• 镜像
    • 开始之前
    • 创建一个默认路由策略
    • 镜像流量到 v2
    • 清理
    • 相关内容

    镜像

    此任务演示了 Istio 的流量镜像功能。

    流量镜像，也称为影子流量，是一个以尽可能低的风险为生产带来变化的强大的功能。镜像会将实时流量的副本发送到镜像服务。镜像流量发生在主服务的关键请求路径之外。

    在此任务中，首先把流量全部路由到 v1 版本的测试服务。然后，执行规则将一部分流量镜像到 v2 版本。

    开始之前

    • 按照安装指南中的说明设置 Istio。

    • 首先部署两个版本的 httpbin 服务，httpbin 服务已开启访问日志：

    httpbin-v1:

    1. $ cat <<EOF | istioctl kube-inject -f - | kubectl create -f -
    2. apiVersion: apps/v1
    3. kind: Deployment
    4. metadata:
    5.   name: httpbin-v1
    6. spec:
    7.   replicas: 1
    8.   template:
    9.     metadata:
    10.       labels:
    11.         app: httpbin
    12.         version: v1
    13.     spec:
    14.       containers:
    15.       - image: docker.io/kennethreitz/httpbin
    16.         imagePullPolicy: IfNotPresent
    17.         name: httpbin
    18.         command: ["gunicorn", "--access-logfile", "-", "-b", "0.0.0.0:80", "httpbin:app"]
    19.         ports:
    20.         - containerPort: 80
    21. EOF

    httpbin-v2:

    1. $ cat <<EOF | istioctl kube-inject -f - | kubectl create -f -
    2. apiVersion: apps/v1
    3. kind: Deployment
    4. metadata:
    5.   name: httpbin-v2
    6. spec:
    7.   replicas: 1
    8.   template:
    9.     metadata:
    10.       labels:
    11.         app: httpbin
    12.         version: v2
    13.     spec:
    14.       containers:
    15.       - image: docker.io/kennethreitz/httpbin
    16.         imagePullPolicy: IfNotPresent
    17.         name: httpbin
    18.         command: ["gunicorn", "--access-logfile", "-", "-b", "0.0.0.0:80", "httpbin:app"]
    19.         ports:
    20.         - containerPort: 80
    21. EOF

    httpbin Kubernetes service:

    1. $ kubectl create -f - <<EOF
    2. apiVersion: v1
    3. kind: Service
    4. metadata:
    5.   name: httpbin
    6.   labels:
    7.     app: httpbin
    8. spec:
    9.   ports:
    10.   - name: http
    11.     port: 8000
    12.     targetPort: 80
    13.   selector:
    14.     app: httpbin
    15. EOF
    • 启动 sleep 服务，这样就可以使用 curl 来提供负载了：

    sleep service:

    1. $ cat <<EOF | istioctl kube-inject -f - | kubectl create -f -
    2. apiVersion: apps/v1
    3. kind: Deployment
    4. metadata:
    5.   name: sleep
    6. spec:
    7.   replicas: 1
    8.   template:
    9.     metadata:
    10.       labels:
    11.         app: sleep
    12.     spec:
    13.       containers:
    14.       - name: sleep
    15.         image: tutum/curl
    16.         command: ["/bin/sleep","infinity"]
    17.         imagePullPolicy: IfNotPresent
    18. EOF

    创建一个默认路由策略

    默认情况下，Kubernetes 在 httpbin 服务的两个版本之间进行负载均衡。在此步骤中会更改该行为，把所有流量都路由到 v1。

    • 创建一个默认路由规则，将所有流量路由到服务的 v1：

    如果安装/配置 Istio 的时候开启了 TLS 认证，在应用 DestinationRule 之前必须将 TLS 流量策略 mode: ISTIO_MUTUAL 添加到 DestinationRule。否则，请求将发生 503 错误，如设置目标规则后出现 503 错误所述。

    1. $ kubectl apply -f - <<EOF
    2. apiVersion: networking.istio.io/v1alpha3
    3. kind: VirtualService
    4. metadata:
    5.   name: httpbin
    6. spec:
    7.   hosts:
    8.     - httpbin
    9.   http:
    10.   - route:
    11.     - destination:
    12.         host: httpbin
    13.         subset: v1
    14.       weight: 100
    15. ---
    16. apiVersion: networking.istio.io/v1alpha3
    17. kind: DestinationRule
    18. metadata:
    19.   name: httpbin
    20. spec:
    21.   host: httpbin
    22.   subsets:
    23.   - name: v1
    24.     labels:
    25.       version: v1
    26.   - name: v2
    27.     labels:
    28.       version: v2
    29. EOF

