Apache Flink 定義

Apache Flink是一個(gè)框架和分布式處理引擎，用于對(duì)無界和有界數(shù)據(jù)流進(jìn)行狀態(tài)計(jì)算。Flink設(shè)計(jì)為在所有常見的集群環(huán)境中運(yùn)行，以內(nèi)存速度和任何規(guī)模執(zhí)行計(jì)算。

設(shè)置：下載并啟動(dòng)Flink

Flink可在Linux，Mac OS X和Windows上運(yùn)行。為了能夠運(yùn)行Flink，唯一的要求是安裝Java 8.x。Windows用戶，請(qǐng)查看Windows上的Flink指南，該指南介紹了如何在Windows上運(yùn)行Flink以進(jìn)行本地設(shè)置。

您可以通過發(fā)出以下命令來檢查Java的正確安裝：

java -version

如果你有Java 8，輸出將如下所示：

java version "1.8.0_201"
Java(TM) SE Runtime Environment (build 1.8.0_201-b09)
Java HotSpot(TM) 64-Bit Server VM (build 25.201-b09, mixed mode)

下載和解壓縮

1. 從下載頁(yè)面下載二進(jìn)制文件。您可以選擇任何您喜歡的Hadoop / Scala組合。如果您打算只使用本地文件系統(tǒng)，任何Hadoop版本都可以正常工作。
2. 轉(zhuǎn)到下載目錄。
3. 解壓縮下載的存檔。

$ cd ~/Downloads # Go to download directory 
$ tar xzf flink-*.tgz # Unpack the downloaded archive 
$ cd flink-1.8.0

對(duì)于MacOS X用戶，可以通過Homebrew安裝Flink 。

$ brew install apache-flink ... 
$ flink --version Version: 1.8.0, Commit ID: 4caec0d

啟動(dòng)本地Flink群集

$ ./bin/start-cluster.sh # Start Flink

檢查分派器的web前端在HTTP：//本地主機(jī)：8081，并確保一切都正常運(yùn)行。Web前端應(yīng)報(bào)告單個(gè)可用的TaskManager實(shí)例。

jobmanager-1.png

您還可以通過檢查logs目錄中的日志文件來驗(yàn)證系統(tǒng)是否正在運(yùn)行：

$ tail log/flink-*-standalonesession-*.log
INFO ... - Rest endpoint listening at localhost:8081
INFO ... - http://localhost:8081 was granted leadership ...
INFO ... - Web frontend listening at http://localhost:8081.
INFO ... - Starting RPC endpoint for StandaloneResourceManager at akka://flink/user/resourcemanager .
INFO ... - Starting RPC endpoint for StandaloneDispatcher at akka://flink/user/dispatcher .
INFO ... - ResourceManager akka.tcp://flink@localhost:6123/user/resourcemanager was granted leadership ...
INFO ... - Starting the SlotManager.
INFO ... - Dispatcher akka.tcp://flink@localhost:6123/user/dispatcher was granted leadership ...
INFO ... - Recovering all persisted jobs.
INFO ... - Registering TaskManager ... under ... at the SlotManager.

閱讀代碼

您可以在Scala中找到此SocketWindowWordCount示例的完整源代碼，并在GitHub上找到Java。

Scala

object SocketWindowWordCount {

    def main(args: Array[String]) : Unit = {

        // the port to connect to
        val port: Int = try {
            ParameterTool.fromArgs(args).getInt("port")
        } catch {
            case e: Exception => {
                System.err.println("No port specified. Please run 'SocketWindowWordCount --port <port>'")
                return
            }
        }

        // get the execution environment
        val env: StreamExecutionEnvironment = StreamExecutionEnvironment.getExecutionEnvironment

        // get input data by connecting to the socket
        val text = env.socketTextStream("localhost", port, '\n')

        // parse the data, group it, window it, and aggregate the counts
        val windowCounts = text
            .flatMap { w => w.split("\\s") }
            .map { w => WordWithCount(w, 1) }
            .keyBy("word")
            .timeWindow(Time.seconds(5), Time.seconds(1))
            .sum("count")

        // print the results with a single thread, rather than in parallel
        windowCounts.print().setParallelism(1)

        env.execute("Socket Window WordCount")
    }

    // Data type for words with count
    case class WordWithCount(word: String, count: Long)
}

JAVA

public class SocketWindowWordCount {

    public static void main(String[] args) throws Exception {

        // the port to connect to
        final int port;
        try {
            final ParameterTool params = ParameterTool.fromArgs(args);
            port = params.getInt("port");
        } catch (Exception e) {
            System.err.println("No port specified. Please run 'SocketWindowWordCount --port <port>'");
            return;
        }

        // get the execution environment
        final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();

        // get input data by connecting to the socket
        DataStream<String> text = env.socketTextStream("localhost", port, "\n");

        // parse the data, group it, window it, and aggregate the counts
        DataStream<WordWithCount> windowCounts = text
            .flatMap(new FlatMapFunction<String, WordWithCount>() {
                @Override
                public void flatMap(String value, Collector<WordWithCount> out) {
                    for (String word : value.split("\\s")) {
                        out.collect(new WordWithCount(word, 1L));
                    }
                }
            })
            .keyBy("word")
            .timeWindow(Time.seconds(5), Time.seconds(1))
            .reduce(new ReduceFunction<WordWithCount>() {
                @Override
                public WordWithCount reduce(WordWithCount a, WordWithCount b) {
                    return new WordWithCount(a.word, a.count + b.count);
                }
            });

        // print the results with a single thread, rather than in parallel
        windowCounts.print().setParallelism(1);

        env.execute("Socket Window WordCount");
    }

    // Data type for words with count
    public static class WordWithCount {

        public String word;
        public long count;

        public WordWithCount() {}

        public WordWithCount(String word, long count) {
            this.word = word;
            this.count = count;
        }

        @Override
        public String toString() {
            return word + " : " + count;
        }
    }
}

運(yùn)行示例

現(xiàn)在，我們將運(yùn)行此Flink應(yīng)用程序。它將從套接字讀取文本，并且每5秒打印一次前5秒內(nèi)每個(gè)不同單詞的出現(xiàn)次數(shù)，即處理時(shí)間的翻滾窗口，只要文字漂浮在其中。

• 首先，我們使用netcat來啟動(dòng)本地服務(wù)器

$ nc -l 9000

• 提交Flink計(jì)劃：

$ ./bin/flink run examples/streaming/SocketWindowWordCount.jar --port 9000 Starting execution of program

程序連接到套接字并等待輸入。您可以檢查Web界面以驗(yàn)證作業(yè)是否按預(yù)期運(yùn)行：

• 單詞在5秒的時(shí)間窗口（處理時(shí)間，翻滾窗口）中計(jì)算并打印到stdout。監(jiān)視TaskManager的輸出文件并寫入一些文本nc（輸入在點(diǎn)擊后逐行發(fā)送到Flink）：

$ nc -l 9000
lorem ipsum
ipsum ipsum ipsum
bye

該.out文件將在每個(gè)時(shí)間窗口結(jié)束時(shí)，只要打印算作字浮在，例如：

$ tail -f log/flink-*-taskexecutor-*.out
lorem : 1
bye : 1
ipsum : 4

要停止Flink 所要做的操作：

$ ./bin/stop-cluster.sh