一、前言

Flume作为当下最风行的大数据采集组件之一。其自身领有分布式/高牢靠/高可用等长处,但相比拟于Flink/Spark/Kafka等大数据组件,其对于本地调试的性能反对度并不高,如果咱们没有把握Flume的近程调试要领,就只能不停的进行打日志,部署,打日志,部署这样低效的工作,而这对于程序员来说无异于折磨。所以明天小编就和大家一起来探索Flume的近程调试办法。

二、环境筹备

  1. flink官网下载上传服务器并解压。
  2. 开发自定义Source,这里以简略的读取mysql表数据为demo,局部代码如下:
package org.bigwinner.flume.sources;import org.apache.flume.Context;import org.apache.flume.Event;import org.apache.flume.EventDeliveryException;import org.apache.flume.PollableSource;import org.apache.flume.conf.Configurable;import org.apache.flume.event.EventBuilder;import org.apache.flume.source.AbstractSource;import org.slf4j.Logger;import org.slf4j.LoggerFactory;import java.sql.*;/** * @author: IT大狮兄 * @date: 2021/8/13 下午9:11 * @version: 1.0.0 * @description: 自定义Source--读取MySQL表的数据 */public class MysqlSource extends AbstractSource implements PollableSource, Configurable {    private static final Logger LOGGER = LoggerFactory.getLogger(MysqlSource.class);    private String mysqlUrl;    private String mysqlUser;    private String mysqlPassword;    private String mysqlTable;    private String mysqlDriver;    private Connection conn = null;    public Status process() throws EventDeliveryException {        String sql = "select * from " + mysqlTable;        try {            PreparedStatement statement = conn.prepareStatement(sql);            ResultSet resultSet = statement.executeQuery();            while (resultSet.next()) {                String id = resultSet.getString(1);                String uuid = resultSet.getString(2);                String iccid = resultSet.getString(3);                byte[] eventBytes = new StringBuilder().append(id).append("--")                        .append(uuid).append("--").append(iccid).toString().getBytes();                Event event = EventBuilder.withBody(eventBytes);                getChannelProcessor().processEvent(event);            }        } catch (SQLException throwables) {            throwables.printStackTrace();        }        return Status.READY;    }    public long getBackOffSleepIncrement() {        return 0;    }    public long getMaxBackOffSleepInterval() {        return 0;    }    @Override    /** Flume生命周期开始,能够做一些初始化的工作 */    public void start() {        LOGGER.info("Mysql source start......");        try {            Class.forName(mysqlDriver);            conn = DriverManager.getConnection(mysqlUrl, mysqlUser, mysqlPassword);        } catch (ClassNotFoundException e) {            LOGGER.error("Driver class is not found!");        } catch (SQLException throwables) {            LOGGER.error("get the connection error: {}", throwables);        }    }    @Override    /** Flume生命周期完结,能够做一些保留等完结前的工作 */    public void stop() {        LOGGER.info("Mysql source stop......");        if (conn != null) {            try {                conn.close();            } catch (SQLException throwables) {                LOGGER.error("连贯敞开异样: {}", throwables);            }        }        super.stop();    }    /** Flume配置文件读取的办法 */    public void configure(Context context) {        mysqlUrl = context.getString("mysql.url", "");        mysqlUser = context.getString("mysql.user", "");        mysqlPassword = context.getString("mysql.password", "");        mysqlTable = context.getString("mysql.table", "");        LOGGER.info("mysql_driver: {} --> mysql_url: {} --> mysql_user: {} --> mysql_password: {} --> mysql_table: {}",                mysqlDriver, mysqlUrl, mysqlUser, mysqlPassword, mysqlTable);    }}
  1. 编辑flume agent配置文件,并上传到flume conf目录下
a1.sources = s1a1.sinks = k1a1.channels = c1##############################          Source##############################自定义MySQL source类a1.sources.s1.type = org.bigwinner.flume.sources.MysqlSourcea1.sources.s1.mysql.driver = com.mysql.jdbc.Drivera1.sources.s1.mysql.url = jdbc:mysql://lsl001:3306/redis_tempa1.sources.s1.mysql.user = superboya1.sources.s1.mysql.password = iamsuperboya1.sources.s1.mysql.table = redis_temp##############################          Channel##############################配置file-channel数据管道a1.channels.c1.type = file#最小需要空间a1.channels.c1.minimumRequiredSpace = 3145728#最大文件大小a1.channels.c1.maxFileSize = 2146435071#flume事件指针检查点备份目录a1.channels.c1.checkpointDir = /opt/soft/flume/flume/data/checkpoint#file-channel对event备份到本地的文件目录a1.channels.c1.dataDirs = /opt/soft/flume/data/file-channel-mysql/data#文件管道中的数据容量,单位条数a1.channels.c1.capacity = 200#文件管道中的事务数据容量,单位条数a1.channels.c1.transactionCapacity = 100#检查点备份flume工夫指针的间隔时间a1.channels.c1.checkpointInterval=60000##############################          Sink##############################本次测试重点在Source,所以sink用null即可,示意不输入到任何中央a1.sinks.k1.type = nulla1.sources.s1.channels = c1a1.sinks.k1.channel = c1
  1. 打包,打成成宽依赖包,即蕴含所有依赖。并上传到flume的lib目录下

三、环境配置

  1. 服务器环境配置

    • 批改flume-ng启动命令文件: vim /opt/soft/flume/bin/flume-ng,批改为如下配置:

// 端口默认8000JAVA_OPTS="-Xmx500m -Xdebug -Xrunjdwp:transport=dt_socket,address=8000,server=y,suspend=y"
  • 如果flume-env.sh文件在应用,须要正文掉flume-env.sh的JAVA_OPTS配置:vim /opt/soft/flume/conf/flume-env.sh,没有应用则可疏忽。

  1. 本地IDE(本例以Idea为准)环境配置

    • 编辑配置界面,增加remote

  • 配置remote

四、验证

  1. 启动flume agent,后果如下图所示,即代表配置没有问题:
  2. 启动debug程序,查看是否失常debug:

由上,咱们看到程序正确的进入了断点,并查问到了mysql的记录。

五、总结

以上就是明天和大家分享的Flume的近程调试办法,如果不晓得的小伙伴连忙实际起来吧,晋升效率,珍视本人!

案例代码参考:flume_demo