就像标题所说的…
我阅读了这篇文章(https://www.elastic.co/blog/changing-mapping-with-zero- downtime),这个概念很棒,但是我很难找到有关如何通过JAVA API进行操作的不错的参考。
我找到了这个插件:https : //github.com/karussell/elasticsearch- reindex,但似乎对我尝试做的事情有些过头
经过在当地星巴克的研究后,我得出了以下结论:
假设我们已经有索引(“ old_index”)并且它有数据…现在让我们将数据移动到我们创建的新索引(“ new_index”)中(对于某个字段,可能具有不同的STRING vs INT,或现在您决定不再希望分析或存储某些字段等)。
这里的基本思想是从已经存在的索引(“ old_index”)中检索所有数据,并将其提取到新索引(“ new_index”)中。但是,您只需要做几件事:
步骤1.您需要执行搜索滚动 https://www.elastic.co/guide/zh/elasticsearch/reference/current/search- request- scroll.html
与常规搜索相比,它所做的一切都可以更高效地检索结果。没有评分,等等。这是文档必须说的:“滚动并不是为了实时用户请求,而是为了处理大量数据,例如,为了将一个索引的内容重新索引为新索引具有不同的配置。”
这是有关如何使用Java API的链接:https : //www.elastic.co/guide/en/elasticsearch/client/java- api/current/scrolling.html
第2步。插入时,必须使用批量摄取。再次出于性能原因完成此操作。这是Bulk Ingest Java API的链接:https : //www.elastic.co/guide/en/elasticsearch/client/java- api/current/bulk.html#_using_bulk_processor
现在到ho上去做吧…
步骤1.设置滚动搜索以从旧索引中“加载”数据
SearchResponse scrollResp = client.prepareSearch("old_index") // Specify index .setSearchType(SearchType.SCAN) .setScroll(new TimeValue(60000)) .setQuery(QueryBuilders.matchAllQuery()) // Match all query .setSize(100).execute().actionGet(); //100 hits per shard will be returned for each scroll
步骤2.设置批量处理器。
int BULK_ACTIONS_THRESHOLD = 1000; int BULK_CONCURRENT_REQUESTS = 1; BulkProcessor bulkProcessor = BulkProcessor.builder(client, new BulkProcessor.Listener() { @Override public void beforeBulk(long executionId, BulkRequest request) { logger.info("Bulk Going to execute new bulk composed of {} actions", request.numberOfActions()); } @Override public void afterBulk(long executionId, BulkRequest request, BulkResponse response) { logger.info("Executed bulk composed of {} actions", request.numberOfActions()); } @Override public void afterBulk(long executionId, BulkRequest request, Throwable failure) { logger.warn("Error executing bulk", failure); } }).setBulkActions(BULK_ACTIONS_THRESHOLD).setConcurrentRequests(BULK_CONCURRENT_REQUESTS).setFlushInterval(TimeValue.timeValueMillis(5)).build();
步骤3.在步骤1中通过创建的滚动搜索器从旧索引中读取,直到剩下mo记录并插入新索引中
//Scroll until no hits are returned while (true) { scrollResp = client.prepareSearchScroll(scrollResp.getScrollId()).setScroll(new TimeValue(600000)).execute().actionGet(); //Break condition: No hits are returned if (scrollResp.getHits().getHits().length == 0) { logger.info("Closing the bulk processor"); bulkProcessor.close(); break; } // Get results from a scan search and add it to bulk ingest for (SearchHit hit: scrollResp.getHits()) { IndexRequest request = new IndexRequest("new_index", hit.type(), hit.id()); Map source = ((Map) ((Map) hit.getSource())); request.source(source); bulkProcessor.add(request); } }
步骤4.现在是将指向旧索引的现有别名分配给新索引的时候了。然后删除对旧索引的别名引用,然后删除旧索引本身。为新索引分配别名
client.admin().indices().prepareAliases().addAlias("new_index", "alias_name").get();
从旧索引中删除别名,然后删除旧索引
client.admin().indices().prepareAliases().removeAlias("old_index", "alias_name").execute().actionGet(); client.admin().indices().prepareDelete("old_index").execute().actionGet();