在学习 MapReduce 的过程中，不少人接触的第一个项目就是单词计数。单词计数通过两个函数 Map 和 Reduce，可以快速地统计出文本文件中每个单词出现的个数，它虽然简单，但也是最能体现 MapReduce 思想的程序之一。而 Serverless 的出现，为 MapReduce 进行大数据处理又提供了一个新的部署方案，Serverless 与 MapReduce 究竟如何结合呢？
\n本文将通过一个简单的教程，指导大家快速开发一个基于 MapReduce 的 WordCount 应用模版，并在 Serverless 应用中心 Registry 里实现复用。
\n已安装 Serverless Framework，并保证您的 Serverless Framework 不低于以下版本：
\n$ serverless –v\nFramework Core: 1.74.1 (standalone)\nPlugin: 3.6.14\nSDK: 2.3.1\nComponents: 2.31.6下面是该实例的实现流程：
\ncos:ObjectCreated:* 事件发布给函数。部署成功后，本模版将会为您创建以下资源：
\n# serverless.yml\norg: serverless\napp: MapReduce_Demo\nstage: dev\ncomponent: cos\nname: destmr\n\ninputs:\n bucket: destmr\n region: ap-guangzhou同理，完成其它两个存储桶配置。
\nMap 函数 yml 文件示例如下：
\ncomponent: scf # (必选) 组件名称，在该实例中为scf\nname: map_function # 必选) 组件实例名称.\norg: serverless # (可选) 用于记录组织信息，\napp: MapReduce_Demo # (可选) 用于您的 AP名称\nstage: dev # (可选) 用于区分环境信息，默认值是 dev\n\ninputs:\n name: map_function\n src: ./\n handler: map_function.main_handler \n runtime: Python2.7 \n region: ap-guangzhou \n description: This is one of the MapReduce function which is map_function\n memorySize: 128 \n timeout: 10\n environment:\n variables:\n Bucket: ${output:${stage}:${app}:middlestagebucket.bucket}\n TENCENT_SECRET_ID: ${env:TENCENT_SECRET_ID}\n TENCENT_SECRET_KEY: ${env:TENCENT_SECRET_KEY}\n \n events: # 触发器\n - cos: # cos触发器\n name: ${output:${stage}:${app}:srcmr.cosOrigin}\n parameters:\n bucket: ${output:${stage}:${app}:srcmr.cosOrigin}\n events: 'cos:ObjectCreated:*'\n enable: trueMap_Reduce_Demo\n|--bucket_destmr\n |--serverless.yml\n|--bucket_middlestage\n |--serverless.yml\n|--bucket_srcmr\n |--serverless.yml\n|--fun_map\n |--serverless.yml\n |--map_function.py\n|--fun_reduce\n |--serverless.yml\n |--reduce_function.py您也可根据您的实际业务逻辑进行更改。
\n完成模版开发后，您可以将您的模版上传至 Registry，供大家公开复用。
\n# serverless.yml\nname: mapreduce-demo # 项目模板的名字\ndisplayName: 基于 MapReduce 统计字数 #项目模板展示在控制台的名称（中文）\nauthor: Tencent Cloud, Inc. # 作者的名字\norg: Tencent Cloud, Inc. # 组织名称，可选\ntype: template #项目类型，可填 template 或 component，此处为模版\ndescription: Deploy a MapReduce wordcount application. # 描述您的项目模板\ndescription-i18n:\n zh-cn: 本示例Demo演示怎么利用COS来做MapReduce，一共需要2个函数：map_function和reduce_function，3个COS Bucket：srcmr、middlestagebucket 和 destmr # 中文描述\nkeywords: tencent, serverless, cos, scf, mapreduce # 关键字\nrepo: # 源代码 Repo\nreadme: # 详细的说明文件\nlicense: MIT # 版权声明\nsrc: # 描述项目中的哪些文件需要作为模板发布\n src: ./ # 指定具体的相对目录，此目录下的文件将作为模板发布\n exclude: #描述在指定的目录内哪些文件应该被排除\n - .env\n - serverless.yml$ sls registry publish\nserverless ⚡ registry\nPublishing "mapreduce-demo@0.0.0"...\n\nServerless › Successfully published mapreduce-demo所有上传到 Registry的模版都支持公开下载与复用的，操作如下：
\n$ sls init -t mapreduce-demo\nserverless ⚡ framework\n\n- Successfully created "mapreduce-demo" instance in the currennt working directory.\n- Don't forget to update serverless.yml and install dependencies if needed.\n- Whenever you're ready, run "serverless deploy" to deploy your new instance.