feat: e2e-aws using tf code

e2e-aws using TF code. Signed-off-by: Noel Georgi <git@frezbo.dev>
siderolabs · Aug 11, 2023 · 79ca1a3 · 79ca1a3
1 parent bf3a5e0
commit 79ca1a3
Show file tree

Hide file tree

Showing 8 changed files with 169 additions and 99 deletions.
diff --git a/.drone.jsonnet b/.drone.jsonnet
@@ -131,7 +131,7 @@ local Step(name, image='', target='', privileged=false, depends_on=[], environme
 
 // TriggerDownstream is a helper function for creating a step that triggers a
 // downstream pipeline. It is used to standardize the creation of these steps.
-local TriggerDownstream(name, target, repositories, image='', params=[], depends_on=[]) = {
+local TriggerDownstream(name, target, repositories, image='', params=[], depends_on=[], when={}) = {
   name: name,
   image: if image == '' then downstream_image else image,
   settings: {
@@ -146,6 +146,7 @@ local TriggerDownstream(name, target, repositories, image='', params=[], depends
     deploy: target,
   },
   depends_on: [x.name for x in depends_on],
+  when: when,
 };
 
 // This provides the docker service.
@@ -326,7 +327,7 @@ local load_artifacts = Step(
   extra_commands=[
     'az login --service-principal -u "$${AZURE_CLIENT_ID}" -p "$${AZURE_CLIENT_SECRET}" --tenant "$${AZURE_TENANT_ID}"',
     'mkdir -p _out/',
-    'az storage blob download-batch --overwrite true -d _out -s ${CI_COMMIT_SHA}${DRONE_TAG//./-}',
+    'az storage blob download-batch --overwrite -d _out -s ${CI_COMMIT_SHA}${DRONE_TAG//./-}',
     'chmod +x _out/clusterctl _out/integration-test-linux-amd64 _out/module-sig-verify-linux-amd64 _out/kubectl _out/kubestr _out/helm _out/cilium _out/talosctl*',
   ]
 );
@@ -340,7 +341,7 @@ local extensions_build = TriggerDownstream(
     std.format('REGISTRY=%s', local_registry),
     'PLATFORM=linux/amd64',
     'BUCKET_PATH=${CI_COMMIT_SHA}${DRONE_TAG//./-}',
-    '_out/talos-metadata',
+    '_out/talos-metadata',  // params passed from file with KEY=VALUE format
   ],
   depends_on=[load_artifacts],
 );
@@ -642,7 +643,65 @@ local capi_docker = Step('e2e-docker', depends_on=[load_artifacts], target='e2e-
   INTEGRATION_TEST_RUN: 'XXX',
 });
 local e2e_capi = Step('e2e-capi', depends_on=[capi_docker], environment=creds_env_vars);
-local e2e_aws = Step('e2e-aws', depends_on=[e2e_capi], environment=creds_env_vars);
+
+local e2e_aws_prepare = Step(
+  'cloud-images',
+  depends_on=[
+    load_artifacts,
+  ],
+  environment=creds_env_vars {
+    CLOUD_IMAGES_EXTRA_ARGS: '--name-prefix talos-e2e --target-clouds aws --architectures amd64 --aws-regions us-east-1',
+  },
+  extra_commands=[
+    'make e2e-aws-prepare',
+    'az login --service-principal -u "$${AZURE_CLIENT_ID}" -p "$${AZURE_CLIENT_SECRET}" --tenant "$${AZURE_TENANT_ID}"',
+    'az storage blob upload-batch --overwrite -s _out --pattern "e2e-aws-generated/*" -d "${CI_COMMIT_SHA}${DRONE_TAG//./-}"',
+  ]
+);
+
+local tf_apply = TriggerDownstream(
+  'tf-apply',
+  'e2e-talos-tf-apply',
+  ['siderolabs/contrib@main'],
+  params=[
+    'BUCKET_PATH=${CI_COMMIT_SHA}${DRONE_TAG//./-}',
+    'TYPE=aws',
+    'AWS_DEFAULT_REGION=us-east-1',
+  ],
+  depends_on=[e2e_aws_prepare],
+);
+
+local e2e_aws_tf_apply_post = Step(
+  'e2e-aws-download-artifacts',
+  with_make=false,
+  environment=creds_env_vars,
+  extra_commands=[
+    'az login --service-principal -u "$${AZURE_CLIENT_ID}" -p "$${AZURE_CLIENT_SECRET}" --tenant "$${AZURE_TENANT_ID}"',
+    'az storage blob download -f _out/e2e-aws-talosconfig -n e2e-aws-talosconfig -c ${CI_COMMIT_SHA}${DRONE_TAG//./-}',
+    'az storage blob download -f _out/e2e-aws-kubeconfig -n e2e-aws-kubeconfig -c ${CI_COMMIT_SHA}${DRONE_TAG//./-}',
+  ],
+  depends_on=[tf_apply],
+);
+
+local e2e_aws = Step('e2e-aws', depends_on=[e2e_aws_tf_apply_post], environment=creds_env_vars);
+
+local tf_destroy = TriggerDownstream(
+  'tf-destroy',
+  'e2e-talos-tf-destroy',
+  ['siderolabs/contrib@main'],
+  params=[
+    'TYPE=aws',
+    'AWS_DEFAULT_REGION=us-east-1',
+  ],
+  depends_on=[e2e_aws],
+  when={
+    status: [
+      'failure',
+      'success',
+    ],
+  },
+);
+
 local e2e_azure = Step('e2e-azure', depends_on=[e2e_capi], environment=creds_env_vars);
 local e2e_gcp = Step('e2e-gcp', depends_on=[e2e_capi], environment=creds_env_vars);
 
