{"id":6467,"date":"2022-04-08T21:04:10","date_gmt":"2022-04-08T13:04:10","guid":{"rendered":"https:\/\/egonlin.com\/?p=6467"},"modified":"2022-04-08T21:04:40","modified_gmt":"2022-04-08T13:04:40","slug":"spark%e8%bf%90%e8%a1%8c%e6%9e%b6%e6%9e%84","status":"publish","type":"post","link":"https:\/\/egonlin.com\/?p=6467","title":{"rendered":"Spark\u8fd0\u884c\u67b6\u6784"},"content":{"rendered":"<h1>\u4e00 Spark\u4e0ehadoop<\/h1>\n<p>Hadoop\u6709\u4e24\u4e2a\u6838\u5fc3\u6a21\u5757\uff0c\u5206\u5e03\u5f0f\u5b58\u50a8\u6a21\u5757HDFS\u548c\u5206\u5e03\u5f0f\u8ba1\u7b97\u6a21\u5757Mapreduce<\/p>\n<p>Spark \u652f\u6301\u591a\u79cd\u7f16\u7a0b\u8bed\u8a00\uff0c\u5305\u62ec Java\u3001Python\u3001R \u548c Scala\uff0c\u540c\u65f6 Spark \u4e5f\u652f\u6301 Hadoop \u7684\u5e95\u5c42\u5b58\u50a8\u7cfb\u7edf HDFS\uff0c\u4f46 Spark \u4e0d\u4f9d\u8d56 Hadoop\u3002<\/p>\n<p>Hadoop\u7684Mapreduce\u4e0espark\u90fd\u53ef\u4ee5\u8fdb\u884c\u6570\u636e\u8ba1\u7b97\uff0c\u800c\u76f8\u6bd4\u4e8eMapreduce\uff0cspark\u7684\u901f\u5ea6\u66f4\u5feb\u5e76\u4e14\u63d0\u4f9b\u7684\u529f\u80fd\u66f4\u52a0\u4e30\u5bcc<\/p>\n<p>\u5173\u7cfb\u56fe\u5982\u4e0b\uff1a<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67841.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67841.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>Hadoop\u3001Hive\u3001Spark \u4e4b\u95f4\u662f\u4ec0\u4e48\u5173\u7cfb\u3001\u4ec0\u4e48\u662fYarn<\/p>\n<h1>\u4e8c spark\u751f\u6001\u7ec4\u6210<\/h1>\n<p>\u901a\u5e38\u5f53\u9700\u8981\u5904\u7406\u7684\u6570\u636e\u91cf\u8d85\u8fc7\u4e86\u5355\u673a\u5c3a\u5ea6(\u6bd4\u5982\u6211\u4eec\u7684\u8ba1\u7b97\u673a\u67094GB\u7684\u5185\u5b58\uff0c\u800c\u6211\u4eec\u9700\u8981\u5904\u7406100GB\u4ee5\u4e0a\u7684\u6570\u636e)\u8fd9\u65f6\u6211\u4eec\u53ef\u4ee5\u9009\u62e9spark\u96c6\u7fa4\u8fdb\u884c\u8ba1\u7b97\uff0c\u6709\u65f6\u6211\u4eec\u53ef\u80fd\u9700\u8981\u5904\u7406\u7684\u6570\u636e\u91cf\u5e76\u4e0d\u5927\uff0c\u4f46\u662f\u8ba1\u7b97\u5f88\u590d\u6742\uff0c\u9700\u8981\u5927\u91cf\u7684\u65f6\u95f4\uff0c\u8fd9\u65f6\u6211\u4eec\u4e5f\u53ef\u4ee5\u9009\u62e9\u5229\u7528spark\u96c6\u7fa4\u5f3a\u5927\u7684\u8ba1\u7b97\u8d44\u6e90\uff0c\u5e76\u884c\u5316\u5730\u8ba1\u7b97\uff0c<\/p>\n<p>Spark \u9664\u4e86 Spark Core \u5916\uff0c\u8fd8\u6709\u5176\u5b83\u7531\u591a\u4e2a\u7ec4\u4ef6\u7ec4\u6210\uff0c\u76ee\u524d\u4e3b\u8981\u6709\u56db\u4e2a\u7ec4\u4ef6\uff1aSpark SQL\u3001Spark Streaming\u3001MLlib\u3001GraphX\u3002\u8fd9\u56db\u4e2a\u7ec4\u4ef6\u52a0\u4e0a Spark Core \u7ec4\u6210\u4e86 Spark \u7684\u751f\u6001\u3002\u901a\u5e38\uff0c\u6211\u4eec\u5728\u7f16\u5199\u4e00\u4e2a Spark \u5e94\u7528\u7a0b\u5e8f\uff0c\u9700\u8981\u7528\u5230 Spark<br \/>\nCore \u548c\u5176\u4f59 4 \u4e2a\u7ec4\u4ef6\u4e2d\u7684\u81f3\u5c11\u4e00\u4e2a\u3002Spark \u7684\u6574\u4f53\u6784\u67b6\u56fe\u5982\u4e0b\u56fe\u6240\u793a\uff1a<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67842.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67842.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>Spark SQL\uff1a\u63d0\u4f9b\u901a\u8fc7Apache Hive\u7684SQL\u53d8\u4f53Hive\u67e5\u8be2\u8bed\u8a00\uff08HiveQL\uff09\u4e0eSpark\u8fdb\u884c\u4ea4\u4e92\u7684API\u3002<strong>\u6bcf\u4e2a\u6570\u636e\u5e93\u8868\u88ab\u5f53\u505a\u4e00\u4e2aRDD(\u5f39\u6027\u5206\u5e03\u5f0f\u6570\u636e\u96c6resilient distributed dataset)<\/strong>\uff0cSpark SQL\u67e5\u8be2\u88ab\u8f6c\u6362\u4e3aSpark\u64cd\u4f5c\u3002<\/p>\n<p>Spark Streaming\uff1a\u8fd9\u4e2a\u6a21\u5757\u4e3b\u8981\u662f\u5bf9\u6d41\u6570\u636e\u7684\u5904\u7406\uff0c\u652f\u6301\u6d41\u6570\u636e\u7684\u53ef\u4f38\u7f29\u548c\u5bb9\u9519\u5904\u7406\uff0cSpark Streaming\u5141\u8bb8\u7a0b\u5e8f\u80fd\u591f\u50cf\u666e\u901aRDD\u4e00\u6837\u5904\u7406\u5b9e\u65f6\u6570\u636e,\u53ef\u4ee5\u4e0e Flume\uff08\u9488\u5bf9\u6570\u636e\u65e5\u5fd7\u8fdb\u884c\u4f18\u5316\u7684\u4e00\u4e2a\u7cfb\u7edf\uff09\u548c Kafka\uff08\u9488\u5bf9\u5206\u5e03\u5f0f\u6d88\u606f\u4f20\u9012\u8fdb\u884c\u4f18\u5316\u7684\u6d41\u5904\u7406\u5e73\u53f0\uff09\u7b49\u5df2\u5efa\u7acb\u7684\u6570\u636e\u6e90\u96c6\u6210\u3002Spark Streaming \u7684\u5b9e\u73b0\uff0c\u4e5f\u4f7f\u7528 RDD \u62bd\u8c61\u7684\u6982\u5ff5\uff0c\u4f7f\u5f97\u5728\u4e3a\u6d41\u6570\u636e\uff08\u5982\u6279\u91cf\u5386\u53f2\u65e5\u5fd7\u6570\u636e\uff09\u7f16\u5199\u5e94\u7528\u7a0b\u5e8f\u65f6\uff0c\u80fd\u591f\u66f4\u7075\u6d3b\uff0c\u4e5f\u66f4\u5bb9\u6613\u5b9e\u73b0\u3002<\/p>\n<p>MLlib\uff1a\u4e3b\u8981\u7528\u4e8e\u673a\u5668\u5b66\u4e60\u9886\u57df\uff0c\u5b83\u5b9e\u73b0\u4e86\u4e00\u7cfb\u5217\u5e38\u7528\u7684\u673a\u5668\u5b66\u4e60\u548c\u7edf\u8ba1\u7b97\u6cd5\uff0c\u5982\u5206\u7c7b\u3001\u56de\u5f52\u3001\u805a\u7c7b\u3001\u4e3b\u6210\u5206\u5206\u6790\u7b49\u7b97\u6cd5\u3002<\/p>\n<p>GraphX\uff1a\u8fd9\u4e2a\u6a21\u5757\u4e3b\u8981\u652f\u6301\u6570\u636e\u56fe\u7684\u5206\u6790\u548c\u8ba1\u7b97\uff0c\u5e76\u652f\u6301\u56fe\u5f62\u5904\u7406\u7684 Pregel API \u7248\u672c\u3002GraphX \u5305\u542b\u4e86\u8bb8\u591a\u88ab\u5e7f\u6cdb\u7406\u89e3\u7684\u56fe\u5f62\u7b97\u6cd5\uff0c\u5982 PageRank\u3002<\/p>\n<p>Spark Core\uff1a\u662f Spark \u7684\u6838\u5fc3\uff0c\u4e3b\u8981\u8d1f\u8d23\u4efb\u52a1\u8c03\u5ea6\u7b49\u7ba1\u7406\u529f\u80fd\u3002Spark<br \/>\nCore \u7684\u5b9e\u73b0\u4f9d\u8d56\u4e8e RDDs\uff08Resilient Distributed Datasets,<br \/>\n\u5f39\u6027\u5206\u5e03\u5f0f\u6570\u636e\u96c6\uff09\u7684\u7a0b\u5e8f\u62bd\u8c61\u6982\u5ff5\u3002<\/p>\n<p>Spark \u5e95\u5c42\u8fd8\u652f\u6301\u591a\u79cd\u6570\u636e\u6e90\uff0c\u80fd\u591f\u4ece\u5176\u5b83\u6587\u4ef6\u7cfb\u7edf\u8bfb\u53d6\u6570\u636e\uff0c\u5982 HDFS\u3001Amazon S3\u3001Hypertable\u3001HBase \u7b49\u3002Spark \u5bf9\u8fd9\u4e9b\u6587\u4ef6\u7cfb\u7edf\u7684\u652f\u6301\uff0c\u540c\u65f6\u4e5f\u4e30\u5bcc\u4e86\u6574\u4e2a Spark \u751f\u6001\u7684\u8fd0\u884c\u73af\u5883\u3002<\/p>\n<p>spark\u67b6\u6784\u7b80\u56fe<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67843.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67843.