\n\nmapreduce-demo › Created# .env\nTENCENT_SECRET_ID=123\nTENCENT_SECRET_KEY=123$ cd mapreduce-demo\n$ sls deploy --all\nserverless ⚡ framework\n\nsrcmr: \n region: ap-guangzhou\n bucket: srcmr-0000000000\n cosOrigin: srcmr-0000000000.cos.ap-guangzhou.myqcloud.com\n url: http://srcmr-0000000000.cos.ap-guangzhou.myqcloud.com\n vendorMessage: null\n\ndestmr: \n region: ap-guangzhou\n bucket: destmr-0000000000\n cosOrigin: destmr-0000000000.cos.ap-guangzhou.myqcloud.com\n url: http://destmr-0000000000.cos.ap-guangzhou.myqcloud.com\n vendorMessage: null\n\nmiddlestagebucket: \n region: ap-guangzhou\n bucket: middlestagebucket-0000000000\n cosOrigin: middlestagebucket-0000000000.cos.ap-guangzhou.myqcloud.com\n url: http://middlestagebucket-0000000000.cos.ap-guangzhou.myqcloud.com\n vendorMessage: null\n\nmap_function: \n functionName: map_function\n description: This is one of the MapReduce function which is map_function\n namespace: default\n runtime: Python2.7\n handler: map_function.main_handler\n memorySize: 128\n lastVersion: $LATEST\n traffic: 1\n triggers: \n cos: \n - srcmr-0000000000.cos.ap-guangzhou.myqcloud.com\n vendorMessage: null\n\nreduce-function: \n functionName: reduce_function\n description: This is one of the MapReduce function which is reduce_function\n namespace: default\n runtime: Python2.7\n handler: reduce_function.main_handler\n memorySize: 128\n lastVersion: $LATEST\n traffic: 1\n triggers: \n cos: \n - middlestagebucket-0000000000.cos.ap-guangzhou.myqcloud.com\n vendorMessage: null\n\n8s › mapreduce-demo › Success可以通过以下命令移除应用
\n$ sls remove --all\n\nserverless ⚡ framework\n\n8s › maprecude › Success\n \n\n传送门：
\n\n
\n- GitHub: github.com/serverless
\n- 官网：serverless.com
\n
欢迎访问：Serverless 中文网，您可以在 最佳实践 里体验更多关于 Serverless 应用的开发！
","tableOfContents":""},"previousBlog":{"id":"40b4ebd6-7c3e-5418-9203-3222d32b7b90","frontmatter":{"thumbnail":"https://img.serverlesscloud.cn/2020731/1596192531300-1596186208253-%E4%BC%81%E4%B8%9A%E5%BE%AE%E4%BF%A1%E6%88%AA%E5%9B%BE_15961861933543.jpg","authors":["粟俊娥"],"categories":["best-practice"],"date":"2020-07-31T00:00:00.000Z","title":"腾讯云 Serverless CI/CD 自动化部署实战","description":"本文将为大家讲解 Serverless 工作原理、架构优势和 Serverless 应用的开发流程，以及如何使用 Serverless CI/CD 能力进行自动化部署。","authorslink":["https://github.com/June1991"],"translators":null,"translatorslink":null,"tags":["CI/CD","Meetup"],"keywords":"Serverless, ServerlessDays, Serverless DevOps","outdated":null},"wordCount":{"words":246,"sentences":47,"paragraphs":47},"fileAbsolutePath":"/opt/build/repo/content/best-practice/2020-07-31-serverless-cicd.md","fields":{"slug":"/best-practice/2020-07-31-serverless-cicd/","keywords":["serverless","无服务器","云函数","Serverless","开发","serverless","部署","video"]}},"nextBlog":{"id":"a95f844d-eae9-5465-b1e8-d5dcd4464453","frontmatter":{"thumbnail":"https://img.serverlesscloud.cn/2020713/1594610687799-Nextjs-93.jpg","authors":["Yugasun"],"categories":["best-practice"],"date":"2020-07-12T00:00:00.000Z","title":"如何优雅地部署一个 Serverless Next.js 应用","description":"本篇专门针对 Next.js 的 SSR 方案进行了探索和优化，一步一步带大家了解，如何基于 Serverless 架构部署一个实际的线上业务","authorslink":["https://github.com/yugasun"],"translators":null,"translatorslink":null,"tags":["Serverless SSR","Next.js"],"keywords":"Serverless SSR,Serverless Egg.js","outdated":null},"wordCount":{"words":478,"sentences":81,"paragraphs":81},"fileAbsolutePath":"/opt/build/repo/content/best-practice/2020-07-12-serverless-nextjs.md","fields":{"slug":"/best-practice/2020-07-12-serverless-nextjs/","keywords":["nextjs","serverless","ssr","云函数","serverless","Next","部署","nextjs","静态","Serverless","Layer"]}}},"pageContext":{"isCreatedByStatefulCreatePages":false,"blogId":"ca884fd4-891f-599b-ac9f-b6f1169df802","previousBlogId":"40b4ebd6-7c3e-5418-9203-3222d32b7b90","nextBlogId":"a95f844d-eae9-5465-b1e8-d5dcd4464453"}}}