前往小程序,Get更优阅读体验!
立即前往
首页
学习
活动
专区
工具
TVP
发布
社区首页 >专栏 >如何开发HBase Endpoint类型的Coprocessor以及部署使用

如何开发HBase Endpoint类型的Coprocessor以及部署使用

作者头像
Fayson
发布2018-07-11 17:56:29
1.8K0
发布2018-07-11 17:56:29
举报
文章被收录于专栏:Hadoop实操

温馨提示:要看高清无码套图,请使用手机打开并单击图片放大查看。

Fayson的github:https://github.com/fayson/cdhproject

提示:代码块部分可以左右滑动查看噢

1.文档编写目的


Fayson在前面的文章介绍了HBase自带的Coprocessor调用示例《如何使用Java调用HBase的 Endpoint Coprocessor》,本篇文章Fayson主要介绍如何开发一个HBase Endpoint类型的协处理器。

本篇文章示例协处理器主要实现了对列的Count、Max、Min、Sum以及Average。前面的文章调用Coprocessor定义的全局的,在本篇文章Fayson介绍另一种实现方式通过代码的方式对指定的表添加Coprocessor。

  • 内容概述

1.环境准备

2.使用Protobuf生成序列化类

3.Endpoint Coprocessor服务端实现

4.Endpoint Coprocessor客户端实现

5.部署及调用

  • 测试环境

1.CM和CDH版本为5.14.3

2.集群未启用Kerberos

2.环境准备


在HMaster、RegionServer内部,创建了RpcServer实例,并与Client三者之间实现了Rpc调用,在HBase0.95版本引入了Google-Protobuf作为中间数据组织方式,并在Protobuf提供的Rpc接口之上,实现了基于服务的Rpc实现。

Protobuf Buffers是一种轻便高效的结构化数据存储格式,可以用于数据序列化。适合做数据存储或RPC数据交换格式。用于通讯协议、数据存储等领域的语言无关、平台无关、可扩展的序列化结构数据格式。

这里Fayson借助于Protobuf来生成HBase RPC数据交换格式类,在HBase中使用的Protobuf版本为2.5.0,所以选择安装相同版本的Protobuf。

1.下载Protobuf2.5.0版本的安装包,地址如下:

代码语言:javascript
复制
https://github.com/google/protobuf/releases/download/v2.5.0/protobuf-2.5.0.tar.gz

(可左右滑动)

2.选择一台服务器安装Protobuf

代码语言:javascript
复制
[root@ip-172-31-5-38 ~]# wget https://github.com/google/protobuf/releases/download/v2.5.0/protobuf-2.5.0.tar.gz

(可左右滑动)

3.执行如下命令安装Protobuf所需要的依赖包

代码语言:javascript
复制
yum install -y autoconf automake libtool curl make g++ unzip gcc-c++

(可左右滑动)

4.解压protobuf-2.5.0.tar.gz,并进入解压目录执行如下命令编译安装

代码语言:javascript
复制
[root@ip-172-31-5-38 ~]# tar -zxvf protobuf-2.5.0.tar.gz
[root@ip-172-31-5-38 ~]# cd protobuf-2.5.0
[root@ip-172-31-5-38 protobuf-2.5.0]# ./configure --prefix=/usr/local/protobuf
[root@ip-172-31-5-38 protobuf-2.5.0]# make && make install

(可左右滑动)

5.配置Protobuf环境变量

代码语言:javascript
复制
export PROTOBUF_HOME=/usr/local/protobuf
export PATH=$PROTOBUF_HOME/bin:$PATH

(可左右滑动)

执行命令使环境变量生效

代码语言:javascript
复制
[root@ip-172-31-5-38 protobuf-2.5.0]# source /etc/profile

(可左右滑动)

6.准备HBase测试表,建表脚本及测试数据如下

代码语言:javascript
复制
create 'fayson_coprocessor', {NAME => 'info'}
put 'fayson_coprocessor','001','info:sales',12.3
put 'fayson_coprocessor','002','info:sales',24.5
put 'fayson_coprocessor','003','info:sales',10.5
put 'fayson_coprocessor','004','info:sales',11.5
put 'fayson_coprocessor','005','info:sales',10.5
put 'fayson_coprocessor','001','info:age',22
put 'fayson_coprocessor','002','info:age',33
put 'fayson_coprocessor','003','info:age',26
put 'fayson_coprocessor','004','info:age',28
put 'fayson_coprocessor','005','info:age',56