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>Cluster Manager\uff1a\u5728standalone\u6a21\u5f0f\u4e2d\u5373\u4e3aMaster\u4e3b\u8282\u70b9\uff0c\u63a7\u5236\u6574\u4e2a\u96c6\u7fa4\uff0c\u76d1\u63a7worker\u3002\u5728YARN\u6a21\u5f0f\u4e2d\u4e3a\u8d44\u6e90\u7ba1\u7406\u5668<\/p>\n<p>Worker\u8282\u70b9\uff1a\u4ece\u8282\u70b9\uff0c\u8d1f\u8d23\u63a7\u5236\u8ba1\u7b97\u8282\u70b9\uff0c\u542f\u52a8Executor\u6216\u8005Driver\u3002<\/p>\n<p>Driver\uff1a \u8fd0\u884cApplication \u7684main()\u51fd\u6570<\/p>\n<p>Executor\uff1a\u6267\u884c\u5668\uff0c\u662f\u4e3a\u67d0\u4e2aApplication\u8fd0\u884c\u5728worker node\u4e0a\u7684\u4e00\u4e2a\u8fdb\u7a0b<\/p>\n<h1>\u4e09 Spark\u7684\u67b6\u6784\u8be6\u89e3<\/h1>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67844.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67844.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<h2>1\u3001Application\uff1a<\/h2>\n<p>Spark Application\u7684\u6982\u5ff5\u548cHadoop MapReduce\u4e2d\u7684\u7c7b\u4f3c\uff0c\u6307\u7684\u662f\u7528\u6237\u7f16\u5199\u7684Spark\u5e94\u7528\u7a0b\u5e8f\uff0c\u5305\u542b\u4e86\u4e00\u4e2aDriver \u529f\u80fd\u7684\u4ee3\u7801\u548c\u5206\u5e03\u5728\u96c6\u7fa4\u4e2d\u591a\u4e2a\u8282\u70b9\u4e0a\u8fd0\u884c\u7684Executor\u4ee3\u7801\uff1b<\/p>\n<h2>2\u3001Driver\uff1a<\/h2>\n<p>\u8d1f\u8d23\u8fd0\u884c\u4e0a\u8ff0Application\u7684main()\u51fd\u6570\u5e76\u4e14\u521b\u5efaSparkContext\uff0c\u5176\u4e2d\u521b\u5efaSparkContext\u7684\u76ee\u7684\u662f\u4e3a\u4e86\u51c6\u5907Spark\u5e94\u7528\u7a0b\u5e8f\u7684\u8fd0\u884c\u73af\u5883\u3002\u5728Spark\u4e2d\u7531SparkContext\u8d1f\u8d23\u548cClusterManager\u901a\u4fe1\uff0c\u8fdb\u884c\u8d44\u6e90\u7684\u7533\u8bf7\u3001\u4efb\u52a1\u7684\u5206\u914d\u548c\u76d1\u63a7\u7b49\uff1b\u5f53Executor\u90e8\u5206\u8fd0\u884c\u5b8c\u6bd5\u540e\uff0cDriver\u8d1f\u8d23\u5c06SparkContext\u5173\u95ed\u3002\u901a\u5e38\u7528SparkContext\u4ee3\u8868Drive\uff1b<\/p>\n<h2>3\u3001Executor\uff1a<\/h2>\n<p>\u6267\u884c\u5668\uff0c\u662f\u4e3a\u4e86\u6267\u884cApplication\u800c\u8fd0\u884c\u5728Worker \u8282\u70b9\u4e0a\u7684\u4e00\u4e2a\u8fdb\u7a0b\uff0c\u8be5\u8fdb\u7a0b\u8d1f\u8d23\u8fd0\u884cTask\uff0c\u5e76\u4e14\u8d1f\u8d23\u5c06\u6570\u636e\u5b58\u5728\u5185\u5b58\u6216\u8005\u78c1\u76d8\u4e0a\uff0c\u6bcf\u4e2aApplication\u90fd\u6709\u5404\u81ea\u72ec\u7acb\u7684\u4e00\u6279Executor\u3002\u5728Spark on Yarn\u6a21\u5f0f\u4e0b\uff0c\u5176\u8fdb\u7a0b\u540d\u79f0\u4e3aCoarseGrainedExecutorBackend\uff0c\u7c7b\u4f3c\u4e8eHadoop MapReduce\u4e2d\u7684YarnChild\u3002\u4e00\u4e2aCoarseGrainedExecutorBackend\u8fdb\u7a0b\u6709\u4e14\u4ec5\u6709\u4e00\u4e2aexecutor\u5bf9\u8c61\uff0c\u5b83\u8d1f\u8d23\u5c06Task\u5305\u88c5\u6210taskRunner\uff0c\u5e76\u4ece\u7ebf\u7a0b\u6c60\u4e2d\u62bd\u53d6\u51fa\u4e00\u4e2a\u7a7a\u95f2\u7ebf\u7a0b\u8fd0\u884cTask\u3002\u6bcf\u4e2aCoarseGrainedExecutorBackend\u80fd\u5e76\u884c\u8fd0\u884cTask\u7684\u6570\u91cf\u5c31\u53d6\u51b3\u4e8e\u5206\u914d\u7ed9\u5b83\u7684CPU\u7684\u4e2a\u6570\u4e86\uff1b<\/p>\n<h2>4\u3001Cluster Manager\uff1a<\/h2>\n<p>\u6307\u7684\u662f\u5728\u96c6\u7fa4\u4e0a\u83b7\u53d6\u8d44\u6e90\u7684\u5916\u90e8\u670d\u52a1\uff0c\u76ee\u524d\u6709\uff1a<\/p>\n<p>\u00d8  Standalone\uff1aSpark\u539f\u751f\u7684\u8d44\u6e90\u7ba1\u7406\uff0c\u7531Master\u8d1f\u8d23\u8d44\u6e90\u7684\u5206\u914d\uff0c\u63a7\u5236\u6574\u4e2a\u96c6\u7fa4\uff0c\u76d1\u63a7worker\uff1b<\/p>\n<p>\u00d8  Hadoop Yarn\uff08\u76ee\u524d\u6700\u6d41\u884c\uff09\uff1a\u7531YARN\u4e2d\u7684ResourceManager\u8d1f\u8d23\u8d44\u6e90\u7684\u5206\u914d\uff1b<\/p>\n<p>Yarn\u4ecb\u7ecd<\/p>\n<blockquote>\n<p>\u5927\u6570\u636e\u751f\u6001\u4e2d\u6709\u5f88\u591a\u4e71\u4e03\u516b\u7cdf\u7684\u5de5\u5177\uff0c\u90fd\u5728\u540c\u4e00\u4e2a\u96c6\u7fa4\u4e0a\u8fd0\u8f6c\uff0c\u5927\u5bb6\u9700\u8981\u4e92\u76f8\u5c0a\u91cd\u6709\u5e8f\u5de5\u4f5c\u3002\u6240\u4ee5\u53e6\u5916\u4e00\u4e2a\u91cd\u8981\u7ec4\u4ef6\u662f\uff0c\u8c03\u5ea6\u7cfb\u7edf\uff1a\u73b0\u5728\u6700\u6d41\u884c\u7684\u662fYarn\u3002\u4f60\u53ef\u4ee5\u628a\u4ed6\u770b\u4f5c\u4e2d\u592e\u7ba1\u7406\uff0c\u597d\u6bd4\u4f60\u5988\u5728\u53a8\u623f\u76d1\u5de5\uff0c\u54ce\uff0c\u4f60\u59b9\u59b9\u5207\u83dc\u5207\u5b8c\u4e86\uff0c\u4f60\u53ef\u4ee5\u628a\u5200\u62ff\u53bb\u6740\u9e21\u4e86\u3002\u53ea\u8981\u5927\u5bb6\u90fd\u670d\u4ece\u4f60\u5988\u5206\u914d\uff0c\u90a3\u5927\u5bb6\u90fd\u80fd\u6109\u5feb\u6ef4\u70e7\u83dc\u3002<\/p>\n<p>\u4f60\u53ef\u4ee5\u8ba4\u4e3a\uff0c\u5927\u6570\u636e\u751f\u6001\u5708\u5c31\u662f\u4e00\u4e2a\u53a8\u623f\u5de5\u5177\u751f\u6001\u5708\u3002\u4e3a\u4e86\u505a\u4e0d\u540c\u7684\u83dc\uff0c\u4e2d\u56fd\u83dc\uff0c\u65e5\u672c\u83dc\uff0c\u6cd5\u56fd\u83dc\uff0c\u4f60\u9700\u8981\u5404\u79cd\u4e0d\u540c\u7684\u5de5\u5177\u3002\u800c\u4e14\u5ba2\u4eba\u7684\u9700\u6c42\u6b63\u5728\u590d\u6742\u5316\uff0c\u4f60\u7684\u53a8\u5177\u4e0d\u65ad\u88ab\u53d1\u660e\uff0c\u4e5f\u6ca1\u6709\u4e00\u4e2a\u4e07\u7528\u7684\u53a8\u5177\u53ef\u4ee5\u5904\u7406\u6240\u6709\u60c5\u51b5\uff0c\u56e0\u6b64\u5b83\u4f1a\u53d8\u7684\u8d8a\u6765\u8d8a\u590d\u6742\u3002\u4f60\u53ef\u4ee5\u8ba4\u4e3a\uff0c\u5927\u6570\u636e\u751f\u6001\u5708\u5c31\u662f\u4e00\u4e2a\u53a8\u623f\u5de5\u5177\u751f\u6001\u5708\u3002\u4e3a\u4e86\u505a\u4e0d\u540c\u7684\u83dc\uff0c\u4e2d\u56fd\u83dc\uff0c\u65e5\u672c\u83dc\uff0c\u6cd5\u56fd\u83dc\uff0c\u4f60\u9700\u8981\u5404\u79cd\u4e0d\u540c\u7684\u5de5\u5177\u3002\u800c\u4e14\u5ba2\u4eba\u7684\u9700\u6c42\u6b63\u5728\u590d\u6742\u5316\uff0c\u4f60\u7684\u53a8\u5177\u4e0d\u65ad\u88ab\u53d1\u660e\uff0c\u4e5f\u6ca1\u6709\u4e00\u4e2a\u4e07\u7528\u7684\u53a8\u5177\u53ef\u4ee5\u5904\u7406\u6240\u6709\u60c5\u51b5\uff0c\u56e0\u6b64\u5b83\u4f1a\u53d8\u7684\u8d8a\u6765\u8d8a\u590d\u6742\u3002<\/p>\n<\/blockquote>\n<h2>5\u3001Worker\uff1a<\/h2>\n<p>\u96c6\u7fa4\u4e2d\u4efb\u4f55\u53ef\u4ee5\u8fd0\u884cApplication\u4ee3\u7801\u7684\u8282\u70b9\uff0c\u7c7b\u4f3c\u4e8eYARN\u4e2d\u7684NodeManager\u8282\u70b9\u3002\u5728Standalone\u6a21\u5f0f\u4e2d\u6307\u7684\u5c31\u662f\u901a\u8fc7Slave\u6587\u4ef6\u914d\u7f6e\u7684Worker\u8282\u70b9\uff0c\u5728Spark on Yarn\u6a21\u5f0f\u4e2d\u6307\u7684\u5c31\u662fNodeManager\u8282\u70b9\uff1b<\/p>\n<h2>6\u3001\u4f5c\u4e1a\uff08Job\uff09\uff1a<\/h2>\n<p>\u5305\u542b\u591a\u4e2aTask\u7ec4\u6210\u7684\u5e76\u884c\u8ba1\u7b97\uff0c\u5f80\u5f80\u7531Spark Action\u50ac\u751f\uff0c \u4e00\u4e2aApplication\u4e2d\u5f80\u5f80\u4f1a\u4ea7\u751f\u591a\u4e2aJob\uff0c\u4e00\u4e2aJOB\u5305\u542b\u591a\u4e2aRDD\u53ca\u4f5c\u7528\u4e8e\u76f8\u5e94RDD\u4e0a\u7684\u5404\u79cdOperation\uff1b<\/p>\n<h2>7\u3001\u9636\u6bb5\uff08Stage\uff09\uff1a<\/h2>\n<p>\u6bcf\u4e2aJob\u4f1a\u88ab\u62c6\u5206\u5f88\u591a\u7ec4Task\uff0c\u4f5c\u4e3a\u4e00\u4e2aTaskSet\uff0c \u5176\u540d\u79f0\u4e3aStage\uff0cStage\u7684\u5212\u5206\u548c\u8c03\u5ea6\u662f\u6709DAGScheduler\u6765\u8d1f\u8d23\u7684\uff0cStage\u6709\u975e\u6700\u7ec8\u7684Stage\uff08Shuffle Map Stage\uff09\u548c\u6700\u7ec8\u7684Stage\uff08Result Stage\uff09\u4e24\u79cd\uff0cStage\u7684\u8fb9\u754c\u5c31\u662f\u53d1\u751fshuffle\u7684\u5730\u65b9<\/p>\n<h2>8\u3001\u4efb\u52a1\uff08Task\uff09\uff1a<\/h2>\n<p>\u88ab\u9001\u5230\u67d0\u4e2aExecutor\u4e0a\u7684\u5de5\u4f5c\u5355\u5143\uff0c\u4f46hadoopMR\u4e2d\u7684MapTask\u548cReduceTask\u6982\u5ff5\u4e00\u6837\uff0c\u662f\u8fd0\u884cApplication\u7684\u57fa\u672c\u5355\u4f4d\uff0c\u591a\u4e2aTask\u7ec4\u6210\u4e00\u4e2aStage\uff0c\u800cTask\u7684\u8c03\u5ea6\u548c\u7ba1\u7406\u7b49\u662f\u7531TaskScheduler\u8d1f\u8d23<\/p>\n<p>\u4e00\u56fe\u8bf4\u660ejob\u3001stage\u3001task\u4e09\u8005\u7684\u5173\u7cfb<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67845.