --masters 3 --api-port 6443 --wait --timeout 360s || failed "could not create cluster multimaster"
+$EXE create cluster "multimaster" --masters 3 --api-port 6443 --wait --timeout 360s || failed "could not create cluster multimaster"
 
 info "Checking that we have access to the cluster..."
-check_k3d_clusters "multimaster" || failed "error checking cluster"
+check_clusters "multimaster" || failed "error checking cluster"
 
 info "Sleeping for 5 seconds to give the cluster enough time to get ready..."
 sleep 5
 
 info "Checking that we have 3 master nodes online..."
-check_multi_master() {
-  for c in "$@" ; do
-    $EXE get kubeconfig "$c" --switch
-    nodeCount=$(kubectl get nodes -o=custom-columns=NAME:.metadata.name --no-headers | wc -l)
-    if [[ $nodeCount == 3 ]]; then
-      passed "cluster $c has 3 nodes, as expected"
-    else
-      warn "cluster $c has incorrect number of nodes: $nodeCount != 3"
-      kubectl get nodes -o=custom-columns=NAME:.metadata.name --no-headers
-      return 1
-    fi
-  done
-  return 0
-}
-check_multi_master "multimaster"
+check_multi_node "multimaster" 3 || failed "failed to verify number of nodes"
 
 info "Deleting cluster multimaster..."
 $EXE delete cluster "multimaster" || failed "could not delete the cluster multimaster"