(可左右滑动)

3.使用Protobuf生成序列化类


1.准备MyFirstCoprocessor.proto文件,内容如下

代码语言:javascript
复制
[root@ip-172-31-5-171 hbase-coprocessor]# vim MyFirstCoprocessor.proto 
syntax = "proto2";
option java_package = "com.cloudera.hbase.coprocessor.server";
option java_outer_classname = "MyFirstCoprocessor";
option java_generic_services = true;
option java_generate_equals_and_hash = true;
option optimize_for = SPEED;
message MyCoprocessRequest {
    required string family = 1;
    required string columns = 2;
}
message MyCoprocessResponse {
    required int64 count = 1;
    required double maxnum = 3;
    required double minnum = 4;
    required double sumnum = 5;
}
service AggregationService {
  rpc getAggregation(MyCoprocessRequest)
    returns (MyCoprocessResponse);
}

(可左右滑动)

2.在命令行执行如下命令生成Java类

代码语言:javascript
复制
[root@ip-172-31-5-38 hbase-coprocessor]# protoc --java_out=./ MyFirstCoprocessor.proto 
[root@ip-172-31-5-38 hbase-coprocessor]# ll
total 4
drwxr-xr-x 3 root root  22 May 14 16:34 com
-rw-r--r-- 1 root root 609 May 14 16:33 MyFirstCoprocessor.proto
[root@ip-172-31-5-38 hbase-coprocessor]# 

(可左右滑动)

在当前目录下根据java_package指定的目录生成Java类。

4.Endpoint Coprocessor服务端实现


1.使用Maven创建Java示例工程,pom.xml文件内容如下

代码语言:javascript
复制
<dependency>
    <groupId>org.apache.hadoop</groupId>
    <artifactId>hadoop-client</artifactId>
    <version>2.6.0-cdh5.11.2</version>
</dependency>
<dependency>
    <groupId>org.apache.hadoop</groupId>
    <artifactId>hadoop-common</artifactId>
    <version>2.6.0-cdh5.11.2</version>
</dependency>
<dependency>
    <groupId>org.apache.hbase</groupId>
    <artifactId>hbase-client</artifactId>
    <version>1.2.0-cdh5.11.2</version>
</dependency>
<dependency>
    <groupId>org.apache.hbase</groupId>
    <artifactId>hbase-examples</artifactId>
    <version>1.2.0-cdh5.11.2</version>
</dependency>
<dependency>
    <groupId>com.google.protobuf</groupId>
    <artifactId>protobuf-java</artifactId>
    <version>2.5.0</version>
</dependency>

(可左右滑动)