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67845.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<h2>9\u3001DAGScheduler:<\/h2>\n<p>\u6839\u636eJob\u6784\u5efa\u57fa\u4e8eStage\u7684DAG\uff08Directed Acyclic Graph\u6709\u5411\u65e0\u73af\u56fe)\uff0c\u5e76\u63d0\u4ea4Stage\u7ed9TASkScheduler\u3002 \u6839\u636eRDD\u548cStage\u4e4b\u95f4\u7684\u5173\u7cfb\u627e\u51fa\u5f00\u9500\u6700\u5c0f\u7684\u8c03\u5ea6\u65b9\u6cd5\uff0c\u7136\u540e\u628aStage\u4ee5TaskSet\u7684\u5f62\u5f0f\u63d0\u4ea4\u7ed9TaskScheduler\uff0c\u4e0b\u56fe\u5c55\u793a\u4e86DAGScheduler\u7684\u4f5c\u7528\uff1a<\/p>\n<p>DAGScheduler\u51b3\u5b9a\u4e86\u8fd0\u884cTask\u7684\u7406\u60f3\u4f4d\u7f6e\uff0c\u5e76\u628a\u8fd9\u4e9b\u4fe1\u606f\u4f20\u9012\u7ed9\u4e0b\u5c42\u7684TaskScheduler\u3002\u6b64\u5916\uff0cDAGScheduler\u8fd8\u5904\u7406\u7531\u4e8eShuffle\u6570\u636e\u4e22\u5931\u5bfc\u81f4\u7684\u5931\u8d25\uff0c\u8fd9\u6709\u53ef\u80fd\u9700\u8981\u91cd\u65b0\u63d0\u4ea4\u8fd0\u884c\u4e4b\u524d\u7684Stage\uff08\u975eShuffle\u6570\u636e\u4e22\u5931\u5bfc\u81f4\u7684Task\u5931\u8d25\u7531TaskScheduler\u5904\u7406\uff09\u3002<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67846.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67846.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<h2>10\u3001TASKSedulter:<\/h2>\n<p>\u5c06TaskSET\u63d0\u4ea4\u7ed9worker\u8fd0\u884c\uff0c\u6bcf\u4e2aExecutor\u8fd0\u884c\u4ec0\u4e48Task\u5c31\u662f\u5728\u6b64\u5904\u5206\u914d\u7684. TaskScheduler\u7ef4\u62a4\u6240\u6709TaskSet\uff0c\u5f53Executor\u5411Driver\u53d1\u751f\u5fc3\u8df3\u65f6\uff0cTaskScheduler\u4f1a\u6839\u636e\u8d44\u6e90\u5269\u4f59\u60c5\u51b5\u5206\u914d\u76f8\u5e94\u7684Task\u3002\u53e6\u5916TaskScheduler\u8fd8\u7ef4\u62a4\u7740\u6240\u6709Task\u7684\u8fd0\u884c\u6807\u7b7e\uff0c\u91cd\u8bd5\u5931\u8d25\u7684Task\u3002\u4e0b\u56fe\u5c55\u793a\u4e86TaskScheduler\u7684\u4f5c\u7528<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67847.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67847.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>\u6b64\u5916\u987b\u77e5\uff0c\u5728\u4e0d\u540c\u8fd0\u884c\u6a21\u5f0f\u4e2dTASKSedulter\u4efb\u52a1\u8c03\u5ea6\u5668\u5177\u4f53\u4e3a\uff1a<\/p>\n<p>\uff081\uff09Spark on Standalone\u6a21\u5f0f\u4e3aTaskScheduler<\/p>\n<p>\uff082\uff09YARN-Client\u6a21\u5f0f\u4e3aYarnClientClusterScheduler<\/p>\n<p>\uff083\uff09YARN-Cluster\u6a21\u5f0f\u4e3aYarnClusterScheduler<\/p>\n<p>\u5c06\u4e0a\u8ff0\u672f\u8bed\u4e32\u8d77\u6765\u7684\u8fd0\u884c\u5c42\u6b21\u56fe\u5982\u4e0b\uff1a<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67848.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u67848.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>Job=\u591a\u4e2astage\uff0cStage=\u591a\u4e2a\u540c\u79cdtask, Task\u5206\u4e3aShuffleMapTask\u548cResultTask\uff0cDependency\u5206\u4e3aShuffleDependency\u548cNarrowDependency<\/p>\n<h1>\u56db Spark\u8fd0\u884c\u57fa\u672c\u6d41\u7a0b<\/h1>\n<h6>1\u3001\u4e0d\u8bba Spark \u4ee5\u4f55\u79cd\u6a21\u5f0f\u8fdb\u884c\u90e8\u7f72\uff0c\u4efb\u52a1\u63d0\u4ea4\u540e\uff0c\u90fd\u4f1a\u5148\u542f\u52a8 Driver \u8fdb\u7a0b\uff0c\u5373\u542f\u52a8SparkContext\uff0c\u6784\u5efaSpark Application\u7684\u8fd0\u884c\u73af\u5883\uff0c\u968f\u540e Driver\uff08SparkContext \uff09\u8fdb\u7a0b\u5411\u8d44\u6e90\u7ba1\u7406\u5668\uff08\u53ef\u4ee5\u662fStandalone\uff0cMesos\uff0cYarn\uff09\u6ce8\u518c\u5e94\u7528\u7a0b\u5e8f\u3001\u7533\u8bf7\u8fd0\u884cExecutor\u8d44\u6e90<\/h6>\n<h6>2\u3001\u8d44\u6e90\u7ba1\u7406\u5668\u5206\u914dExecutor\u8d44\u6e90\u5e76\u542f\u52a8StandaloneExecutorBackend\uff0cExecutor\u8fd0\u884c\u60c5\u51b5\u5c06\u968f\u7740\u5fc3\u8df3\u53d1\u9001\u5230\u8d44\u6e90\u7ba1\u7406\u5668\u4e0a\uff1b<\/h6>\n<h6>3\u3001Worker \u4e0a\u7684 Executor \u542f\u52a8\u540e\u4f1a\u5411 Driver \u53cd\u5411\u6ce8\u518c\uff0cSparkContext\u77e5\u6653\u5177\u4f53\u7684\u6267\u884c\u8005\u6709\u54ea\u4e9b\u4e4b\u540e\uff0c\u4f1a\u4f9d\u6b64\u6784\u5efa\u751f\u6210DAG\u56fe\uff0c\u5c06DAG\u56fe\u5206\u89e3\u6210Stage\uff0c\u5e76\u628aTaskset\u53d1\u9001\u7ed9Task Scheduler\u3002Executor\u5411SparkContext\u7533\u8bf7Task\uff0cTask Scheduler\u5c06Task\u53d1\u653e\u7ed9Executor\u8fd0\u884c\u540c\u65f6SparkContext\u5c06\u5e94\u7528\u7a0b\u5e8f\u4ee3\u7801\u53d1\u653e\u7ed9Executor\u3002<\/h6>\n<h6>4\u3001Task\u5728Executor\u4e0a\u8fd0\u884c\uff0c\u5728\u4efb\u52a1\u6267\u884c\u7684\u8fc7\u7a0b\u4e2d\uff0cExecutor \u4e5f\u4f1a\u4e0d\u65ad\u4e0e Driver \u8fdb\u884c\u901a\u4fe1\uff0c\u62a5\u544a\u4efb\u52a1\u8fd0\u884c\u60c5\u51b5\u3002\u8fd0\u884c\u5b8c\u91ca\u653e\u6240\u6709\u8d44\u6e90<\/h6>\n<p>\u8be6\u7ec6\u6267\u884c\u56fe<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678410.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678410.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>\u603b\u7ed3spark\u7684\u8fd0\u884c\u7279\u70b9\uff1a<\/p>\n<p>1.\u6bcf\u4e2aApplication\u83b7\u53d6\u4e13\u5c5e\u7684executor\u8fdb\u7a0b\uff0c\u8be5\u8fdb\u7a0b\u5728Application\u671f\u95f4\u4e00\u76f4\u9a7b\u7559\uff0c\u5e76\u4ee5\u591a\u7ebf\u7a0b\u65b9\u5f0f\u8fd0\u884cTask\u3002\u8fd9\u79cdApplication\u9694\u79bb\u673a\u5236\u662f\u6709\u4f18\u52bf\u7684\uff0c\u65e0\u8bba\u662f\u4ece\u8c03\u5ea6\u89d2\u5ea6\u770b\uff08\u6bcf\u4e2aDriver\u8c03\u5ea6\u4ed6\u81ea\u5df1\u7684\u4efb\u52a1\uff09\uff0c\u8fd8\u662f\u4ece\u8fd0\u884c\u89d2\u5ea6\u770b\uff08\u6765\u81ea\u4e0d\u540cApplication\u7684Task\u8fd0\u884c\u5728\u4e0d\u540cJVM\u4e2d\uff09\uff0c\u5f53\u7136\u8fd9\u6837\u610f\u5473\u7740Spark Application\u4e0d\u80fd\u8de8\u5e94\u7528\u7a0b\u5e8f\u5171\u4eab\u6570\u636e\uff0c\u9664\u975e\u5c06\u6570\u636e\u5199\u5165\u5916\u90e8\u5b58\u50a8\u7cfb\u7edf<\/p>\n<p>2.Spark\u4e0e\u8d44\u6e90\u7ba1\u7406\u5668\u65e0\u5173\uff0c\u53ea\u8981\u80fd\u591f\u83b7\u53d6executor\u8fdb\u7a0b\uff0c\u5e76\u80fd\u4fdd\u6301\u76f8\u4e92\u901a\u4fe1\u5c31\u53ef\u4ee5\u4e86<\/p>\n<p>3.