weizijun
weizijun
write节点判断hashId是ok的，但是并未写入指定hashId的分片，searcher上有多个分片时，看起来会随机写入分片
Add docs in support of AlibabaCloud AI Search support in the Inference API: https://github.com/elastic/elasticsearch/pull/111181
### Description Currently, knn search does not use LRUQueryCache for acceleration. The knn search is heavy when k is big. Can query caching be supported in knn search? In semantic...