2.将Protobuf生成的java类拷贝至指定的包目录下

与MyFirstCoprocessor.proto文件指定的java_package包目录一致。

3.在com.cloudera.hbase.coprocessor.server包下新建MyFirstCoprocessorEndpoint实现类,内容如下

代码语言:javascript
复制
package com.cloudera.hbase.coprocessor.server;
import com.google.protobuf.RpcCallback;
import com.google.protobuf.RpcController;
import com.google.protobuf.Service;
import org.apache.commons.collections.map.HashedMap;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.hbase.Cell;
import org.apache.hadoop.hbase.CellUtil;
import org.apache.hadoop.hbase.Coprocessor;
import org.apache.hadoop.hbase.CoprocessorEnvironment;
import org.apache.hadoop.hbase.client.Scan;
import org.apache.hadoop.hbase.coprocessor.CoprocessorException;
import org.apache.hadoop.hbase.coprocessor.CoprocessorService;
import org.apache.hadoop.hbase.coprocessor.RegionCoprocessorEnvironment;
import org.apache.hadoop.hbase.protobuf.ResponseConverter;
import org.apache.hadoop.hbase.regionserver.InternalScanner;
import org.apache.hadoop.hbase.util.Bytes;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
/**
 * package: com.cloudera.hbase.coprocessor.server
 * describe: HBase RegionServer上Endpoint Coprocessor实现,主要实现对指定列的Count、MAX、MIN、SUM聚合操作
 * creat_user: Fayson
 * email: htechinfo@163.com
 * creat_date: 2018/5/13
 * creat_time: 下午11:11
 * 公众号:Hadoop实操
 */
public class MyFirstCoprocessorEndPoint extends MyFirstCoprocessor.AggregationService implements Coprocessor, CoprocessorService {
    protected static final Log log = LogFactory.getLog(MyFirstCoprocessorEndPoint.class);
    private RegionCoprocessorEnvironment env;
    @Override
    public void getAggregation(RpcController controller, MyFirstCoprocessor.MyCoprocessRequest request, RpcCallback<MyFirstCoprocessor.MyCoprocessResponse> done) {
        Scan scan = new Scan();
        scan.addFamily(Bytes.toBytes(request.getFamily()));
        //传入列的方式   sales:MAX,sales:MIN,sales:AVG,slaes:SUM,sales:COUNT
        String colums = request.getColumns();
        //记录所有要扫描的列
        Map<String, List<String>> columnMaps = new HashedMap();
        for (String columnAndType : colums.split(",")) {
            String column = columnAndType.split(":")[0];
            String type = columnAndType.split(":")[1];
            List<String> typeList = null;
            if (columnMaps.containsKey(column)) {
                typeList = columnMaps.get(column);
            } else {
                typeList = new ArrayList<>();
                //将column添加到Scan中
                scan.addColumn(Bytes.toBytes(request.getFamily()), Bytes.toBytes(column));
            }
            typeList.add(type);
            columnMaps.put(column, typeList);
        }
        InternalScanner scanner = null;
        MyFirstCoprocessor.MyCoprocessResponse response = null;
        Double max = null;
        Double min = null;
        Double sumVal = null;
        long counter = 0L;
        try {
            scanner = this.env.getRegion().getScanner(scan);
            List<Cell> results = new ArrayList<>();
            boolean hasMore = false;
            scanner = env.getRegion().getScanner(scan);
            do {
                hasMore = scanner.next(results);
                if (results.size() > 0) {
                    ++counter;
                }
                log.info("counter:" + counter);
                log.info("results size:" + results.size());
                for (Cell cell : results) {
                    String column = Bytes.toString(CellUtil.cloneQualifier(cell));
                    log.info("Column Name: " + column);
                    log.info("Cell Value:" + new String(CellUtil.cloneValue(cell)));
                    Double temp = Double.parseDouble(new String(CellUtil.cloneValue(cell)));
                    if (columnMaps.containsKey(column)) {
                        List<String> types = columnMaps.get(column);
                        for (String type : types) {
                            switch (type.toUpperCase()) {
                                case "MIN":
                                    min = min != null && (temp == null || compare(temp, min) >= 0) ? min : temp;
                                    log.info("MIN Value: " + min.doubleValue());
                                    break;
                                case "MAX":
                                    max = max != null && (temp == null || compare(temp, max) <= 0) ? max : temp;
                                    break;
                                case "SUM":
                                    if (temp != null) {
                                        sumVal = add(sumVal, temp);
                                    }
                                    break;
                                default:
                                    break;
                            }
                        }
                    }
                }
                results.clear();
            } while (hasMore);
            response = MyFirstCoprocessor.MyCoprocessResponse.newBuilder()
                    .setMaxnum(max!=null?max.doubleValue():Double.MAX_VALUE)
                    .setMinnum(min!=null?min.doubleValue():Double.MIN_NORMAL)
                    .setCount(counter)
                    .setSumnum(sumVal!=null?sumVal.doubleValue():Double.MIN_NORMAL).build();
        } catch (IOException e) {
            e.printStackTrace();
            ResponseConverter.setControllerException(controller, e);
        } finally {
            if (scanner != null) {
                try {
                    scanner.close();
                } catch (IOException e) {
                    e.printStackTrace();
                }
            }
        }
        done.run(response);
    }
    public static int compare(Double l1, Double l2) {
        if (l1 == null ^ l2 == null) {
            return l1 == null ? -1 : 1; // either of one is null.
        } else if (l1 == null)
            return 0; // both are null
        return l1.compareTo(l2); // natural ordering.
    }
    public double divideForAvg(Double d1, Long l2) {
        return l2 != null && d1 != null?d1.doubleValue() / l2.doubleValue():0.0D / 0.0;
    }
    public Double add(Double d1, Double d2) {
        return d1 != null && d2 != null ? Double.valueOf(d1.doubleValue() + d2.doubleValue()) : (d1 == null ? d2 : d1);
    }
    @Override
    public void start(CoprocessorEnvironment coprocessorEnvironment) throws IOException {
        if (coprocessorEnvironment instanceof RegionCoprocessorEnvironment) {
            this.env = (RegionCoprocessorEnvironment) coprocessorEnvironment;
        } else {
            throw new CoprocessorException("Must be loaded on a table region!");
        }
    }
    @Override
    public void stop(CoprocessorEnvironment coprocessorEnvironment) throws IOException {
    }
    @Override
    public Service getService() {
        return this;
    }
}