\u63d0\u4ea4SparkContext\u7684Client\u5e94\u8be5\u9760\u8fd1Worker\u8282\u70b9\uff08\u8fd0\u884cExecutor\u7684\u8282\u70b9\uff09\uff0c\u6700\u597d\u662f\u5728\u540c\u4e00\u4e2aRack\u91cc\uff0c\u56e0\u4e3aSpark Application\u8fd0\u884c\u8fc7\u7a0b\u4e2dSparkContext\u548cExecutor\u4e4b\u95f4\u6709\u5927\u91cf\u7684\u4fe1\u606f\u4ea4\u6362<\/p>\n<p>4.Task\u91c7\u7528\u4e86\u6570\u636e\u672c\u5730\u6027\u548c\u63a8\u6d4b\u6267\u884c\u7684\u4f18\u5316\u673a\u5236<\/p>\n<h3>RDD\u8fd0\u884c\u539f\u7406<\/h3>\n<p>\u90a3\u4e48 RDD\u5728Spark\u67b6\u6784\u4e2d\u662f\u5982\u4f55\u8fd0\u884c\u7684\u5462\uff1f\u603b\u9ad8\u5c42\u6b21\u6765\u770b\uff0c\u4e3b\u8981\u5206\u4e3a\u4e09\u6b65\uff1a<\/p>\n<p>1.\u521b\u5efa RDD \u5bf9\u8c61<\/p>\n<p>2.DAGScheduler\u6a21\u5757\u4ecb\u5165\u8fd0\u7b97\uff0c\u8ba1\u7b97RDD\u4e4b\u95f4\u7684\u4f9d\u8d56\u5173\u7cfb\u3002RDD\u4e4b\u95f4\u7684\u4f9d\u8d56\u5173\u7cfb\u5c31\u5f62\u6210\u4e86DAG<\/p>\n<p>3.\u6bcf\u4e00\u4e2aJOB\u88ab\u5206\u4e3a\u591a\u4e2aStage\uff0c\u5212\u5206Stage\u7684\u4e00\u4e2a\u4e3b\u8981\u4f9d\u636e\u662f\u5f53\u524d\u8ba1\u7b97\u56e0\u5b50\u7684\u8f93\u5165\u662f\u5426\u662f\u786e\u5b9a\u7684\uff0c\u5982\u679c\u662f\u5219\u5c06\u5176\u5206\u5728\u540c\u4e00\u4e2aStage\uff0c\u907f\u514d\u591a\u4e2aStage\u4e4b\u95f4\u7684\u6d88\u606f\u4f20\u9012\u5f00\u9500\u3002<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/hadoop.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/hadoop.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div>RDD\u8fd0\u884c\u539f\u7406<\/p>\n<p>\u4ee5\u4e0b\u9762\u4e00\u4e2a\u6309 A-Z \u9996\u5b57\u6bcd\u5206\u7c7b\uff0c\u67e5\u627e\u76f8\u540c\u9996\u5b57\u6bcd\u4e0b\u4e0d\u540c\u59d3\u540d\u603b\u4e2a\u6570\u7684\u4f8b\u5b50\u6765\u770b\u4e00\u4e0b RDD \u662f\u5982\u4f55\u8fd0\u884c\u8d77\u6765\u7684\u3002<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678412.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678412.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>\u6b65\u9aa4 1 \uff1a\u521b\u5efa RDD  \u4e0a\u9762\u7684\u4f8b\u5b50\u9664\u53bb\u6700\u540e\u4e00\u4e2a collect \u662f\u4e2a\u52a8\u4f5c\uff0c\u4e0d\u4f1a\u521b\u5efa RDD \u4e4b\u5916\uff0c\u524d\u9762\u56db\u4e2a\u8f6c\u6362\u90fd\u4f1a\u521b\u5efa\u51fa\u65b0\u7684 RDD \u3002\u56e0\u6b64\u7b2c\u4e00\u6b65\u5c31\u662f\u521b\u5efa\u597d\u6240\u6709 RDD( \u5185\u90e8\u7684\u4e94\u9879\u4fe1\u606f ) \u3002<\/p>\n<p>\u6b65\u9aa4 2 \uff1a\u521b\u5efa\u6267\u884c\u8ba1\u5212 Spark \u4f1a\u5c3d\u53ef\u80fd\u5730\u7ba1\u9053\u5316\uff0c\u5e76\u57fa\u4e8e\u662f\u5426\u8981\u91cd\u65b0\u7ec4\u7ec7\u6570\u636e\u6765\u5212\u5206 \u9636\u6bb5 (stage) \uff0c\u4f8b\u5982\u672c\u4f8b\u4e2d\u7684 groupBy() \u8f6c\u6362\u5c31\u4f1a\u5c06\u6574\u4e2a\u6267\u884c\u8ba1\u5212\u5212\u5206\u6210\u4e24\u9636\u6bb5\u6267\u884c\u3002\u6700\u7ec8\u4f1a\u4ea7\u751f\u4e00\u4e2a DAG(directed acyclic graph \uff0c\u6709\u5411\u65e0\u73af\u56fe ) \u4f5c\u4e3a\u903b\u8f91\u6267\u884c\u8ba1\u5212\u3002<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678413.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678413.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>\u6b65\u9aa4 3 \uff1a\u8c03\u5ea6\u4efb\u52a1  \u5c06\u5404\u9636\u6bb5\u5212\u5206\u6210\u4e0d\u540c\u7684 \u4efb\u52a1 (task) \uff0c\u6bcf\u4e2a\u4efb\u52a1\u90fd\u662f\u6570\u636e\u548c\u8ba1\u7b97\u7684\u5408\u4f53\u3002\u5728\u8fdb\u884c\u4e0b\u4e00\u9636\u6bb5\u524d\uff0c\u5f53\u524d\u9636\u6bb5\u7684\u6240\u6709\u4efb\u52a1\u90fd\u8981\u6267\u884c\u5b8c\u6210\u3002\u56e0\u4e3a\u4e0b\u4e00\u9636\u6bb5\u7684\u7b2c\u4e00\u4e2a\u8f6c\u6362\u4e00\u5b9a\u662f\u91cd\u65b0\u7ec4\u7ec7\u6570\u636e\u7684\uff0c\u6240\u4ee5\u5fc5\u987b\u7b49\u5f53\u524d\u9636\u6bb5\u6240\u6709\u7ed3\u679c\u6570\u636e\u90fd\u8ba1\u7b97\u51fa\u6765\u4e86\u624d\u80fd\u7ee7\u7eed\u3002<\/p>\n<p>\u5047\u8bbe\u672c\u4f8b\u4e2d\u7684 hdfs:\/\/names \u4e0b\u6709\u56db\u4e2a\u6587\u4ef6\u5757\uff0c\u90a3\u4e48 HadoopRDD \u4e2d partitions \u5c31\u4f1a\u6709\u56db\u4e2a\u5206\u533a\u5bf9\u5e94\u8fd9\u56db\u4e2a\u5757\u6570\u636e\uff0c\u540c\u65f6 preferedLocations \u4f1a\u6307\u660e\u8fd9\u56db\u4e2a\u5757\u7684\u6700\u4f73\u4f4d\u7f6e\u3002\u73b0\u5728\uff0c\u5c31\u53ef\u4ee5\u521b\u5efa\u51fa\u56db\u4e2a\u4efb\u52a1\uff0c\u5e76\u8c03\u5ea6\u5230\u5408\u9002\u7684\u96c6\u7fa4\u7ed3\u70b9\u4e0a\u3002<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678414.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678414.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<h1>\u4e94 Spark\u8fd0\u884c\u6a21\u5f0f<\/h1>\n<h3>spark\u652f\u6301\u56db\u79cd\u8fd0\u884c\u6a21\u5f0f<\/h3>\n<h4>1\u3001\u672c\u5730\u8fd0\u884c\u6a21\u5f0f\uff08local\u6a21\u5f0f\uff09\uff0c\u5e38\u7528\u4e8e\u672c\u5730\u5f00\u53d1\u6d4b\u8bd5<\/h4>\n<h4>2\u3001\u5206\u5e03\u5f0f\u96c6\u7fa4\u6a21\u5f0f\u4e4b\uff1a\u72ec\u7acb\u8fd0\u884c\u6a21\u5f0f\uff08Standalone\u6a21\u5f0f\uff09<\/h4>\n<p>\u662fSpark \u81ea\u5e26\u7684\u4e00\u79cd\u96c6\u7fa4\u7ba1\u7406\u6a21\u5f0f\uff0c\u5373\u72ec\u7acb\u6a21\u5f0f\uff0c\u81ea\u5e26\u5b8c\u6574\u7684\u670d\u52a1\uff0c\u53ef\u5355\u72ec\u90e8\u7f72\u5230\u4e00\u4e2a\u96c6\u7fa4\u4e2d\uff0c\u65e0\u9700\u4f9d\u8d56\u4efb\u4f55\u5176\u4ed6\u8d44\u6e90\u7ba1\u7406\u7cfb\u7edf\u3002\u5b83\u662f Spark \u5b9e\u73b0\u7684\u8d44\u6e90\u8c03\u5ea6\u6846\u67b6\uff0c\u5176\u4e3b\u8981\u7684\u8282\u70b9\u6709 Driver \u8282\u70b9\u3001Master \u8282\u70b9\u548c Worker \u8282\u70b9\u3002Standalone\u6a21\u5f0f\u4e5f\u662f\u6700\u7b80\u5355\u6700\u5bb9\u6613\u90e8\u7f72\u7684\u4e00\u79cd\u6a21\u5f0f\u3002<\/p>\n<h4>3\u3001\u5206\u5e03\u5f0f\u96c6\u7fa4\u6a21\u5f0f\u4e4b\uff1aSpark on Mesos\u6a21\u5f0f<\/h4>\n<p>\u5373 Spark \u8fd0\u884c\u5728Apache Mesos\u6846\u67b6\u4e4b\u4e0a\u7684\u4e00\u79cd\u6a21\u5f0f\u3002Apache