@@ -656,7 +715,7 @@ local e2e_trigger(names) = {
 
 local e2e_pipelines = [
   // regular pipelines, triggered on promote events
-  Pipeline('e2e-aws', default_pipeline_steps + [capi_docker, e2e_capi, e2e_aws]) + e2e_trigger(['e2e-aws']),
+  Pipeline('e2e-aws', default_pipeline_steps + [e2e_aws_prepare, tf_apply, e2e_aws_tf_apply_post, e2e_aws, tf_destroy]) + e2e_trigger(['e2e-aws']),
   Pipeline('e2e-gcp', default_pipeline_steps + [capi_docker, e2e_capi, e2e_gcp]) + e2e_trigger(['e2e-gcp']),
 
   // cron pipelines, triggered on schedule events

diff --git a/hack/cloud-image-uploader.sh b/hack/cloud-image-uploader.sh
@@ -4,4 +4,4 @@ set -e
 
 cd hack/cloud-image-uploader
 
-go run . --artifacts-path="../../${ARTIFACTS}" --tag="${TAG}" --abbrev-tag="${ABBREV_TAG}"
+go run . --artifacts-path="../../${ARTIFACTS}" --tag="${TAG}" --abbrev-tag="${ABBREV_TAG}" "$@"
diff --git a/hack/cloud-image-uploader/aws.go b/hack/cloud-image-uploader/aws.go
@@ -253,6 +253,10 @@ func (au *AWSUploader) registerAMIArch(ctx context.Context, region string, svc *
 
 	imageName := fmt.Sprintf("talos-%s-%s-%s", au.Options.Tag, region, arch)
 
+	if au.Options.NamePrefix != "" {
+		imageName = fmt.Sprintf("%s-%s-%s-%s", au.Options.NamePrefix, au.Options.Tag, region, arch)
+	}
+
 	imageResp, err := svc.DescribeImagesWithContext(ctx, &ec2.DescribeImagesInput{
 		Filters: []*ec2.Filter{
 			{

diff --git a/hack/cloud-image-uploader/main.go b/hack/cloud-image-uploader/main.go
@@ -61,10 +61,12 @@ func run() error {
 		log.Printf("failed to get a list of enabled AWS regions: %s, ignored", err)
 	}
 