(可左右滑动)

5.Endpoint Coprocessor客户端实现


1.编写MyFirstCoprocessExample.java类,代码如下:

代码语言:javascript
复制
package com.cloudera.hbase.coprocessor.client;
import com.cloudera.hbase.coprocessor.server.MyFirstCoprocessor;
import com.cloudera.hbase.coprocessor.server.MyFirstCoprocessorEndPoint;
import com.google.common.util.concurrent.AtomicDouble;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hbase.Coprocessor;
import org.apache.hadoop.hbase.HBaseConfiguration;
import org.apache.hadoop.hbase.HTableDescriptor;
import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.client.Connection;
import org.apache.hadoop.hbase.client.ConnectionFactory;
import org.apache.hadoop.hbase.client.Table;
import org.apache.hadoop.hbase.client.coprocessor.Batch;
import org.apache.hadoop.hbase.client.coprocessor.DoubleColumnInterpreter;
import org.apache.hadoop.hbase.ipc.BlockingRpcCallback;
import java.io.IOException;
import java.util.concurrent.atomic.AtomicLong;
/**
 * package: com.cloudera.hbase.coprocessor.client
 * describe: 调用HBase RegionServer端的协处理器
 * creat_user: Fayson
 * email: htechinfo@163.com
 * creat_date: 2018/5/14
 * creat_time: 下午6:36
 * 公众号:Hadoop实操
 */
public class MyFirstCoprocessExample {
    public static void main(String[] args) {
        String table_name = "fayson_coprocessor";
        //初始化HBase配置
        Configuration configuration = HBaseConfiguration.create();
        configuration.set("hbase.zookeeper.property.clientPort", "2181");
        configuration.setStrings("hbase.zookeeper.quorum", "ip-172-31-5-38.ap-southeast-1.compute.internal,ip-172-31-8-230.ap-southeast-1.compute.internal,ip-172-31-5-171.ap-southeast-1.compute.internal");
        try {
            //创建一个HBase的Connection
            Connection connection = ConnectionFactory.createConnection(configuration);
            TableName tableName = TableName.valueOf(table_name);
            if(!connection.getAdmin().tableExists(tableName)) {
                System.out.println(table_name + "does not exist....");
                System.exit(0);
            }
            Table table = connection.getTable(tableName);
            //删除表上的协处理器
            deleteCoprocessor(connection, table, MyFirstCoprocessorEndPoint.class);
            //为指定的表添加协处理器
            String hdfspath = "hdfs://nameservice3/hbase/coprocessor/hbase-demo-1.0-SNAPSHOT.jar";
            setupToExistTable(connection, table, hdfspath, MyFirstCoprocessorEndPoint.class);
            //客户端调用Region端的协处理器
            execFastEndpointCoprocessor(table, "info", "sales:MAX,sales:MIN,sales:AVG,sales:SUM,sales:COUNT");
            //关闭连接
            connection.close();
        } catch (IOException e) {
            e.printStackTrace();
        }
    }
    /**
     * 删除HBase表上的协处理器
     * @param connection
     * @param table
     * @param cls
     */
    public static void deleteCoprocessor(Connection connection, Table table, Class<?>... cls) {
        System.out.println("begin delete " + table.getName().toString() + " Coprocessor......");
        try {
            HTableDescriptor hTableDescriptor = table.getTableDescriptor();
            for(Class cass : cls) {
                hTableDescriptor.removeCoprocessor(cass.getCanonicalName());
            }
            connection.getAdmin().modifyTable(table.getName(), hTableDescriptor);
        } catch (IOException e) {
            e.printStackTrace();
        }
        System.out.println("end delete " + table.getName().toString() + " Coprocessor......");
    }
    /**
     *
     * @param connection
     * @param table
     * @param jarPath
     * @param cls
     */
    public static void setupToExistTable(Connection connection, Table table, String jarPath, Class<?>... cls) {
