Elasticsearch缓存源码分析

官网：query-cache
官网：shard-request-cache
官网：fileddata

1.NodeQueryCache，没有这个类，对应的是ElasticsearchLRUQueryCache

a.IndicesService类里的loadOrExecuteQueryPhase方法判断是否可以缓存，可以缓存的分支执行loadIntoContext方法。

    /**
     * Try to load the query results from the cache or execute the query phase directly if the cache cannot be used.
     */
    private void loadOrExecuteQueryPhase(final ShardSearchRequest request, final SearchContext context) throws Exception {
        final boolean canCache = indicesService.canCache(request, context);
        context.getQueryShardContext().freezeContext();
        if (canCache) {
            indicesService.loadIntoContext(request, context, queryPhase);
        } else {
            queryPhase.execute(context);
        }
    }

b.loadIntoContext方法执行cacheShardLevelResult，将结果缓存。

    public void loadIntoContext(ShardSearchRequest request, SearchContext context, QueryPhase queryPhase) throws Exception {
        assert canCache(request, context);
        final DirectoryReader directoryReader = context.searcher().getDirectoryReader();

        boolean[] loadedFromCache = new boolean[] { true };
        BytesReference bytesReference = cacheShardLevelResult(context.indexShard(), directoryReader, request.cacheKey(), out -> {
            queryPhase.execute(context);
            try {
                context.queryResult().writeToNoId(out);

            } catch (IOException e) {
                throw new AssertionError("Could not serialize response", e);
            }
            loadedFromCache[0] = false;
        });
        ......
    }

c.IndicesRequestCache的getOrCompute方法，存入本地缓存。

    private BytesReference cacheShardLevelResult(IndexShard shard, DirectoryReader reader, BytesReference cacheKey, Consumer<StreamOutput> loader)
            throws Exception {
        IndexShardCacheEntity cacheEntity = new IndexShardCacheEntity(shard);
        Supplier<BytesReference> supplier = () -> {
            /* BytesStreamOutput allows to pass the expected size but by default uses
             * BigArrays.PAGE_SIZE_IN_BYTES which is 16k. A common cached result ie.
             * a date histogram with 3 buckets is ~100byte so 16k might be very wasteful
             * since we don't shrink to the actual size once we are done serializing.
             * By passing 512 as the expected size we will resize the byte array in the stream
             * slowly until we hit the page size and don't waste too much memory for small query
             * results.*/
            final int expectedSizeInBytes = 512;
            try (BytesStreamOutput out = new BytesStreamOutput(expectedSizeInBytes)) {
                loader.accept(out);
                // for now, keep the paged data structure, which might have unused bytes to fill a page, but better to keep
                // the memory properly paged instead of having varied sized bytes
                return out.bytes();
            }
        };
        return indicesRequestCache.getOrCompute(cacheEntity, supplier, reader, cacheKey);
    }

d.computeIfAbsent 这个方法执行存储。

e.segment是cache使用map，避免竞争，使用分段技术，hash方法如下。

private CacheSegment<K, V> getCacheSegment(K key) {
        return segments[key.hashCode() & 0xff];
    }

官方介绍网址

2.ShardRequestCache

a.官方说默认指挥存储空结果，除开聚合和建议。https://www.elastic.co/guide/en/elasticsearch/reference/5.6/shard-request-cache.html

b.从源码看在返回结果存在的话，缓存NodeQueryCache，没有结果的话缓存shardReuestCache.

3.FieldDataCache:

a.Fielddata使用可以通过下面的方式来监控：

对于单个索引使用 {ref}indices-stats.html[indices-stats API]:
GET /_stats/fielddata?fields=*

对于单个节点使用 {ref}cluster-nodes-stats.html[nodes-stats API]:
GET /_nodes/stats/indices/fielddata?fields=*

或者甚至单个节点单个索引
GET /_nodes/stats/indices/fielddata?level=indices&fields=*

https://www.elastic.co/guide/cn/elasticsearch/guide/current/heap-sizing.html