getStatus() open() read() …() Client Master Meta Cache First Access LRU listStatus() 2. Alluxio缓存行为控制 参数 取值 含义 alluxio.user.ufs.block.read.location.policy LocalFirstAvoidEvictionPolicy All x 0.99=99GiB触发驱逐 alluxio.user.block.avoid.eviction.policy.reserved.size .bytes 1056MB 当本地节点的空间少于1056MB时，数据缓存的调度器不会选择该 节点；转而选择其他节点。 alluxio.user.file.passive.cache.enabled false 当从Alluxio远程worker读 当从Alluxio远程worker读文件时，是否缓存文件到Alluxio的本 地worker。 alluxio.user.file.readtype.default CACHE 默认的CACHE_PROMOTE会带来显著的性能开销 策略：1.优先本地加载缓存 2.避免数据震荡 3.避免数据冗余 1 2 3 3. Fuse性能调优 • 选择更高版本的kernel • 设置max_read=131072 • 定制lib

Operator conveys readiness of application or managed resources to the user leveraging the status block of the Custom Resource Configuration of the workload • Operator provides configuration via the spec Operator 基础模型 第二部分 K8s 架构 Cache Informer 机制 Cache 如何获取到本地（内存中） Informer 启动后会通过 reflector 的 list & watch 机制获取某种资源的 全量 objects。list 可以简单理解为一个 HTTP GET 请求，watch 为一 个 HTTP/2 长连接 Cache 如何保持与 API Server 一致性 录 ResourceVersion 版本信息，watch 从 ResourceVersion 开始，获取后 续的增量数据。 watch 通过网络异步（asynchronous）获取增量数据，所以 cache 提供 的是最终一致性（eventual consistency）。 期间遇到网络、API Server 报错等异常时，会有重试机制 Controller-runtime 的 Informer

链路RT/QPS 服务异常 队列长度 gRPC监控 长连接分布 请求分布 限流 Authorization Authenticatio n 序列化 压缩 版本转换 Admission Cache Storage Filter Chain API 存储 Kube-APIServer Webhook ETCD 数据构建 压测场景 压测环境 压测报告 压测平台 监控&大盘• APIServer Watch优化 ETCD Cache Pod A V1 Pod A V2 Pod A V3 Reflector APIServer Watch Cache List & Watch Informer Reflector Store List & Watch• 网络抖动造成informer重新List & Watch List & Watch优化 Cache APIServer 13 Watch Cache Informer Store Kubelets Watch (rv=3 node=x) Too old version err rv=3 FIFO• 网络抖动造成informer重新List & Watch List & Watch优化 Cache APIServer 5 9 11 13 Watch Cache Informer

speed 7. Switch to use a Maven plugin Download and install Docker Order of layers to optimize for cache hits github.com/GoogleContainerTools/jib What did we do? 1. Write first Dockerfile 2. Reduce speed 7. Switch to use a Maven plugin Download and install Docker Order of layers to optimize for cache hits Use of multi-stage builds github.com/GoogleContainerTools/jib What did we do? 1. Write first plugin Download and install Docker Order of layers to optimize for cache hits Use of multi-stage builds Understanding Docker cache mechanism and quirks github.com/GoogleContainerTools/jib What did

: "qa" , "environment" : "production" "tier" : "frontend" , "tier" : "backend" , "tier" : "cache" "partition" : "customerA" , "partition" : "customerB" "track" : "daily" , "track" : "weekly" component: redis matchExpressions: 09-Label和Selector 28 - {key: tier, operator: In, values: [cache]} - {key: environment, operator: NotIn, values: [dev]} matchLabels 是 { key,value } 的映射。 apps/v1beta1 # for versions before 1.6.0 use extensions/v1beta1 kind: Deployment metadata: name: redis-cache spec: replicas: 3 template: metadata: labels: app: store spec:

基于Kubernetes的 CI/CD� 构建Docker镜像最佳实践 预期 ������ ������ ������ 实践 ����� ������ �������� ��Cache ��Volume ��yum/apk cache 用Helm管理Kubernetes应用 • K8s服务编排唯一开源子项目，K8s包管理工具 • 用于部署复杂K8s应用，处理复杂的服务间依赖 • 查看发布历史与某一次发布的具体配置

Logic A DB Server Business Logic A Business Logic A File Server Business Logic A Business Logic A Cache 需要多少 Services, Deployments, StatefulSets DaemonSets, CronJobs? 容器编排的困境 Kubernetes 为我们带来了什么？ 1 无状态应用 有状态应用 守护型应用 批处理任务 应用编排架构 应用编排架构 API Gateway APP API Service A APP API Service B APP Cache APP API Service D APP File Server APP 应用编排架构 应用编排架构 基本架构 Kubernetes Cluster Application Registry

scheduler policy Build tasks and the dependent environments（sidecar） Share files between containers, or cache build files Container Image - Image of build / dependent environment [] Command - glisters pvc pvc pvc registry credential using secret - resources Memory / CPU / GPU Data cache CI/CD Examples - Artifact Management user scripts using ConfigMap Job - pod template • upload

Swapping? This basically comes down to a choice of whether you would rather have a fast cache, or a slower cache that is larger. Many popular application runtimes (e.g. Java jre) have similar NUMA related

implications. This basically comes down to a choice of whether you would rather have a fast cache, or a slower cache that is larger. Many popular application runtimes (e.g. Java jre) have similar NUMA related