        try {
            if(jarPath != null && !jarPath.isEmpty()) {
                Path path = new Path(jarPath);
                HTableDescriptor hTableDescriptor = table.getTableDescriptor();
                for(Class cass : cls) {
                    hTableDescriptor.addCoprocessor(cass.getCanonicalName(), path, Coprocessor.PRIORITY_USER, null);
                }
                connection.getAdmin().modifyTable(table.getName(), hTableDescriptor);
            }
        } catch (IOException e) {
            e.printStackTrace();
        }
    }
    /**
     * 效率最高的方式,在方式二的基础上优化
     * 通过HBase的coprocessorService(Class, byte[],byte[],Batch.Call,Callback<R>)方法获取表的总条数
     * @param table HBase表名
     * @return 返回表的总条数
     */
    public static long execFastEndpointCoprocessor(Table table, String family, String columns) {
        long start_t = System.currentTimeMillis();
        //定义总的 rowCount 变量
        AtomicLong totalRowCount = new AtomicLong();
        AtomicDouble maxValue = new AtomicDouble(Double.MIN_VALUE);
        AtomicDouble minValue = new AtomicDouble(Double.MAX_VALUE);
        AtomicDouble sumValue = new AtomicDouble();
        try {
            Batch.Callback<MyFirstCoprocessor.MyCoprocessResponse> callback = new Batch.Callback<MyFirstCoprocessor.MyCoprocessResponse>() {
                @Override
                public void update(byte[] bytes, byte[] bytes1, MyFirstCoprocessor.MyCoprocessResponse myCoprocessResponse) {
                    //更新Count值
                    totalRowCount.getAndAdd(myCoprocessResponse.getCount());
                    //更新最大值
                    if(myCoprocessResponse.getMaxnum() > maxValue.doubleValue()) {
                        maxValue.compareAndSet(maxValue.doubleValue(), myCoprocessResponse.getMaxnum());
                    }
                    //更新最小值
                    if(myCoprocessResponse.getMinnum() < minValue.doubleValue()) {
                        minValue.compareAndSet(minValue.doubleValue(), myCoprocessResponse.getMinnum());
                    }
                    //更新求和
                    sumValue.getAndAdd(myCoprocessResponse.getSumnum());
                }
            };
            table.coprocessorService(MyFirstCoprocessor.AggregationService.class, null, null, new Batch.Call<MyFirstCoprocessor.AggregationService, MyFirstCoprocessor.MyCoprocessResponse>() {
                @Override
                public MyFirstCoprocessor.MyCoprocessResponse call(MyFirstCoprocessor.AggregationService aggregationService) throws IOException {
                    MyFirstCoprocessor.MyCoprocessRequest requet = MyFirstCoprocessor.MyCoprocessRequest.newBuilder().setFamily(family).setColumns(columns).build();
                    BlockingRpcCallback<MyFirstCoprocessor.MyCoprocessResponse> rpcCallback = new BlockingRpcCallback<>();
                    aggregationService.getAggregation(null, requet, rpcCallback);
                    MyFirstCoprocessor.MyCoprocessResponse response = rpcCallback.get();
                    return response;
                }
            }, callback);
        } catch (Throwable throwable) {
            throwable.printStackTrace();
        }
        System.out.println("耗时:" + (System.currentTimeMillis() - start_t));
        System.out.println("totalRowCount:" + totalRowCount.longValue());
        System.out.println("maxValue:" + maxValue.doubleValue());
        System.out.println("minValue:" + minValue.doubleValue());
        System.out.println("sumValue:" + sumValue.doubleValue());
        System.out.println("avg:" + new DoubleColumnInterpreter().divideForAvg(sumValue.doubleValue(), totalRowCount.longValue()));
        return totalRowCount.longValue();
    }
}