+	pflag.StringSliceVar(&DefaultOptions.TargetClouds, "target-clouds", DefaultOptions.TargetClouds, "cloud targets to upload to")
 	pflag.StringSliceVar(&DefaultOptions.Architectures, "architectures", DefaultOptions.Architectures, "list of architectures to process")
 	pflag.StringVar(&DefaultOptions.ArtifactsPath, "artifacts-path", DefaultOptions.ArtifactsPath, "artifacts path")
 	pflag.StringVar(&DefaultOptions.Tag, "tag", DefaultOptions.Tag, "tag (version) of the uploaded image")
 	pflag.StringVar(&DefaultOptions.AzureAbbrevTag, "abbrev-tag", DefaultOptions.AzureAbbrevTag, "abbreviated tag (version) of the uploaded image")
+	pflag.StringVar(&DefaultOptions.NamePrefix, "name-prefix", DefaultOptions.NamePrefix, "prefix for the name of the uploaded image")
 
 	pflag.StringSliceVar(&DefaultOptions.AWSRegions, "aws-regions", DefaultOptions.AWSRegions, "list of AWS regions to upload to")
 
@@ -84,21 +86,28 @@ func run() error {
 
 	g, ctx = errgroup.WithContext(ctx)
 
-	g.Go(func() error {
-		aws := AWSUploader{
-			Options: DefaultOptions,
+	for _, target := range DefaultOptions.TargetClouds {
+		switch target {
+		case "aws":
+			g.Go(func() error {
+				aws := AWSUploader{
+					Options: DefaultOptions,
+				}
+
+				return aws.Upload(ctx)
+			})
+		case "azure":
+			g.Go(func() error {
+				azure := AzureUploader{
+					Options: DefaultOptions,
+				}
+
+				return azure.AzureGalleryUpload(ctx)
+			})
+		default:
+			return fmt.Errorf("unknown target: %s", target)
 		}
-
-		return aws.Upload(ctx)
-	})
-
-	g.Go(func() error {
-		azure := AzureUploader{
-			Options: DefaultOptions,
-		}
-
-		return azure.AzureGalleryUpload(ctx)
-	})
+	}
 
 	if err = g.Wait(); err != nil {
 		return fmt.Errorf("failed: %w", err)

diff --git a/hack/cloud-image-uploader/options.go b/hack/cloud-image-uploader/options.go
@@ -13,7 +13,9 @@ import (
 type Options struct {
 	Tag           string
 	ArtifactsPath string
+	NamePrefix    string
 	Architectures []string
+	TargetClouds  []string
 
 	// AWS options.
 	AWSRegions []string
@@ -36,6 +38,7 @@ type Location struct {
 var DefaultOptions = Options{
 	ArtifactsPath: "_out/",
 	Architectures: []string{"amd64", "arm64"},
+	TargetClouds:  []string{"aws", "azure"},
 }
 
 // AWSImage returns path to AWS pre-built image.

diff --git a/hack/test/e2e-aws-prepare.sh b/hack/test/e2e-aws-prepare.sh
@@ -0,0 +1,16 @@
+#!/usr/bin/env bash
+
+set -eou pipefail
+
+source ./hack/test/e2e.sh
+
+REGION="us-east-1"
+
+AMI_ID=$(jq -r ".[] | select(.region == \"${REGION}\") | select (.arch == \"amd64\") | .id" "${ARTIFACTS}/cloud-images.json")
+
+mkdir -p "${ARTIFACTS}/e2e-aws-generated"
+
+NAME_PREFIX="talos-e2e-${SHA}-aws"
+
+jq --null-input --arg AMI_ID "${AMI_ID}" --arg CLUSTER_NAME "${NAME_PREFIX}" --arg KUBERNETES_VERSION "${KUBERNETES_VERSION}" '{ami_id: $AMI_ID, cluster_name: $CLUSTER_NAME, kubernetes_version: $KUBERNETES_VERSION}' \
+  | jq -f hack/test/tfvars/aws.jq > "${ARTIFACTS}/e2e-aws-generated/vars.json"
diff --git a/hack/test/e2e-aws.sh b/hack/test/e2e-aws.sh
@@ -4,85 +4,49 @@ set -eou pipefail
 