Mesos\u662f\u4e00\u4e2a\u66f4\u5f3a\u5927\u7684\u5206\u5e03\u5f0f\u8d44\u6e90\u7ba1\u7406\u6846\u67b6\uff0c\u8d1f\u8d23\u96c6\u7fa4\u8d44\u6e90\u7684\u5206\u914d\uff0c\u5b83\u5141\u8bb8\u591a\u79cd\u4e0d\u540c\u7684\u6846\u67b6\u90e8\u7f72\u5728\u5176\u4e0a\uff0c\u5305\u62ecYARN\u3002\u5b83\u88ab\u79f0\u4e3a\u662f\u5206\u5e03\u5f0f\u7cfb\u7edf\u7684\u5185\u6838\u3002<\/p>\n<h4>4\u3001\u5206\u5e03\u5f0f\u96c6\u7fa4\u6a21\u5f0f\u4e4b\uff1aSpark on YARN\u6a21\u5f0f<\/h4>\n<p>\u5373 Spark \u8fd0\u884c\u5728Hadoop YARN\u6846\u67b6\u4e4b\u4e0a\u7684\u4e00\u79cd\u6a21\u5f0f\u3002Hadoop YARN\uff08Yet Another Resource<br \/>\nNegotiator\uff0c\u53e6\u4e00\u79cd\u8d44\u6e90\u534f\u8c03\u8005\uff09\u662f\u4e00\u79cd\u65b0\u7684 Hadoop \u8d44\u6e90\u7ba1\u7406\u5668\uff0c\u5b83\u662f\u4e00\u4e2a\u901a\u7528\u8d44\u6e90\u7ba1\u7406\u7cfb\u7edf\uff0c\u53ef\u4e3a\u4e0a\u5c42\u5e94\u7528\u63d0\u4f9b\u7edf\u4e00\u7684\u8d44\u6e90\u7ba1\u7406\u548c\u8c03\u5ea6\u3002<\/p>\n<p>\u9488\u5bf9\u4e09\u79cd\u5206\u5e03\u5f0f\u96c6\u7fa4\u6a21\u5f0f\uff0c\u90fd\u91c7\u7528\u4e86Master\/Worker(Slave)\u7684\u67b6\u6784\uff0cSpark \u5206\u5e03\u5f0f\u8fd0\u884c\u67b6\u6784\u5927\u81f4\u5982\u4e0b\uff1a<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678415.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678415.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>Spark\u5ba2\u6237\u7aef\u76f4\u63a5\u8fde\u63a5Mesos,\u4e0d\u9700\u8981\u989d\u5916\u6784\u5efaSpark\u96c6\u7fa4\uff0c\u56fd\u5185\u7528\u7684\u5c11\uff1b<\/p>\n<p>Spark\u5ba2\u6237\u7aef\u76f4\u63a5\u8fde\u63a5Yarn,\u4e0d\u9700\u8981\u989d\u5916\u6784\u5efaSpark\u96c6\u7fa4\u3002\u56fd\u5185\u751f\u4ea7\u4e0a\u7528\u7684\u591a\u3002<\/p>\n<p><strong>\u800c\u96c6\u7fa4\u6a21\u5f0f\u53c8\u6839\u636eDriver\u8fd0\u884c\u5728\u54ea\u53c8\u5206\u4e3a\u5ba2Client\u6a21\u5f0f\u548cCluster\u6a21\u5f0f<\/strong>\u3002\u7528\u6237\u5728\u63d0\u4ea4\u4efb\u52a1\u7ed9 Spark \u5904\u7406\u65f6\uff0c\u4ee5\u4e0b\u4e24\u4e2a\u53c2\u6570\u5171\u540c\u51b3\u5b9a\u4e86 Spark \u7684\u8fd0\u884c\u65b9\u5f0f\u3002<\/p>\n<p>\u00b7 \u2013master MASTER_URL \uff1a\u51b3\u5b9a\u4e86 Spark \u4efb\u52a1\u63d0\u4ea4\u7ed9\u54ea\u79cd\u96c6\u7fa4\u5904\u7406\u3002<br \/>\n\u00b7 \u2013deploy-mode DEPLOY_MODE\uff1a\u51b3\u5b9a\u4e86 Driver \u7684\u8fd0\u884c\u65b9\u5f0f\uff0c\u53ef\u9009\u503c\u4e3a Client\u6216\u8005 Cluster\u3002<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678416.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678416.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<h2>5.1 standalone\u72ec\u7acb\u96c6\u7fa4\u8fd0\u884c\u6a21\u5f0f<\/h2>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678417.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678417.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>Standalone\u6a21\u5f0f\u4f7f\u7528Spark\u81ea\u5e26\u7684\u8d44\u6e90\u8c03\u5ea6\u6846\u67b6\uff0cStandalone \u96c6\u7fa4\u6709\u56db\u4e2a\u91cd\u8981\u7ec4\u6210\u90e8\u5206\uff0c\u5206\u522b\u662f\uff1a<\/p>\n<h4>1\u3001Driver\uff1a<\/h4>\n<p>\u662f\u4e00\u4e2a\u8fdb\u7a0b\uff0c\u6211\u4eec\u7f16\u5199\u7684 Spark \u5e94\u7528\u7a0b\u5e8f\u5c31\u8fd0\u884c\u5728 Driver \u4e0a\uff0c\u7531Driver \u8fdb\u7a0b\u6267\u884c\uff1b<\/p>\n<blockquote>\n<p>Driver\u65e2\u53ef\u4ee5\u8fd0\u884c\u5728Master\u8282\u70b9\u4e0a\u4e2d\uff0c\u4e5f\u53ef\u4ee5\u8fd0\u884c\u5728\u672c\u5730Client\u7aef\u3002<br \/>\n\u5f53\u7528spark-shell\u4ea4\u4e92\u5f0f\u5de5\u5177\u63d0\u4ea4Spark\u7684Job\u65f6\uff0cDriver\u5728Master\u8282\u70b9\u4e0a\u8fd0\u884c\uff1b<br \/>\n\u5f53\u4f7f\u7528spark-submit\u5de5\u5177\u63d0\u4ea4Job\u6216\u8005\u5728Eclips\u3001IDEA\u7b49\u5f00\u53d1\u5e73\u53f0\u4e0a\u4f7f\u7528<br \/>\n\u201dnew SparkConf.setManager(\u201cspark:\/\/master:7077\u201d)\u201d\u65b9\u5f0f\u8fd0\u884cSpark\u4efb\u52a1\u65f6\uff0cDriver\u662f\u8fd0\u884c\u5728\u672c\u5730Client\u7aef\u4e0a\u7684\u3002<\/p>\n<\/blockquote>\n<h4>2\u3001aster\uff1a<\/h4>\n<p>\u662f\u4e00\u4e2a\u8fdb\u7a0b\uff0c\u4e3b\u8981\u8d1f\u8d23\u8d44\u6e90\u7684\u8c03\u5ea6\u548c\u5206\u914d\uff0c\u5e76\u8fdb\u884c\u96c6\u7fa4\u7684\u76d1\u63a7\u7b49\u804c\u8d23\uff1b<\/p>\n<h4>3\u3001Worker\uff1a<\/h4>\n<p>\u662f\u4e00\u4e2a\u8fdb\u7a0b\uff0c\u4e00\u4e2a Worker \u8fd0\u884c\u5728\u96c6\u7fa4\u4e2d\u7684\u4e00\u53f0\u670d\u52a1\u5668\u4e0a\uff0c\u4e3b\u8981\u8d1f\u8d23\u4e24\u4e2a\u804c\u8d23\uff0c\u4e00\u4e2a\u662f\u7528\u81ea\u5df1\u7684\u5185\u5b58\u5b58\u50a8 RDD \u7684\u67d0\u4e2a\u6216\u67d0\u4e9b partition\uff1b\u53e6\u4e00\u4e2a\u662f\u542f\u52a8\u5176\u4ed6\u8fdb\u7a0b\u548c\u7ebf\u7a0b\uff08Executor\uff09\uff0c\u5bf9 RDD \u4e0a\u7684 partition \u8fdb\u884c\u5e76\u884c\u7684\u5904\u7406\u548c\u8ba1\u7b97\u3002<\/p>\n<h4>4\u3001Executor\uff1a<\/h4>\n<p>\u662f\u4e00\u4e2a\u8fdb\u7a0b\uff0c\u4e00\u4e2a Worker \u4e0a\u53ef\u4ee5\u8fd0\u884c\u591a\u4e2a Executor\uff0cExecutor\u901a\u8fc7\u542f\u52a8\u591a\u4e2a\u7ebf\u7a0b\uff08task\uff09\u6765\u6267\u884c\u5bf9 RDD \u7684 partition \u8fdb\u884c\u5e76\u884c\u8ba1\u7b97\uff0c\u4e5f\u5c31\u662f\u6267\u884c\u6211\u4eec\u5bf9 RDD \u5b9a\u4e49\u7684\u4f8b\u5982 map\u3001flatMap\u3001reduce \u7b49\u7b97\u5b50\u64cd\u4f5c\u3002<\/p>\n<h3>5.1.1 Standalone\u7684client\u6a21\u5f0f<\/h3>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678418.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678418.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>\u89e3\u6790\uff1a<br \/>\n\u5728 Standalone Client \u6a21\u5f0f\u4e0b\uff0cDriver \u5728\u4efb\u52a1\u63d0\u4ea4\u7684\u672c\u5730\u673a\u5668\u4e0a\u8fd0\u884c\uff0cDriver \u542f\u52a8\u540e\u5411 Master \u6ce8\u518c\u5e94\u7528\u7a0b\u5e8f\uff0cMaster \u6839\u636e submit \u811a\u672c\u7684\u8d44\u6e90\u9700\u6c42\u627e\u5230\u5185\u90e8\u8d44\u6e90\u81f3\u5c11\u53ef\u4ee5\u542f\u52a8\u4e00\u4e2a Executor \u7684\u6240\u6709 Worker\uff0c\u7136\u540e\u5728\u8fd9\u4e9b Worker \u4e4b\u95f4\u5206\u914d Executor\uff0cWorker\u4e0a\u7684 Executor \u542f\u52a8\u540e\u4f1a\u5411 Driver \u53cd\u5411\u6ce8\u518c\uff0c\u6240\u6709\u7684 Executor \u6ce8\u518c\u5b8c\u6210\u540e\uff0cDriver \u5f00\u59cb\u6267\u884c main \u51fd\u6570\uff0c\u4e4b\u540e\u6267\u884c\u5230 Action \u7b97\u5b50\u65f6\uff0c\u5f00\u59cb\u5212\u5206 stage\uff0c\u6bcf\u4e2a stage \u751f\u6210\u5bf9\u5e94\u7684 taskSet\uff0c\u4e4b\u540e\u5c06 task \u5206\u53d1\u5230\u5404\u4e2a Executor \u4e0a\u6267\u884c\u3002<\/p>\n<p>\u8fd0\u884c\u6d41\u7a0b\u8be6\u89e3<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678419.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678419.