前言

批量写入到Elasticsearch会提高写入性能,减少Elasticsearch io压力。


一、Elasticsearch是什么?

Elasticsearch是一个实时的分布式开放源代码全文本搜索和分析引擎。可从RESTful Web服务界面访问它,并使用无模式的JSON(JavaScript对象表示法)文档存储数据。它基于Java编程语言构建,因此Elasticsearch可以在不同平台上运行。它使用户能够以很高的速度浏览大量的数据。

二、实现步骤

1.创建BulkProcessor

BulkProcessor是一个线程安全的批量处理类,允许方便地设置每次写入ES的最大数量,以及超时时间。所谓超时时间,就是在规定的时间内,如果没有请求进来,就把之前累积的请求直接写到ES,不必等待请求数量累积到你规定的最大数量。
代码如下(示例):


import org.elasticsearch.action.ActionRequest;
import org.elasticsearch.action.DocWriteRequest;
import org.elasticsearch.action.bulk.BulkItemResponse;
import org.elasticsearch.action.bulk.BulkProcessor;
import org.elasticsearch.action.bulk.BulkRequest;
import org.elasticsearch.action.bulk.BulkResponse;
import org.elasticsearch.rest.RestStatus;

public class BulkProListener implements BulkProcessor.Listener{
    @Override
    public void beforeBulk(long l, BulkRequest bulkRequest) {
        System.out.println("执行前");
    }

    @Override
    public void afterBulk(long l, BulkRequest request, BulkResponse response) {
        System.out.println("执行后");
        if (response.hasFailures()) {
            BulkItemResponse itemResponse;
            Throwable failure;
            RestStatus restStatus;
            DocWriteRequest actionRequest;

            try {
                for (int i = 0; i < response.getItems().length; i++) {
                    itemResponse = response.getItems()[i];
                    if (itemResponse.isFailed()) {
                        failure = itemResponse.getFailure().getCause();
                        if (failure != null) {
                            restStatus = itemResponse.getFailure().getStatus();
                            actionRequest = request.requests().get(i);
                            if (restStatus == null) {
                                if (actionRequest instanceof ActionRequest) {
                                    System.out.println("Failed Elasticsearch item request:  " + failure.getCause().getMessage());
                                } else {
                                    throw new UnsupportedOperationException(
                                            "The sink currently only supports ActionRequests");
                                }
                            }else{
                                if (actionRequest instanceof ActionRequest) {
                                    System.out.println("Failed sink item request:  " + failure.getCause().getMessage()+" status: "+restStatus.getStatus());
                                    failure.printStackTrace();
                                } else {
                                    throw new UnsupportedOperationException(
                                            "The sink currently only supports ActionRequests");
                                }
                            }
                        }
                    }

                }
            }catch (Throwable t){
                t.printStackTrace();
            }
        }
    }

    @Override
    public void afterBulk(long l, BulkRequest request, Throwable failure) {
        System.out.println("有错误");
        try {
            for (DocWriteRequest writeRequest : request.requests()) {
                if (writeRequest instanceof ActionRequest) {
                    System.out.println("Failed Elasticsearch item request:  " + failure.getMessage());
                    failure.printStackTrace();
                } else {
                    throw new UnsupportedOperationException(
                            "The sink currently only supports ActionRequests");
                }
            }
        } catch (Throwable t) {
            // fail the sink and skip the rest of the items
            // if the failure handler decides to throw an exception
            t.printStackTrace();
        }
    }
}

2.批量写入数据

在es中建立索引batch,类型my_type,结构为"user_name",“user_id”,“age”,“user_note”
//设置满5000条提交,时间间隔10秒
bulkProcessor.setBulkActions(5000).setFlushInterval(TimeValue.timeValueSeconds(10)).build();
代码如下(示例):


import com.alibaba.fastjson.JSONObject;
import org.elasticsearch.action.bulk.BulkProcessor;
import org.elasticsearch.action.bulk.BulkRequestBuilder;
import org.elasticsearch.action.index.IndexRequestBuilder;
import org.elasticsearch.client.transport.TransportClient;
import org.elasticsearch.common.network.NetworkModule;
import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.common.transport.InetSocketTransportAddress;
import org.elasticsearch.common.unit.TimeValue;
import org.elasticsearch.transport.Netty3Plugin;
import org.elasticsearch.transport.client.PreBuiltTransportClient;

