"openservice.name" . }} helm.sh/chart: {{ include "openservice.chart" . }} app.kubernetes.io/instance: {{ .Release.Name }} app.kubernetes.io/managed-by: {{ .Release.Service }} data: config ReportStorageMode = redis AttrStorageMode = mongodb TiDB = influxdb [MQTT] Broker = tcp://emqx-headless:1883 ClientID = server0 UserName = dashboard Secret = iSurpass2018 "openservice.name" . }} helm.sh/chart: {{ include "openservice.chart" . }} app.kubernetes.io/instance: {{ .Release.Name }} app.kubernetes.io/managed-by: {{ .Release.Service }} spec: serviceName:

termination Deployments Manages the release rollout of new versions of Pods Singletons A single instance of a Pod which is not replicated or scaled 19 Developers Admin Kubernetes & PKS $kubectl create Services API Cluster3 NSX-T vSphere PKS • Includes • PKS Controller, NSX-T • CFCR, Harbor, Broker • Deploys & Configures - CFCR - vSphere - NSX-T Integration - Harbor • Manages Cluster Day 2 K8s-2 Worker Worker Kubernetes Cluster Services API K8s-3 Worker Worker Worker GCP Service Broker 25 API PKS Controller #pks create-cluster K8s-1 n=3 #pks create-cluster K8s-2 n=3 #pks create-cluster

为什什么做基于 Kubernetes 的 Kafka 平台 基于 Kubernetes 的 Kafka 平台实践 Apache Kafka 分布式的流式数据平台 ⾼高吞吐 容错性 Broker Broker Broker Producer Consumer Topic Topic Topic Producer Consumer 平台承载知乎业务⽇日志、数据传输和消息队列列服务 平台线上稳定运⾏行行 的集群细分 • Topic 服务等级、容量量和规模划分 资源规划 多变需求引发集群规模增⻓长 • Broker, Topic 规模 服务器器资源利利⽤用率 • 单机运⾏行行多 Broker ⽅方式 多集群问题 磁盘是不不得不不考虑的问题 • ⽇日志落盘，⽇日志失效 ⽅方案 • Broker 之间物理理磁盘隔离 磁盘是个⼤大问题 服务器器选型 ⾼高密度存储服务器器 • 多磁盘, 单盘 单盘 RAID • 服务器器使⽤用率⾼高 如何管理理集群 集群数增加，导致 Broker 扩张 • 如何调度它们 • 如何管理理它们 服务器器如何管理理 Kubernetes 集群资源管理理和调度 容器器技术提供资源隔离 应⽤用程序管理理 Kafka on Kubernetes 设计 Kafka 容器器 • 内存、CPU、⽹网络和存储 调度 Kafka 容器器 内存

NameSpace NameSpace NameSpace Kubernetes NameSpace NameSpace NameSpace NameSpace VMware PKS Broker UAA VMware PKS API BOSH CredHub ©2019 VMware, Inc. 15 PKS已經做好11大領域超過百項的安全強化措施 PKS對

kubelet ECI: Elastic Container Instance • Run Containers without Managing Infrastructure • 容器成为云上的一等公民 • 安全隔离的容器运行环境 • 支持CPU 0.25c – 64c，GPU，按需创建按秒收费 • Spot Instance：极大降低计算成本 • Startup time: ~10s 7 layer • Knative serving on ASK：automatic scaling in knative • 集成ARMS, SLS Elastic Container Instance ECI ECI ECI ECI ECI ECI ECI ECI ECI ECI ECI ECI ECI ECI ECI ASK Cluster Serverless Kubernetes resource change ECI Two-way sync of resources K8S resources CRUD K8S Client Elastic Container Instance Pod Viking agent Container Container Pod Get Pod status Service/Ingress DNS Entry SLB Private

Aggregation Architecture ETCD API Server Kubernetes Core Extend API Server Service Instance-1 Instance-2 Instance-3 Expose service & registry API ETCD/ Other Storage © 2017 Rancher Labs, Inc .

difference Machine and Node state, manage Node softwares and configure files. • Machine: the instance of Machine CRD with node basic information, which represent a node desired in the Kubernetes

and Helm • Spin up pods for each variation of hyperparameters • One centralized TensorBoard instance • Autoscaling will create / remove VMs as needed to save cost Demo: Create End to End ML Pipelines

• Scaling based on business growth is very easy. Scalable • In the case of traffic spike, for instance, we can spin up new containers much more quickly than setting up new VMs. Faster MTTR Project

https://kubernetes.slack.com/messages/sig-vmware Or simply join #sig-vmware on the Kubernetes Slack instance SIG VMware serves to bring together members of the VMware and Kubernetes community to maintain