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<h5>1.SparkContext\u8fde\u63a5\u5230Master\uff0c\u5411Master\u6ce8\u518c\u5e76\u7533\u8bf7\u8d44\u6e90\uff08CPU Core \u548cMemory\uff09\uff1b<\/h5>\n<h5>2.Master\u6839\u636eSparkContext\u7684\u8d44\u6e90\u7533\u8bf7\u8981\u6c42\u548cWorker\u5fc3\u8df3\u5468\u671f\u5185\u62a5\u544a\u7684\u4fe1\u606f\u51b3\u5b9a\u5728\u54ea\u4e2aWorker\u4e0a\u5206\u914d\u8d44\u6e90\uff0c\u7136\u540e\u5728\u8be5Worker\u4e0a\u83b7\u53d6\u8d44\u6e90\uff0c\u7136\u540e\u542f\u52a8StandaloneExecutorBackend\uff1b<\/h5>\n<h5>3.StandaloneExecutorBackend\u5411SparkContext\u6ce8\u518c\uff1b<\/h5>\n<h5>4.SparkContext\u5c06Applicaiton\u4ee3\u7801\u53d1\u9001\u7ed9StandaloneExecutorBackend\uff1b\u5e76\u4e14SparkContext\u89e3\u6790Applicaiton\u4ee3\u7801\uff0c\u6784\u5efaDAG\u56fe\uff0c\u5e76\u63d0\u4ea4\u7ed9DAG Scheduler\u5206\u89e3\u6210Stage\uff08\u5f53\u78b0\u5230Action\u64cd\u4f5c\u65f6\uff0c\u5c31\u4f1a\u50ac\u751fJob\uff1b\u6bcf\u4e2aJob\u4e2d\u542b\u67091\u4e2a\u6216\u591a\u4e2aStage\uff0cStage\u4e00\u822c\u5728\u83b7\u53d6\u5916\u90e8\u6570\u636e\u548cshuffle\u4e4b\u524d\u4ea7\u751f\uff09\uff0c\u7136\u540e\u4ee5Stage\uff08\u6216\u8005\u79f0\u4e3aTaskSet\uff09\u63d0\u4ea4\u7ed9Task Scheduler\uff0cTask Scheduler\u8d1f\u8d23\u5c06Task\u5206\u914d\u5230\u76f8\u5e94\u7684Worker\uff0c\u6700\u540e\u63d0\u4ea4\u7ed9StandaloneExecutorBackend\u6267\u884c\uff1b<\/h5>\n<h5>5.StandaloneExecutorBackend\u4f1a\u5efa\u7acbExecutor\u7ebf\u7a0b\u6c60\uff0c\u5f00\u59cb\u6267\u884cTask\uff0c\u5e76\u5411SparkContext\u62a5\u544a\uff0c\u76f4\u81f3Task\u5b8c\u6210\u3002<\/h5>\n<h5>6.\u6240\u6709Task\u5b8c\u6210\u540e\uff0cSparkContext\u5411Master\u6ce8\u9500\uff0c\u91ca\u653e\u8d44\u6e90\u3002<\/h5>\n<h3>5.1.2 Standalone\u7684cluster\u6a21\u5f0f<\/h3>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678420.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678420.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>\u89e3\u6790<br \/>\n\u5728 Standalone Cluster \u6a21\u5f0f\u4e0b\uff0c\u4efb\u52a1\u63d0\u4ea4\u540e\uff0cMaster \u4f1a\u627e\u5230\u4e00\u4e2a Worker \u542f\u52a8 Driver\u8fdb\u7a0b\uff0c Driver \u542f\u52a8\u540e\u5411 Master \u6ce8\u518c\u5e94\u7528\u7a0b\u5e8f\uff0cMaster \u6839\u636e submit \u811a\u672c\u7684\u8d44\u6e90\u9700\u6c42\u627e\u5230\u5185\u90e8\u8d44\u6e90\u81f3\u5c11\u53ef\u4ee5\u542f\u52a8\u4e00\u4e2a Executor \u7684\u6240\u6709 Worker\uff0c\u7136\u540e\u5728\u8fd9\u4e9b Worker \u4e4b\u95f4\u5206\u914d Executor\uff0cWorker \u4e0a\u7684 Executor \u542f\u52a8\u540e\u4f1a\u5411 Driver \u53cd\u5411\u6ce8\u518c\uff0c\u6240\u6709\u7684 Executor\u6ce8\u518c\u5b8c\u6210\u540e\uff0cDriver \u5f00\u59cb\u6267\u884c main \u51fd\u6570\uff0c\u4e4b\u540e\u6267\u884c\u5230 Action \u7b97\u5b50\u65f6\uff0c\u5f00\u59cb\u5212\u5206 stage\uff0c\u6bcf\u4e2a stage \u751f\u6210\u5bf9\u5e94\u7684 taskSet\uff0c\u4e4b\u540e\u5c06 task \u5206\u53d1\u5230\u5404\u4e2a Executor \u4e0a\u6267\u884c\u3002\u6ce8\u610f\uff0cStandalone \u7684\u4e24\u79cd\u6a21\u5f0f\u4e0b\uff08client\/Cluster\uff09\uff0cMaster \u5728\u63a5\u5230 Driver \u6ce8\u518cSpark \u5e94\u7528\u7a0b\u5e8f\u7684\u8bf7\u6c42\u540e\uff0c\u4f1a\u83b7\u53d6\u5176\u6240\u7ba1\u7406\u7684\u5269\u4f59\u8d44\u6e90\u80fd\u591f\u542f\u52a8\u4e00\u4e2a Executor \u7684\u6240\u6709 Worker\uff0c\u7136\u540e\u5728\u8fd9\u4e9b Worker \u4e4b\u95f4\u5206\u53d1 Executor\uff0c\u6b64\u65f6\u7684\u5206\u53d1\u53ea\u8003\u8651 Worker \u4e0a\u7684\u8d44\u6e90\u662f\u5426\u8db3\u591f\u4f7f\u7528\uff0c\u76f4\u5230\u5f53\u524d\u5e94\u7528\u7a0b\u5e8f\u6240\u9700\u7684\u6240\u6709 Executor \u90fd\u5206\u914d\u5b8c\u6bd5\uff0cExecutor\u53cd\u5411\u6ce8\u518c\u5b8c\u6bd5\u540e\uff0cDriver \u5f00\u59cb\u6267\u884c main \u7a0b\u5e8f\u3002<\/p>\n<h3>5.1.3 YARN\u6846\u67b6\u539f\u7406<\/h3>\n<p>\u4f55\u6846\u67b6\u4e0eYARN\u7684\u7ed3\u5408\uff0c\u90fd\u5fc5\u987b\u9075\u5faaYARN\u7684\u5f00\u53d1\u6a21\u5f0f\u3002\u5728\u5206\u6790Spark on YARN\u7684\u5b9e\u73b0\u7ec6\u8282\u4e4b\u524d\uff0c\u6709\u5fc5\u8981\u5148\u5206\u6790\u4e00\u4e0bYARN\u6846\u67b6\u7684\u4e00\u4e9b\u57fa\u672c\u539f\u7406\u3002<\/p>\n<p>Yarn\u6846\u67b6\u7684\u57fa\u672c\u8fd0\u884c\u6d41\u7a0b\u56fe\u4e3a\uff1a<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678421.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678421.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>\u5176\u4e2d\uff0cResourceManager\u8d1f\u8d23\u5c06\u96c6\u7fa4\u7684\u8d44\u6e90\u5206\u914d\u7ed9\u5404\u4e2a\u5e94\u7528\u4f7f\u7528\uff0c\u800c\u8d44\u6e90\u5206\u914d\u548c\u8c03\u5ea6\u7684\u57fa\u672c\u5355\u4f4d\u662fContainer\uff0c\u5176\u4e2d\u5c01\u88c5\u4e86\u673a\u5668\u8d44\u6e90\uff0c\u5982\u5185\u5b58\u3001CPU\u3001\u78c1\u76d8\u548c\u7f51\u7edc\u7b49\uff0c\u6bcf\u4e2a\u4efb\u52a1\u4f1a\u88ab\u5206\u914d\u4e00\u4e2aContainer\uff0c\u8be5\u4efb\u52a1\u53ea\u80fd\u5728\u8be5Container\u4e2d\u6267\u884c\uff0c\u5e76\u4f7f\u7528\u8be5Container\u5c01\u88c5\u7684\u8d44\u6e90\u3002NodeManager\u662f\u4e00\u4e2a\u4e2a\u7684\u8ba1\u7b97\u8282\u70b9\uff0c\u4e3b\u8981\u8d1f\u8d23\u542f\u52a8Application\u6240\u9700\u7684Container\uff0c\u76d1\u63a7\u8d44\u6e90\uff08\u5185\u5b58\u3001CPU\u3001\u78c1\u76d8\u548c\u7f51\u7edc\u7b49\uff09\u7684\u4f7f\u7528\u60c5\u51b5\u5e76\u5c06\u4e4b\u6c47\u62a5\u7ed9ResourceManager\u3002ResourceManager\u4e0eNodeManagers\u5171\u540c\u7ec4\u6210\u6574\u4e2a\u6570\u636e\u8ba1\u7b97\u6846\u67b6\uff0cApplicationMaster\u4e0e\u5177\u4f53\u7684Application\u76f8\u5173\uff0c\u4e3b\u8981\u8d1f\u8d23\u540cResourceManager\u534f\u5546\u4ee5\u83b7\u53d6\u5408\u9002\u7684Container\uff0c\u5e76\u8ddf\u8e2a\u8fd9\u4e9bContainer\u7684\u72b6\u6001\u548c\u76d1\u63a7\u5176\u8fdb\u5ea6\u3002<\/p>\n<h3>5.1.4 YARN\u7684client\u8fd0\u884c\u6a21\u5f0f<\/h3>\n<p>Yarn-Client\u6a21\u5f0f\u4e2d\uff0cDriver\u5728\u5ba2\u6237\u7aef\u672c\u5730\u8fd0\u884c\uff0c\u8fd9\u79cd\u6a21\u5f0f\u53ef\u4ee5\u4f7f\u5f97Spark Application\u548c\u5ba2\u6237\u7aef\u8fdb\u884c\u4ea4\u4e92\uff0c\u56e0\u4e3aDriver\u5728\u5ba2\u6237\u7aef\uff0c\u6240\u4ee5\u53ef\u4ee5\u901a\u8fc7webUI\u8bbf\u95eeDriver\u7684\u72b6\u6001\uff0c\u9ed8\u8ba4\u662fhttp:\/\/hadoop1:4040\u8bbf\u95ee\uff0c\u800cYARN\u901a\u8fc7http:\/\/ hadoop1:8088\u8bbf\u95ee\u3002<\/p>\n<p>YARN-client\u7684\u5de5\u4f5c\u6d41\u7a0b\u5206\u4e3a\u4ee5\u4e0b\u51e0\u4e2a\u6b65\u9aa4\uff1a<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678422.