[ 
https://issues.apache.org/jira/browse/FLINK-8101?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16347610#comment-16347610
 ] 

ASF GitHub Bot commented on FLINK-8101:
---------------------------------------

Github user cjolif commented on a diff in the pull request:

    https://github.com/apache/flink/pull/5374#discussion_r165186422
  
    --- Diff: 
flink-connectors/flink-connector-elasticsearch5.3/src/main/java/org/apache/flink/streaming/connectors/elasticsearch53/BulkProcessorIndexer.java
 ---
    @@ -0,0 +1,57 @@
    +/*
    + * Licensed to the Apache Software Foundation (ASF) under one
    + * or more contributor license agreements.  See the NOTICE file
    + * distributed with this work for additional information
    + * regarding copyright ownership.  The ASF licenses this file
    + * to you under the Apache License, Version 2.0 (the
    + * "License"); you may not use this file except in compliance
    + * with the License.  You may obtain a copy of the License at
    + *
    + *    http://www.apache.org/licenses/LICENSE-2.0
    + *
    + * Unless required by applicable law or agreed to in writing, software
    + * distributed under the License is distributed on an "AS IS" BASIS,
    + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    + * See the License for the specific language governing permissions and
    + * limitations under the License.
    + */
    +
    +package org.apache.flink.streaming.connectors.elasticsearch53;
    +
    +import org.apache.flink.streaming.connectors.elasticsearch.RequestIndexer;
    +
    +import org.elasticsearch.action.ActionRequest;
    +import org.elasticsearch.action.DocWriteRequest;
    +import org.elasticsearch.action.bulk.BulkProcessor;
    +
    +import java.util.concurrent.atomic.AtomicLong;
    +
    +/**
    + * Implementation of a {@link RequestIndexer}, using a {@link 
BulkProcessor}.
    + * {@link ActionRequest ActionRequests} will be converted to {@link 
DocWriteRequest}
    + * and will be buffered before sending a bulk request to the Elasticsearch 
cluster.
    + */
    +public class BulkProcessorIndexer implements RequestIndexer {
    +
    +   private final BulkProcessor bulkProcessor;
    +   private final boolean flushOnCheckpoint;
    +   private final AtomicLong numPendingRequestsRef;
    +
    +   public BulkProcessorIndexer(BulkProcessor bulkProcessor,
    +                                                           boolean 
flushOnCheckpoint,
    +                                                           AtomicLong 
numPendingRequests) {
    +           this.bulkProcessor = bulkProcessor;
    +           this.flushOnCheckpoint = flushOnCheckpoint;
    +           this.numPendingRequestsRef = numPendingRequests;
    +   }
    +
    +   @Override
    +   public void add(ActionRequest... actionRequests) {
    +           for (ActionRequest actionRequest : actionRequests) {
    +                   if (flushOnCheckpoint) {
    +                           numPendingRequestsRef.getAndIncrement();
    +                   }
    +                   this.bulkProcessor.add((DocWriteRequest) actionRequest);
    --- End diff --
    
    This is actually from the commit I brought into the PR from orignal 
@zjureel's PR. That said I think the answer is definitely yes in the case that 
matters for Flink. Indeed:
    
    * The ActionRequest values here are actually coming from the implementation 
of the `ElasticsearchSinkFunction.process` method which should create 
`ActionRequest` and add them to the indexer.
    * The idea here is not to create any sort of `ActionRequest` you would 
possibly dream of but indexing requests?
    * The way to create `ActionRequest` for indexing in Elasticsearch is to use 
`org.elasticsearch.action.index.IndexRequest` 
    * starting with Elasticsearch 5.3 IndexRequest inherits from 
`DocWriteRequest` while it was not before 5.3.
    
    See: 
    
    
![image](https://user-images.githubusercontent.com/623171/35646706-5723ab78-06d0-11e8-8d50-5b4545047a1f.png)
    
    vs
    
    
![image](https://user-images.githubusercontent.com/623171/35646719-63d7f1b2-06d0-11e8-8308-c330b3c11dad.png)
    
    So the only case I see where this could not be a `DocWriteRequest` would be 
if someone in the `ElasticsearchSinkFunction` would create something else than 
an index request. But I don't really see why? 
    
    That said this raises the question of why from the origin the API was not 
typed against `IndexRequest` instead of `ActionRequest` as this would avoid 
those questions and force the user to return a `IndexRequest`?
    
    In every case there is little choice because starting with 5.3 
Elasticsearch does not accept ActionRequest in BulkProcessor anymore but just 
IndexRequest/DocWriteRequest.
    
    Do you have a suggestion on how to handle this better? Obviously I can add 
documentation saying starting with 5.3 the sink function MUST return 
DocWriteRequest? But is that enough for you?



> Elasticsearch 6.x support
> -------------------------
>
>                 Key: FLINK-8101
>                 URL: https://issues.apache.org/jira/browse/FLINK-8101
>             Project: Flink
>          Issue Type: New Feature
>          Components: ElasticSearch Connector
>    Affects Versions: 1.4.0
>            Reporter: Hai Zhou UTC+8
>            Assignee: Flavio Pompermaier
>            Priority: Major
>             Fix For: 1.5.0
>
>
> Recently, elasticsearch 6.0.0 was released: 
> https://www.elastic.co/blog/elasticsearch-6-0-0-released  
> The minimum version of ES6 compatible Elasticsearch Java Client is 5.6.0



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to