(可左右滑动)

6.部署及调用


1.使用mvn编译工程

代码语言:javascript
复制
mvn clean package

(可左右滑动)

2.将编译好的jar包,上传HDFS的/hbase/coprocessor目录下

代码语言:javascript
复制
[root@ip-172-31-5-38 ~]# export HADOOP_USER_NAME=hbase         
[root@ip-172-31-5-38 ~]# hadoop fs -mkdir -p /hbase/coprocessor
[root@ip-172-31-5-38 ~]# hadoop fs -put hbase-demo-1.0-SNAPSHOT.jar /hbase/coprocessor
[root@ip-172-31-5-38 ~]# hadoop fs -ls /hbase/coprocessor

(可左右滑动)

在客户端调用的示例代码中使用的是代码为指定的表添加Coprocessor操作,所以这里不需要在HBase中配置全局的Coprocessor。

3.运行MyFirstCoprocessorExample代码,查看运行结果

统计的值与我们写入的数据一致。

7.总结


  • 在开发HBase的Coprocessor借助于Protobuf生成RPC请求数据交互类,我们只需要在生成的类基础上实现业务即可。
  • 本篇文章主要介绍了怎么样通过代码的方式为指定的HBase表添加Coprocessor,这种方式使用更灵活,不需要重启HBase服务。
  • 将编写好的Coprocessor jar上传至HDFS,确保文件的目录属主。
  • HBase自带的也有AggregateImplementation类实现列的聚合,原生的不能同时对多个列进行聚合处理,如果需要多次聚合则需要多次调用RPC请求,HBase数据在不断的写入会出现每次聚合的结果有偏差,本示例将聚合放在一个RPC中处理可以减少RPC的请求次数并确保查询条件相同的情况下不会出现数据不一致问题。

GitHub地址:

https://github.com/fayson/cdhproject/blob/master/hbasedemo/proto/MyFirstCoprocessor.proto

https://github.com/fayson/cdhproject/blob/master/hbasedemo/src/main/java/com/cloudera/hbase/coprocessor/server/MyFirstCoprocessorEndPoint.java

https://github.com/fayson/cdhproject/blob/master/hbasedemo/src/main/java/com/cloudera/hbase/coprocessor/server/MyFirstCoprocessor.java

https://github.com/fayson/cdhproject/blob/master/hbasedemo/src/main/java/com/cloudera/hbase/coprocessor/client/MyFirstCoprocessExample.java

提示:代码块部分可以左右滑动查看噢

为天地立心,为生民立命,为往圣继绝学,为万世开太平。 温馨提示:要看高清无码套图,请使用手机打开并单击图片放大查看。

推荐关注Hadoop实操,第一时间,分享更多Hadoop干货,欢迎转发和分享。

原创文章,欢迎转载,转载请注明:转载自微信公众号Hadoop实操

本文参与 腾讯云自媒体同步曝光计划,分享自微信公众号。
原始发表:2018-05-15,如有侵权请联系 cloudcommunity@tencent.com 删除

本文分享自 Hadoop实操 微信公众号,前往查看

如有侵权,请联系 cloudcommunity@tencent.com 删除。

本文参与 腾讯云自媒体同步曝光计划  ,欢迎热爱写作的你一起参与!

评论
登录后参与评论
0 条评论
热度
最新
推荐阅读
相关产品与服务
领券
问题归档专栏文章快讯文章归档关键词归档开发者手册归档开发者手册 Section 归档