    现在所有流量都转到httpbin:v1服务。

    • 向服务发送一下流量：
    1. $ export SLEEP_POD=$(kubectl get pod -l app=sleep -o jsonpath={.items..metadata.name})
    2. $ kubectl exec -it $SLEEP_POD -c sleep -- sh -c 'curl  http://httpbin:8000/headers' | python -m json.tool
    3. {
    4.   "headers": {
    5.     "Accept": "*/*",
    6.     "Content-Length": "0",
    7.     "Host": "httpbin:8000",
    8.     "User-Agent": "curl/7.35.0",
    9.     "X-B3-Sampled": "1",
    10.     "X-B3-Spanid": "eca3d7ed8f2e6a0a",
    11.     "X-B3-Traceid": "eca3d7ed8f2e6a0a",
    12.     "X-Ot-Span-Context": "eca3d7ed8f2e6a0a;eca3d7ed8f2e6a0a;0000000000000000"
    13.   }
    14. }
    • 分别查看 httpbin 服务 v1 和 v2 两个 pods 的日志，您可以看到访问日志进入 v1，而 v2 中没有日志，显示为 <none>：
    1. $ export V1_POD=$(kubectl get pod -l app=httpbin,version=v1 -o jsonpath={.items..metadata.name})
    2. $ kubectl logs -f $V1_POD -c httpbin
    3. 127.0.0.1 - - [07/Mar/2018:19:02:43 +0000] "GET /headers HTTP/1.1" 200 321 "-" "curl/7.35.0"
    1. $ export V2_POD=$(kubectl get pod -l app=httpbin,version=v2 -o jsonpath={.items..metadata.name})
    2. $ kubectl logs -f $V2_POD -c httpbin
    3. <none>

    镜像流量到 v2

    • 改变流量规则将流量镜像到 v2：
    1. $ kubectl apply -f - <<EOF
    2. apiVersion: networking.istio.io/v1alpha3
    3. kind: VirtualService
    4. metadata:
    5.   name: httpbin
    6. spec:
    7.   hosts:
    8.     - httpbin
    9.   http:
    10.   - route:
    11.     - destination:
    12.         host: httpbin
    13.         subset: v1
    14.       weight: 100
    15.     mirror:
    16.       host: httpbin
    17.       subset: v2
    18.     mirror_percent: 100
    19. EOF

    这个路由规则发送 100% 流量到 v1。最后一段表示你将镜像流量到 httpbin:v2 服务。当流量被镜像时，请求将发送到镜像服务中，并在 headers 中的 Host/Authority 属性值上追加 -shadow。例如 cluster-1 变为 cluster-1-shadow。

    此外，重点注意这些被镜像的流量是『即发即弃』的，就是说镜像请求的响应会被丢弃。

    您可以使用 mirror_percent 属性来设置镜像流量的百分比，而不是镜像全部请求。为了兼容老版本，如果这个属性不存在，将镜像所有流量。

    • 发送流量：
    1. $ kubectl exec -it $SLEEP_POD -c sleep -- sh -c 'curl  http://httpbin:8000/headers' | python -m json.tool