 source ./hack/test/e2e.sh
 
-REGION="us-east-1"
-BUCKET="talos-ci-e2e"
+cp "${ARTIFACTS}/e2e-aws-talosconfig" "${TALOSCONFIG}"
+cp "${ARTIFACTS}/e2e-aws-kubeconfig" "${KUBECONFIG}"
+
+# Wait for nodes to check in
+timeout=$(($(date +%s) + TIMEOUT))
+until ${KUBECTL} get nodes -o go-template='{{ len .items }}' | grep ${NUM_NODES} >/dev/null; do
+  [[ $(date +%s) -gt $timeout ]] && exit 1
+  ${KUBECTL} get nodes -o wide && :
+  sleep 10
+done
+
+# Wait for nodes to be ready
+timeout=$(($(date +%s) + TIMEOUT))
+until ${KUBECTL} wait --timeout=1s --for=condition=ready=true --all nodes > /dev/null; do
+  [[ $(date +%s) -gt $timeout ]] && exit 1
+  ${KUBECTL} get nodes -o wide && :
+  sleep 10
+done
+
+# Verify that we have an HA controlplane
+timeout=$(($(date +%s) + TIMEOUT))
+until ${KUBECTL} get nodes -l node-role.kubernetes.io/control-plane='' -o go-template='{{ len .items }}' | grep 3 > /dev/null; do
+  [[ $(date +%s) -gt $timeout ]] && exit 1
+  ${KUBECTL} get nodes -l node-role.kubernetes.io/control-plane='' && :
+  sleep 10
+done
+
+CONTROLPLANE0_NODE_NAME=$(${KUBECTL} get nodes -l node-role.kubernetes.io/control-plane='' -o jsonpath='{.items[0].metadata.name}')
+
+# Wait until we have an IP for first controlplane node
+timeout=$(($(date +%s) + TIMEOUT))
+until [ -n "$(${KUBECTL} get nodes "${CONTROLPLANE0_NODE_NAME}" -o go-template --template='{{range .status.addresses}}{{if eq .type "ExternalIP"}}{{.address}}{{end}}{{end}}')" ]; do
+  [[ $(date +%s) -gt $timeout ]] && exit 1
+  sleep 10
+done
+
+
+# lets get the ip of the first controlplane node
+CONTROLPLANE0_NODE=$(${KUBECTL} get nodes "${CONTROLPLANE0_NODE_NAME}" -o go-template --template='{{range .status.addresses}}{{if eq .type "ExternalIP"}}{{.address}}{{end}}{{end}}')
+
+# set the talosconfig to use the first controlplane ip
+${TALOSCTL} config endpoint "${CONTROLPLANE0_NODE}"
+${TALOSCTL} config node "${CONTROLPLANE0_NODE}"
 