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678422.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<h5>1.Spark Yarn Client\u5411YARN\u7684ResourceManager\u7533\u8bf7\u542f\u52a8Application Master\u3002<\/h5>\n<p>\u540c\u65f6\u5728SparkContent\u521d\u59cb\u5316\u4e2d\u5c06\u521b\u5efaDAGScheduler\u548cTASKScheduler\u7b49\uff0c\u7531\u4e8e\u6211\u4eec\u9009\u62e9\u7684\u662fYarn-Client\u6a21\u5f0f\uff0c\u7a0b\u5e8f\u4f1a\u9009\u62e9YarnClientClusterScheduler\u548cYarnClientSchedulerBackend\uff1b<\/p>\n<h5>2.ResourceManager\u6536\u5230\u8bf7\u6c42\u540e<\/h5>\n<p>\uff0c\u5728\u96c6\u7fa4\u4e2d\u9009\u62e9\u4e00\u4e2aNodeManager\uff0c\u4e3a\u8be5\u5e94\u7528\u7a0b\u5e8f\u5206\u914d\u7b2c\u4e00\u4e2aContainer\uff0c\u8981\u6c42\u5b83\u5728\u8fd9\u4e2aContainer\u4e2d\u542f\u52a8\u5e94\u7528\u7a0b\u5e8f\u7684ApplicationMaster\uff0c\u4e0eYARN-Cluster\u533a\u522b\u7684\u662f\u5728\u8be5ApplicationMaster\u4e0d\u8fd0\u884cSparkContext\uff0c\u53ea\u4e0eSparkContext\u8fdb\u884c\u8054\u7cfb\u8fdb\u884c\u8d44\u6e90\u7684\u5206\u6d3e\uff1b<\/p>\n<h5>3.Client\u4e2d\u7684SparkContext\u521d\u59cb\u5316\u5b8c\u6bd5\u540e<\/h5>\n<p>\uff0c\u4e0eApplicationMaster\u5efa\u7acb\u901a\u8baf\uff0c\u5411ResourceManager\u6ce8\u518c\uff0c\u6839\u636e\u4efb\u52a1\u4fe1\u606f\u5411ResourceManager\u7533\u8bf7\u8d44\u6e90\uff08Container\uff09\uff1b<\/p>\n<h5>4.\u4e00\u65e6ApplicationMaster\u7533\u8bf7\u5230\u8d44\u6e90\uff08\u4e5f\u5c31\u662fContainer\uff09\u540e\uff0c<\/h5>\n<p>\u4fbf\u4e0e\u5bf9\u5e94\u7684NodeManager\u901a\u4fe1\uff0c\u8981\u6c42\u5b83\u5728\u83b7\u5f97\u7684Container\u4e2d\u542f\u52a8\u542f\u52a8CoarseGrainedExecutorBackend\uff0cCoarseGrainedExecutorBackend\u542f\u52a8\u540e\u4f1a\u5411Client\u4e2d\u7684SparkContext\u6ce8\u518c\u5e76\u7533\u8bf7Task\uff1b<\/p>\n<h5>5.Client\u4e2d\u7684SparkContext\u5206\u914dTask\u7ed9CoarseGrainedExecutorBackend\u6267\u884c,<\/h5>\n<p>CoarseGrainedExecutorBackend\u8fd0\u884cTask\u5e76\u5411Driver\u6c47\u62a5\u8fd0\u884c\u7684\u72b6\u6001\u548c\u8fdb\u5ea6\uff0c\u4ee5\u8ba9Client\u968f\u65f6\u638c\u63e1\u5404\u4e2a\u4efb\u52a1\u7684\u8fd0\u884c\u72b6\u6001\uff0c\u4ece\u800c\u53ef\u4ee5\u5728\u4efb\u52a1\u5931\u8d25\u65f6\u91cd\u65b0\u542f\u52a8\u4efb\u52a1\uff1b<\/p>\n<h5>6.\u5e94\u7528\u7a0b\u5e8f\u8fd0\u884c\u5b8c\u6210\u540e\uff0c<\/h5>\n<p>Client\u7684SparkContext\u5411ResourceManager\u7533\u8bf7\u6ce8\u9500\u5e76\u5173\u95ed\u81ea\u5df1\u3002<\/p>\n<h3>5.1.5 YARN\u7684cluster\u8fd0\u884c\u6a21\u5f0f<\/h3>\n<p>\u5728YARN-Cluster\u6a21\u5f0f\u4e2d\uff0c\u5f53\u7528\u6237\u5411YARN\u4e2d\u63d0\u4ea4\u4e00\u4e2a\u5e94\u7528\u7a0b\u5e8f\u540e\uff0cYARN\u5c06\u5206\u4e24\u4e2a\u9636\u6bb5\u8fd0\u884c\u8be5\u5e94\u7528\u7a0b\u5e8f\uff1a<\/p>\n<pre><code>1.\u7b2c\u4e00\u4e2a\u9636\u6bb5\u662f\u628aSpark\u7684Driver\u4f5c\u4e3a\u4e00\u4e2aApplicationMaster\u5728YARN\u96c6\u7fa4\u4e2d\u5148\u542f\u52a8\uff1b\n\n2.\u7b2c\u4e8c\u4e2a\u9636\u6bb5\u662f\u7531ApplicationMaster\u521b\u5efa\u5e94\u7528\u7a0b\u5e8f\uff0c\u7136\u540e\u4e3a\u5b83\u5411ResourceManager\u7533\u8bf7\u8d44\u6e90\uff0c\u5e76\u542f\u52a8Executor\u6765\u8fd0\u884cTask\uff0c\u540c\u65f6\u76d1\u63a7\u5b83\u7684\u6574\u4e2a\u8fd0\u884c\u8fc7\u7a0b\uff0c\u76f4\u5230\u8fd0\u884c\u5b8c\u6210<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678423.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678423.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<h5>1.   Spark Yarn Client\u5411YARN\u4e2d\u63d0\u4ea4\u5e94\u7528\u7a0b\u5e8f\uff0c<\/h5>\n<p>\u5305\u62ecApplicationMaster\u7a0b\u5e8f\u3001\u542f\u52a8ApplicationMaster\u7684\u547d\u4ee4\u3001\u9700\u8981\u5728Executor\u4e2d\u8fd0\u884c\u7684\u7a0b\u5e8f\u7b49\uff1b<\/p>\n<h5>2.   ResourceManager\u6536\u5230\u8bf7\u6c42\u540e\uff0c<\/h5>\n<p>\u5728\u96c6\u7fa4\u4e2d\u9009\u62e9\u4e00\u4e2aNodeManager\uff0c\u4e3a\u8be5\u5e94\u7528\u7a0b\u5e8f\u5206\u914d\u7b2c\u4e00\u4e2aContainer\uff0c\u8981\u6c42\u5b83\u5728\u8fd9\u4e2aContainer\u4e2d\u542f\u52a8\u5e94\u7528\u7a0b\u5e8f\u7684ApplicationMaster\uff0c\u5176\u4e2dApplicationMaster\u8fdb\u884cSparkContext\u7b49\u7684\u521d\u59cb\u5316\uff1b<\/p>\n<h5>3. ApplicationMaster\u5411ResourceManager\u6ce8\u518c\uff0c<\/h5>\n<p>\u8fd9\u6837\u7528\u6237\u53ef\u4ee5\u76f4\u63a5\u901a\u8fc7ResourceManage\u67e5\u770b\u5e94\u7528\u7a0b\u5e8f\u7684\u8fd0\u884c\u72b6\u6001\uff0c\u7136\u540e\u5b83\u5c06\u91c7\u7528\u8f6e\u8be2\u7684\u65b9\u5f0f\u901a\u8fc7RPC\u534f\u8bae\u4e3a\u5404\u4e2a\u4efb\u52a1\u7533\u8bf7\u8d44\u6e90\uff0c\u5e76\u76d1\u63a7\u5b83\u4eec\u7684\u8fd0\u884c\u72b6\u6001\u76f4\u5230\u8fd0\u884c\u7ed3\u675f\uff1b<\/p>\n<h5>4. \u4e00\u65e6ApplicationMaster\u7533\u8bf7\u5230\u8d44\u6e90\uff08\u4e5f\u5c31\u662fContainer\uff09\u540e\uff0c<\/h5>\n<p>\u4fbf\u4e0e\u5bf9\u5e94\u7684NodeManager\u901a\u4fe1\uff0c\u8981\u6c42\u5b83\u5728\u83b7\u5f97\u7684Container\u4e2d\u542f\u52a8\u542f\u52a8CoarseGrainedExecutorBackend\uff0cCoarseGrainedExecutorBackend\u542f\u52a8\u540e\u4f1a\u5411ApplicationMaster\u4e2d\u7684SparkContext\u6ce8\u518c\u5e76\u7533\u8bf7Task\u3002\u8fd9\u4e00\u70b9\u548cStandalone\u6a21\u5f0f\u4e00\u6837\uff0c\u53ea\u4e0d\u8fc7SparkContext\u5728Spark Application\u4e2d\u521d\u59cb\u5316\u65f6\uff0c\u4f7f\u7528CoarseGrainedSchedulerBackend\u914d\u5408YarnClusterScheduler\u8fdb\u884c\u4efb\u52a1\u7684\u8c03\u5ea6\uff0c\u5176\u4e2dYarnClusterScheduler\u53ea\u662f\u5bf9TaskSchedulerImpl\u7684\u4e00\u4e2a\u7b80\u5355\u5305\u88c5\uff0c\u589e\u52a0\u4e86\u5bf9Executor\u7684\u7b49\u5f85\u903b\u8f91\u7b49\uff1b<\/p>\n<h5>5. ApplicationMaster\u4e2d\u7684SparkContext\u5206\u914dTask\u7ed9CoarseGrainedExecutorBackend\u6267\u884c\uff0c<\/h5>\n<p>CoarseGrainedExecutorBackend\u8fd0\u884cTask\u5e76\u5411ApplicationMaster\u6c47\u62a5\u8fd0\u884c\u7684\u72b6\u6001\u548c\u8fdb\u5ea6\uff0c\u4ee5\u8ba9ApplicationMaster\u968f\u65f6\u638c\u63e1\u5404\u4e2a\u4efb\u52a1\u7684\u8fd0\u884c\u72b6\u6001\uff0c\u4ece\u800c\u53ef\u4ee5\u5728\u4efb\u52a1\u5931\u8d25\u65f6\u91cd\u65b0\u542f\u52a8\u4efb\u52a1\uff1b<\/p>\n<h5>6. \u5e94\u7528\u7a0b\u5e8f\u8fd0\u884c\u5b8c\u6210\u540e,<\/h5>\n<p>ApplicationMaster\u5411ResourceManager\u7533\u8bf7\u6ce8\u9500\u5e76\u5173\u95ed\u81ea\u5df1\u3002<\/p>\n<h3>5.1.6 YARN-Client \u4e0e YARN-Cluster \u533a\u522b<\/h3>\n<p>\u7406\u89e3YARN-Client\u548cYARN-Cluster\u6df1\u5c42\u6b21\u7684\u533a\u522b\u4e4b\u524d\u5148\u6e05\u695a\u4e00\u4e2a\u6982\u5ff5\uff1aApplication