import java.net.InetAddress;
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.TimeUnit;

public class EsBatchWriterTest {
    public static void main(String[] args) throws  Exception {
        Settings settings = Settings.builder().put(NetworkModule.HTTP_TYPE_KEY, Netty3Plugin.NETTY_HTTP_TRANSPORT_NAME)
                .put(NetworkModule.TRANSPORT_TYPE_KEY, Netty3Plugin.NETTY_TRANSPORT_NAME).build();
//            Settings settings = Settings.EMPTY;
        //创建client
        TransportClient client = new PreBuiltTransportClient(settings)
                .addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("10.68.8.60"), 9300));
        batch2(client);
//        update(client);
        client.close();
    }
    public static void batch2(TransportClient client) throws InterruptedException {
        BulkProcessor.Builder bulkProcessor = BulkProcessor.builder(
                client,new BulkProListener());
        BulkProcessor processor = bulkProcessor.setBulkActions(5000).setFlushInterval(TimeValue.timeValueSeconds(10)).build();

        int count = 1;
        List<JSONObject> list = getData();
        System.out.println(list.size());
        for(JSONObject obj:list) {
            System.out.println(obj.toJSONString());
            IndexRequestBuilder builder = client.prepareIndex("batch", "my_type").setId(obj.getString("user_id")).setSource(obj);
            processor.add(builder.request());
        }
        processor.awaitClose(2, TimeUnit.MINUTES);
//        processor.close();
    }
    private static List<JSONObject> getData(){
        List<JSONObject> list =new ArrayList<>();
        JSONObject j=new JSONObject();
        j.put("user_name","name7");
        j.put("user_id","7");
        j.put("age","34");
        j.put("user_note","note");
        list.add(j);

        j=new JSONObject();
        j.put("user_name","name8");
        j.put("user_id","8");
        j.put("age","24");
        j.put("user_note","note");
        list.add(j);
        j=new JSONObject();
        j.put("user_name","name9");
        j.put("user_id","9");
        j.put("age","24");
        j.put("user_note","note");
        list.add(j);
        j=new JSONObject();
        j.put("user_name","name10");
        j.put("user_id","10");
        j.put("age","14");
        j.put("user_note","note");
        list.add(j);
        j=new JSONObject();
        j.put("user_name","name11");
        j.put("user_id","11");
        j.put("age","54b");
        j.put("user_note","note");
        list.add(j);

        j=new JSONObject();
        j.put("user_name","name20");
        j.put("user_id","20");
        j.put("age","34a");
        j.put("user_note","note");
        list.add(j);
        j=new JSONObject();
        j.put("user_name","name30");
        j.put("user_id","30");
        j.put("age","30");
        j.put("user_note","note");
        list.add(j);
        return list;
    }
    public static void batch(TransportClient client){
        int count = 1;
        //开启批量插入
        BulkRequestBuilder bulkRequest = client.prepareBulk();
        List<JSONObject> list =new ArrayList<>();

        JSONObject j=new JSONObject();
        j.put("user_name","name1");
        j.put("user_id","1");
        list.add(j);
        j=new JSONObject();
        j.put("user_name","name3");
        j.put("user_id","3");
        list.add(j);
        j=new JSONObject();
        j.put("user_name","name2");
        j.put("user_id","2");
        list.add(j);
        for(JSONObject obj:list){
            IndexRequestBuilder builder = client.prepareIndex("batch", "my_type").setId(obj.getString("user_id")).setSource(obj);
            bulkRequest.add(builder);
            //每一千条提交一次
            if (count% 1000==0) {
                bulkRequest.execute().actionGet();
                System.out.println("提交了:" + count);
            }
            count++;
        }
        bulkRequest.execute().actionGet();
    }
}


总结

执行文档批量请求时,首先需要初始化 Elasticsearch Client,其次创建 BulkProcessor ,
设置 BulkProcessor 参数,最后关闭processor。

01-05 10:02