-function setup {
-  # Setup svc account
-  mkdir -p ${TMP}
-
-  # Uncompress image
-  xz -d < ${ARTIFACTS}/aws-amd64.raw.xz > ${TMP}/disk.raw
-
-  # Upload Image
-  echo "uploading image to s3"
-  aws s3 cp --quiet ${TMP}/disk.raw s3://${BUCKET}/aws-${TAG}.raw
-
-  # Create snapshot from image
-  echo "importing snapshot from s3"
-  import_task_id=$(aws ec2 import-snapshot --region ${REGION} --description "talos e2e ${TAG}" --disk-container "Format=raw,UserBucket={S3Bucket=${BUCKET},S3Key=aws-${TAG}.raw}" | jq -r '.ImportTaskId')
-  echo ${import_task_id}
-
-  # Wait for import to complete
-  echo "waiting for snapshot import to complete"
-  snapshot_status=$(aws ec2 describe-import-snapshot-tasks --region ${REGION} --import-task-ids ${import_task_id} | \
-                    jq -r --arg image_name "aws-${TAG}.raw" '.ImportSnapshotTasks[] | select(.SnapshotTaskDetail.UserBucket.S3Key == $image_name) | .SnapshotTaskDetail.Status')
-  while [ ${snapshot_status} != "completed" ]; do
-    sleep 5
-    snapshot_status=$(aws ec2 describe-import-snapshot-tasks --region ${REGION} --import-task-ids ${import_task_id} | \
-                      jq -r --arg image_name "aws-${TAG}.raw" '.ImportSnapshotTasks[] | select(.SnapshotTaskDetail.UserBucket.S3Key == $image_name) | .SnapshotTaskDetail.Status')
-  done
-  snapshot_id=$(aws ec2 describe-import-snapshot-tasks --region ${REGION} --import-task-ids ${import_task_id} | \
-                jq -r --arg image_name "aws-${TAG}.raw" '.ImportSnapshotTasks[] | select(.SnapshotTaskDetail.UserBucket.S3Key == $image_name) | .SnapshotTaskDetail.SnapshotId')
-  echo ${snapshot_id}
-
-  # Create AMI
-  image_id=$(aws ec2 describe-images --region ${REGION} --filters="Name=name,Values=talos-e2e-${TAG}" | jq -r '.Images[0].ImageId') || true
-
-  if [[ ${image_id} != "null" ]]; then
-    aws ec2 deregister-image --region ${REGION} --image-id ${image_id}
-  fi
-
-  ami=$(aws ec2 register-image --region ${REGION} \
-          --block-device-mappings "DeviceName=/dev/xvda,VirtualName=talostest,Ebs={DeleteOnTermination=true,SnapshotId=${snapshot_id},VolumeSize=20,VolumeType=gp2}" \
-          --root-device-name /dev/xvda \
-          --virtualization-type hvm \
-          --architecture x86_64 \
-          --ena-support \
-          --name talos-e2e-${TAG} | jq -r '.ImageId')
-
-  ## Cluster-wide vars
-  export CLUSTER_NAME=${NAME_PREFIX}
-  export AWS_REGION=us-east-1
-  export AWS_SSH_KEY_NAME=talos-e2e
-  export AWS_VPC_ID=vpc-ff5c5687
-  export AWS_SUBNET=subnet-c4e9b3a0
-  export AWS_SUBNET_AZ=us-east-1a
-  export CALICO_VERSION=v3.24.1
-  export AWS_CLOUD_PROVIDER_VERSION=v1.20.0-alpha.0
-
-  ## Control plane vars
-  export CONTROL_PLANE_MACHINE_COUNT=3
-  export AWS_CONTROL_PLANE_MACHINE_TYPE=t3.large
-  export AWS_CONTROL_PLANE_VOL_SIZE=50
-  export AWS_CONTROL_PLANE_AMI_ID=${ami}
-  export AWS_CONTROL_PLANE_ADDL_SEC_GROUPS='[{id: sg-ebe8e59f}]'
-  export AWS_CONTROL_PLANE_IAM_PROFILE=CAPI_AWS_ControlPlane
-
-  ## Worker vars
-  export WORKER_MACHINE_COUNT=3
-  export AWS_NODE_MACHINE_TYPE=t3.large
-  export AWS_NODE_VOL_SIZE=50
-  export AWS_NODE_AMI_ID=${ami}
-  export AWS_NODE_ADDL_SEC_GROUPS='[{id: sg-ebe8e59f}]'
-  export AWS_NODE_IAM_PROFILE=CAPI_AWS_Worker
-
-  ${CLUSTERCTL} generate cluster ${NAME_PREFIX} \
-    --kubeconfig /tmp/e2e/docker/kubeconfig \
-    --from https://github.com/siderolabs/cluster-api-templates/blob/main/aws/standard/standard.yaml > ${TMP}/cluster.yaml
-}
-
-setup
-create_cluster_capi aws
 run_talos_integration_test
 run_kubernetes_integration_test
diff --git a/hack/test/tfvars/aws.jq b/hack/test/tfvars/aws.jq
@@ -0,0 +1,15 @@
+{
+    "cluster_name": .cluster_name,
+    "num_control_planes": 3,
+    "num_workers": 3,
+    "ami_id": .ami_id,
+    "ccm": true,
+    "kubernetes_version": .kubernetes_version,
+    "instance_type_control_plane": "t3.large",
+    "instance_type_worker": "t3.large",
+    "extra_tags": {
+        "Name": .cluster_name,
+        "Project": "talos-e2e-ci",
+        "Environment": "ci"
+    }
+}