    现在就可以看到 v1 和 v2 中都有了访问日志。v2 中的访问日志就是由镜像流量产生的，这些请求的实际目标是 v1。

    1. $ kubectl logs -f $V1_POD -c httpbin
    2. 127.0.0.1 - - [07/Mar/2018:19:02:43 +0000] "GET /headers HTTP/1.1" 200 321 "-" "curl/7.35.0"
    3. 127.0.0.1 - - [07/Mar/2018:19:26:44 +0000] "GET /headers HTTP/1.1" 200 321 "-" "curl/7.35.0"
    1. $ kubectl logs -f $V2_POD -c httpbin
    2. 127.0.0.1 - - [07/Mar/2018:19:26:44 +0000] "GET /headers HTTP/1.1" 200 361 "-" "curl/7.35.0"
    • 如果要检查流量内部，请在另一个控制台上运行以下命令：
    1. $ export SLEEP_POD=$(kubectl get pod -l app=sleep -o jsonpath={.items..metadata.name})
    2. $ export V1_POD_IP=$(kubectl get pod -l app=httpbin -l version=v1 -o jsonpath={.items..status.podIP})
    3. $ export V2_POD_IP=$(kubectl get pod -l app=httpbin -l version=v2 -o jsonpath={.items..status.podIP})
    4. $ kubectl exec -it $SLEEP_POD -c istio-proxy -- sudo tcpdump -A -s 0 host $V1_POD_IP or host $V2_POD_IP
    5. tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
    6. listening on eth0, link-type EN10MB (Ethernet), capture size 262144 bytes
    7. 05:47:50.159513 IP sleep-7b9f8bfcd-2djx5.38836 > 10-233-75-11.httpbin.default.svc.cluster.local.80: Flags [P.], seq 4039989036:4039989832, ack 3139734980, win 254, options [nop,nop,TS val 77427918 ecr 76730809], length 796: HTTP: GET /headers HTTP/1.1
    8. E..P2.X.X.X.
    9. .K.
    10. .K....P..W,.$.......+.....
    11. ..t.....GET /headers HTTP/1.1
    12. host: httpbin:8000
    13. user-agent: curl/7.35.0
    14. accept: */*
    15. x-forwarded-proto: http
    16. x-request-id: 571c0fd6-98d4-4c93-af79-6a2fe2945847
    17. x-envoy-decorator-operation: httpbin.default.svc.cluster.local:8000/*
    18. x-b3-traceid: 82f3e0a76dcebca2
    19. x-b3-spanid: 82f3e0a76dcebca2
    20. x-b3-sampled: 0
    21. x-istio-attributes: Cj8KGGRlc3RpbmF0aW9uLnNlcnZpY2UuaG9zdBIjEiFodHRwYmluLmRlZmF1bHQuc3ZjLmNsdXN0ZXIubG9jYWwKPQoXZGVzdGluYXRpb24uc2VydmljZS51aWQSIhIgaXN0aW86Ly9kZWZhdWx0L3NlcnZpY2VzL2h0dHBiaW4KKgodZGVzdGluYXRpb24uc2VydmljZS5uYW1lc3BhY2USCRIHZGVmYXVsdAolChhkZXN0aW5hdGlvbi5zZXJ2aWNlLm5hbWUSCRIHaHR0cGJpbgo6Cgpzb3VyY2UudWlkEiwSKmt1YmVybmV0ZXM6Ly9zbGVlcC03YjlmOGJmY2QtMmRqeDUuZGVmYXVsdAo6ChNkZXN0aW5hdGlvbi5zZXJ2aWNlEiMSIWh0dHBiaW4uZGVmYXVsdC5zdmMuY2x1c3Rlci5sb2NhbA==
    22. content-length: 0
    25. 05:47:50.159609 IP sleep-7b9f8bfcd-2djx5.49560 > 10-233-71-7.httpbin.default.svc.cluster.local.80: Flags [P.], seq 296287713:296288571, ack 4029574162, win 254, options [nop,nop,TS val 77427918 ecr 76732809], length 858: HTTP: GET /headers HTTP/1.1
    26. E.....X.X...
    27. .K.
    28. .G....P......l......e.....
    29. ..t.....GET /headers HTTP/1.1
    30. host: httpbin-shadow:8000
    31. user-agent: curl/7.35.0
    32. accept: */*
    33. x-forwarded-proto: http
    34. x-request-id: 571c0fd6-98d4-4c93-af79-6a2fe2945847
    35. x-envoy-decorator-operation: httpbin.default.svc.cluster.local:8000/*
    36. x-b3-traceid: 82f3e0a76dcebca2
    37. x-b3-spanid: 82f3e0a76dcebca2
    38. x-b3-sampled: 0
    39. x-istio-attributes: Cj8KGGRlc3RpbmF0aW9uLnNlcnZpY2UuaG9zdBIjEiFodHRwYmluLmRlZmF1bHQuc3ZjLmNsdXN0ZXIubG9jYWwKPQoXZGVzdGluYXRpb24uc2VydmljZS51aWQSIhIgaXN0aW86Ly9kZWZhdWx0L3NlcnZpY2VzL2h0dHBiaW4KKgodZGVzdGluYXRpb24uc2VydmljZS5uYW1lc3BhY2USCRIHZGVmYXVsdAolChhkZXN0aW5hdGlvbi5zZXJ2aWNlLm5hbWUSCRIHaHR0cGJpbgo6Cgpzb3VyY2UudWlkEiwSKmt1YmVybmV0ZXM6Ly9zbGVlcC03YjlmOGJmY2QtMmRqeDUuZGVmYXVsdAo6ChNkZXN0aW5hdGlvbi5zZXJ2aWNlEiMSIWh0dHBiaW4uZGVmYXVsdC5zdmMuY2x1c3Rlci5sb2NhbA==
    40. x-envoy-internal: true
    41. x-forwarded-for: 10.233.75.12
    42. content-length: 0
