【路径2】query分片查询返回InitialSearchPhase#maybeFork-CSDN博客

本文链接：https://blog.csdn.net/weixin_41871193/article/details/124765218

1.AbstractRunnable
run:38, AbstractRunnable (org.elasticsearch.common.util.concurrent)
doRun:638, ThreadContext$ContextPreservingAbstractRunnable (org.elasticsearch.common.util.concurrent)
2.SearchService
doRun:1000, SearchService$3 (org.elasticsearch.search)
onResponse:271, SearchService$2 (org.elasticsearch.search)
onResponse:278, SearchService$2 (org.elasticsearch.search)
3.SearchTransportService
onResponse:364, SearchTransportService$6$1 (org.elasticsearch.action.search)
onResponse:368, SearchTransportService$6$1 (org.elasticsearch.action.search)
4.RequestHandlerRegistry

sendResponse:108, RequestHandlerRegistry$TransportChannelWrapper (org.elasticsearch.transport)

5.DelegatingTransportChannel
sendResponse:69, DelegatingTransportChannel (org.elasticsearch.transport)
6.TransportService
sendResponse:1135, TransportService $D i r e c t R e s p o n s e C h a n n e l (o r g . e l a s t i c s e a r c h . t r a n s p o r t) s e n d R e s p o n s e : 1159, T r a n s p o r t S e r v i c e$ DirectResponseChannel (org.elasticsearch.transport)
processResponse:1176, TransportService $D i r e c t R e s p o n s e C h a n n e l (o r g . e l a s t i c s e a r c h . t r a n s p o r t) h a n d l e R e s p o n s e : 1082, T r a n s p o r t S e r v i c e$ ContextRestoreResponseHandler (org.elasticsearch.transport)
7.ActionListenerResponseHandler
handleResponse:46, ActionListenerResponseHandler (org.elasticsearch.action)
8.SearchExecutionStatsCollector
onResponse:36, SearchExecutionStatsCollector (org.elasticsearch.action.search)
onResponse:68, SearchExecutionStatsCollector (org.elasticsearch.action.search)
9.SearchActionListener
onResponse:32, SearchActionListener (org.elasticsearch.action.search)
onResponse:51, SearchActionListener (org.elasticsearch.action.search)
10.InitialSearchPhase
innerOnResponse:223, InitialSearchPhase$2 (org.elasticsearch.action.search)
access$000:50, InitialSearchPhase (org.elasticsearch.action.search)
maybeFork:176, InitialSearchPhase (org.elasticsearch.action.search)

private void performPhaseOnShard(. ..) {
    executePhaseOnShard(.. .) {
        //收到执行成功的回复
        public void inne rOnResponse (FirstResult result) {
            maybeFork (thread, () -> onShardResult (result，shardIt) );
        }
        //收到执行失败的回复
        public void onFailure (Exception t) {
            maybeFork(thread, () -> onShardFailure (shardIndex, shard, shard. currentNodeId()，shardIt, t));
        }
    });
}

run:-1, 86025122 (org.elasticsearch.action.search.InitialSearchPhase $2$ $Lambda $1798) c o n s u m e R e s u l t : 385, I n i t i a l S e a r c h P h a s e$ ArraySearchPhaseResults (org.elasticsearch.action.search)
lambda$innerOnResponse$0:223, InitialSearchPhase$2 (org.elasticsearch.action.search)
access$200:50, InitialSearchPhase (org.elasticsearch.action.search)
onShardResult:250, InitialSearchPhase (org.elasticsearch.action.search)
11.AbstractSearchAsyncAction
onShardSuccess:230, AbstractSearchAsyncAction (org.elasticsearch.action.search)
收集返回结果
本过程在search线程池中执行：

private void onShardResult (FirstResult result, SearchShardIterator shardIt) {
    onShardSuccess(result);
    success fulShardExecution(shardIt);
}

onShardSuccess对收集到的结果进行合并。successfulShardExecution方法检查是否所有请求都已收到回复，是否进入

private void successfulShardExecution (SearchShardIterator shardsIt) {
    //计数器累加.
    final int xTotalOps = totalOps.addAndGet (remainingOpsOnIterator);
    //检查是否收到全部回复
    if (xTotalOps == expectedTotalOps) {
        onPhaseDone ();
    } else if (xTota1Ops > expectedTotal0ps) {
        throw new AssertionError(. ..);
    }
}

12.返回InitialSearchPhase
successfulShardExecution:269, InitialSearchPhase (org.elasticsearch.action.search)
13.AbstractSearchAsyncAction
onPhaseDone:248, AbstractSearchAsyncAction (org.elasticsearch.action.search)
executeNextPhase:139, AbstractSearchAsyncAction (org.elasticsearch.action.search)
executePhase:145, AbstractSearchAsyncAction (org.elasticsearch.action.search)

下一步：执行SearchPhase

private void executePhase(SearchPhase phase) {
        try {
            phase.run();
        } catch (Exception e) {
       
    }

doRun:88, FetchSearchPhase$1 (org.elasticsearch.action.search)

   @Override
    public void run() throws IOException {
        context.execute(new ActionRunnable<SearchResponse>(context) {
            @Override
            public void doRun() throws IOException {
            // we do the heavy lifting in this inner run method where we reduce aggs etc. that's why we fork this phase
            // off immediately instead of forking when we send back the response to the user since there we only need
            // to merge together the fetched results which is a linear operation.
                innerRun();
            }
        });
    }

access$000:46, FetchSearchPhase (org.elasticsearch.action.search)
innerRun:106, FetchSearchPhase (org.elasticsearch.action.search)
收集结果
收集器的定义在innerRun中，包括收到的shard数据存放在哪里，收集完成后谁来处理：

final CountedCollector<FetchSearchResult> counter = new CountedCollector<>
    (r -> fetchResults.set(r.getShardIndex()，r), docIdsToLoad.length, finishPhase，context);

fetchResults用于存储从某个shard 收集到的结果，每收到一个shard 的数据就执行一次 counter.countDown()。当所有shard数据收集完毕后，countDown 会出触发执行finishPhase：

final Runnable finishPhase = () -> moveToNextPhase(searchPhaseController, scrollId,
    reducedQueryPhase, queryAndFetchOptimization ? queryResults : fetchResults);

moveToNextPhase方法执行下一阶段，下-阶段要执行的任务定义在FetchSearchPhase构造函数中，主要是触发ExpandSearchPhase：

FetchSearchPhase(InitialSearchPhase.SearchPhaseResults<SearchPhaseResult> resultConsumer, SearchPhaseController searchPhaseController, SearchPhaseContext context) { 
    this(resul tConsumer, searchPhaseController, context,(response, scrollId) -> newExpandSearchPhase(context, response,
        // collapse only happens if the request has inner hits
        (finalResponse)->sendResponsePhase(finalResponse, scrollId, context)));
}