Master\u3002\u5728YARN\u4e2d\uff0c\u6bcf\u4e2aApplication\u5b9e\u4f8b\u90fd\u6709\u4e00\u4e2aApplicationMaster\u8fdb\u7a0b\uff0c\u5b83\u662fApplication\u542f\u52a8\u7684\u7b2c\u4e00\u4e2a\u5bb9\u5668\u3002\u5b83\u8d1f\u8d23\u548cResourceManager\u6253\u4ea4\u9053\u5e76\u8bf7\u6c42\u8d44\u6e90\uff0c\u83b7\u53d6\u8d44\u6e90\u4e4b\u540e\u544a\u8bc9NodeManager\u4e3a\u5176\u542f\u52a8Container\u3002\u4ece\u6df1\u5c42\u6b21\u7684\u542b\u4e49\u8bb2YARN-Cluster\u548cYARN-Client\u6a21\u5f0f\u7684\u533a\u522b\u5176\u5b9e\u5c31\u662fApplicationMaster\u8fdb\u7a0b\u7684\u533a\u522b\u3002<\/p>\n<h5>l  YARN-Cluster\u6a21\u5f0f\u4e0b\uff0c<\/h5>\n<p>Driver\u8fd0\u884c\u5728AM(Application Master)\u4e2d\uff0c\u5b83\u8d1f\u8d23\u5411YARN\u7533\u8bf7\u8d44\u6e90\uff0c\u5e76\u76d1\u7763\u4f5c\u4e1a\u7684\u8fd0\u884c\u72b6\u51b5\u3002\u5f53\u7528\u6237\u63d0\u4ea4\u4e86\u4f5c\u4e1a\u4e4b\u540e\uff0c\u5c31\u53ef\u4ee5\u5173\u6389Client\uff0c\u4f5c\u4e1a\u4f1a\u7ee7\u7eed\u5728YARN\u4e0a\u8fd0\u884c\uff0c\u56e0\u800cYARN-Cluster\u6a21\u5f0f\u4e0d\u9002\u5408\u8fd0\u884c\u4ea4\u4e92\u7c7b\u578b\u7684\u4f5c\u4e1a\uff1b<\/p>\n<p>l  YARN-Client\u6a21\u5f0f\u4e0b\uff0cApplication Master\u4ec5\u4ec5\u5411YARN\u8bf7\u6c42Executor\uff0cClient\u4f1a\u548c\u8bf7\u6c42\u7684Container\u901a\u4fe1\u6765\u8c03\u5ea6\u4ed6\u4eec\u5de5\u4f5c\uff0c\u4e5f\u5c31\u662f\u8bf4Client\u4e0d\u80fd\u79bb\u5f00\u3002<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678424.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678424.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p>\u8865\u5145<\/p>\n<blockquote>\n<p>\u56e0\u4e3a\u5728Spark\u4f5c\u4e1a\u8fd0\u884c\u8fc7\u7a0b\u4e2d\uff0c\u4e00\u822c\u60c5\u51b5\u4e0b\u4f1a\u6709\u5927\u91cf\u6570\u636e\u5728Driver\u548c\u96c6\u7fa4\u4e2d\u8fdb\u884c\u4ea4\u4e92\uff0c<br \/>\n\u6240\u4ee5\u5982\u679c\u662f\u57fa\u4e8eyarn-client\u7684\u6a21\u5f0f\uff0c\u5219\u4f1a\u5728\u7a0b\u5e8f\u8fd0\u884c\u8fc7\u7a0b\u4e2d\u4ea7\u751f\u5927\u91cf\u7684\u7f51\u7edc\u6570\u636e\u4f20\u8f93\uff0c<br \/>\n\u9020\u6210\u7f51\u5361\u6d41\u91cf\u6fc0\u589e\uff1b\u800c\u57fa\u4e8eyarn-cluster\u8fd9\u79cd\u6a21\u5f0f\uff0c\u56e0\u4e3adriver\u672c\u8eab\u5c31\u5728\u96c6\u7fa4\u5185\u90e8\uff0c<br \/>\n\u6240\u4ee5\u6570\u636e\u7684\u4f20\u8f93\u4e5f\u662f\u5728\u96c6\u7fa4\u5185\u90e8\u6765\u5b8c\u6210\uff0c\u90a3\u4e48\u7f51\u7edc\u4f20\u8f93\u538b\u529b\u76f8\u5bf9\u8981\u5c0f\uff1b<br \/>\n\u6240\u4ee5\u5728\u4f01\u4e1a\u751f\u4ea7\u73af\u5883\u4e0b\u591a\u4f7f\u7528yarn-cluster\u8fd9\u79cd\u6a21\u5f0f\uff0c\u6d4b\u8bd5\u591a\u7528yarn-client\u8fd9\u79cd\u6a21\u5f0f\u3002<br \/>\n\u4f46\u662f\u5e26\u6765\u4e00\u4e2a\u95ee\u9898\uff0c\u5c31\u662f\u4e0d\u65b9\u4fbf\u76d1\u63a7\u65e5\u5fd7\uff0cyarn-cluster\u8fd9\u79cd\u6a21\u5f0f\u8981\u60f3\u76d1\u63a7\u65e5\u5fd7\uff0c<br \/>\n\u5fc5\u987b\u8981\u5230\u6bcf\u4e00\u53f0\u673a\u5668\u4e0a\u9762\u53bb\u67e5\u770b\uff0c\u4f46\u8fd9\u90fd\u4e0d\u662f\u95ee\u9898\uff0c\u56e0\u4e3a\u6211\u4eec\u6709sparkUI\uff0c\u540c\u65f6\u4e5f\u6709\u5404\u79cd\u5404\u6837\u7684\u65e5\u5fd7\u76d1\u63a7\u7ec4\u4ef6<\/p>\n<\/blockquote>\n<h1>\u516d spark\u4f7f\u7528\u4e0e\u4e0d\u540c\u6a21\u5f0f\u7684\u90e8\u7f72<\/h1>\n<p>\u5b98\u65b9\uff1a<a href=\"https:\/\/spark-reference-doc-cn.readthedocs.io\/zh_CN\/latest\/deploy-guide\/spark-standalone.html\">https:\/\/spark-reference-doc-cn.readthedocs.io\/zh_CN\/latest\/deploy-guide\/spark-standalone.html<\/a><\/p>\n<p>Standalone spark\u81ea\u5e26\u7684\u96c6\u7fa4\u6a21\u5f0f\u3002\u9700\u8981\u6784\u5efa\u4e00\u4e2a\u7531Master+Slave\u6784\u6210\u7684Spark\u96c6\u7fa4\uff0c\u9009\u7528ZooKeeper\u6765\u5b9e\u73b0Master\u7684HA\uff0c\u6846\u67b6\u7ed3\u6784\u56fe\u5982\u4e0b:<br \/>\n<div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678425.png'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/egonlin.com\/wp-content\/uploads\/2022\/04\/Spark\u8fd0\u884c\u67b6\u678425.png\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\" \/><\/div><\/p>\n<p><a href=\"https:\/\/zhuanlan.zhihu.com\/p\/99398378\">https:\/\/zhuanlan.zhihu.com\/p\/99398378<\/a><\/p>\n<p>\u63a8\u8350\u9605\u8bfb\uff1a<\/p>\n<p><a href=\"https:\/\/www.cnblogs.com\/shishanyuan\/p\/4721326.html\">https:\/\/www.cnblogs.com\/shishanyuan\/p\/4721326.html<\/a><\/p>\n<p><a href=\"https:\/\/blog.csdn.net\/shuimofengyang\/article\/details\/100124601\">https:\/\/blog.csdn.net\/shuimofengyang\/article\/details\/100124601<\/a> <\/p>\n<p><a href=\"https:\/\/www.cnblogs.com\/jinggangshan\/p\/8063970.html\">https:\/\/www.cnblogs.com\/jinggangshan\/p\/8063970.html<\/a><\/p>\n<p><a href=\"https:\/\/blog.csdn.net\/github_26054561\/article\/details\/46344889\">https:\/\/blog.csdn.net\/github_26054561\/article\/details\/46344889<\/a><\/p>\n<p><a href=\"https:\/\/juejin.cn\/post\/6844904058612875277\">https:\/\/juejin.cn\/post\/6844904058612875277<\/a><\/p>\n<p><a href=\"https:\/\/zhuanlan.zhihu.com\/p\/99398378\">https:\/\/zhuanlan.zhihu.com\/p\/99398378<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u4e00 Spark\u4e0ehadoop Hadoop\u6709\u4e24\u4e2a\u6838\u5fc3\u6a21\u5757\uff0c\u5206\u5e03\u5f0f\u5b58\u50a8\u6a21\u5757HDFS\u548c\u5206\u5e03\u5f0f\u8ba1\u7b97\u6a21\u5757Mapredu [&hellip;]<\/p>\n","protected":false},"author":3,"featured_media":6492,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[411],"tags":[],"_links":{"self":[{"href":"https:\/\/egonlin.com\/index.php?rest_route=\/wp\/v2\/posts\/6467"}],"collection":[{"href":"https:\/\/egonlin.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/egonlin.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/egonlin.com\/index.php?rest_route=\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/egonlin.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=6467"}],"version-history":[{"count":0,"href":"https:\/\/egonlin.com\/index.php?rest_route=\/wp\/v2\/posts\/6467\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/egonlin.com\/index.php?rest_route=\/wp\/v2\/media\/6492"}],"wp:attachment":[{"href":"https:\/\/egonlin.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6467"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/egonlin.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6467"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/egonlin.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6467"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}