    45. 05:47:50.166734 IP 10-233-75-11.httpbin.default.svc.cluster.local.80 > sleep-7b9f8bfcd-2djx5.38836: Flags [P.], seq 1:472, ack 796, win 276, options [nop,nop,TS val 77427925 ecr 77427918], length 471: HTTP: HTTP/1.1 200 OK
    46. E....3X.?...
    47. .K.
    48. .K..P...$....ZH...........
    49. ..t...t.HTTP/1.1 200 OK
    50. server: envoy
    51. date: Fri, 15 Feb 2019 05:47:50 GMT
    52. content-type: application/json
    53. content-length: 241
    54. access-control-allow-origin: *
    55. access-control-allow-credentials: true
    56. x-envoy-upstream-service-time: 3
    58. {
    59.   "headers": {
    60.     "Accept": "*/*",
    61.     "Content-Length": "0",
    62.     "Host": "httpbin:8000",
    63.     "User-Agent": "curl/7.35.0",
    64.     "X-B3-Sampled": "0",
    65.     "X-B3-Spanid": "82f3e0a76dcebca2",
    66.     "X-B3-Traceid": "82f3e0a76dcebca2"
    67.   }
    68. }
    70. 05:47:50.166789 IP sleep-7b9f8bfcd-2djx5.38836 > 10-233-75-11.httpbin.default.svc.cluster.local.80: Flags [.], ack 472, win 262, options [nop,nop,TS val 77427925 ecr 77427925], length 0
    71. E..42.X.X.\.
    72. .K.
    73. .K....P..ZH.$.............
    74. ..t...t.
    75. 05:47:50.167234 IP 10-233-71-7.httpbin.default.svc.cluster.local.80 > sleep-7b9f8bfcd-2djx5.49560: Flags [P.], seq 1:512, ack 858, win 280, options [nop,nop,TS val 77429926 ecr 77427918], length 511: HTTP: HTTP/1.1 200 OK
    76. E..3..X.>...
    77. .G.
    78. .K..P....l....;...........
    79. ..|...t.HTTP/1.1 200 OK
    80. server: envoy
    81. date: Fri, 15 Feb 2019 05:47:49 GMT
    82. content-type: application/json
    83. content-length: 281
    84. access-control-allow-origin: *
    85. access-control-allow-credentials: true
    86. x-envoy-upstream-service-time: 3
    88. {
    89.   "headers": {
    90.     "Accept": "*/*",
    91.     "Content-Length": "0",
    92.     "Host": "httpbin-shadow:8000",
    93.     "User-Agent": "curl/7.35.0",
    94.     "X-B3-Sampled": "0",
    95.     "X-B3-Spanid": "82f3e0a76dcebca2",
    96.     "X-B3-Traceid": "82f3e0a76dcebca2",
    97.     "X-Envoy-Internal": "true"
    98.   }
    99. }
    101. 05:47:50.167253 IP sleep-7b9f8bfcd-2djx5.49560 > 10-233-71-7.httpbin.default.svc.cluster.local.80: Flags [.], ack 512, win 262, options [nop,nop,TS val 77427926 ecr 77429926], length 0
    102. E..4..X.X...
    103. .K.
    104. .G....P...;..n............
    105. ..t...|.

    您可以看到流量​​的请求和响应内容。

    清理

    • 删除规则：
    1. $ kubectl delete virtualservice httpbin
    2. $ kubectl delete destinationrule httpbin
    • 关闭 httpbin 服务和客户端：
    1. $ kubectl delete deploy httpbin-v1 httpbin-v2 sleep
    2. $ kubectl delete svc httpbin

    相关内容

    Traffic Mirroring with Istio for Testing in Production

    An introduction to safer, lower-risk deployments and release to production.

    Istio as a Proxy for External Services

    Configure Istio ingress gateway to act as a proxy for external services.

    Multi-mesh deployments for isolation and boundary protection

    Deploy environments that require isolation into separate meshes and enable inter-mesh communication by mesh federation.

    Secure Control of Egress Traffic in Istio, part 3

    Comparison of alternative solutions to control egress traffic including performance considerations.

    Secure Control of Egress Traffic in Istio, part 2

    Use Istio Egress Traffic Control to prevent attacks involving egress traffic.

    Secure Control of Egress Traffic in Istio, part 1

    Attacks involving egress traffic and requirements for egress traffic control.