elasticsearch - 未能将映射放在索引上导致elasticsearch崩溃[5.4.1]

Question

ElasticSearch [5.4.1]1 我们有一些索引，如 ics_pcs_2017.06.062 我们删除一些旧的索引使用jave代码: getClient.admin().indices().prepareDelete(indexname).execute().actionGet()

I think the index ics_pcs_2017.06.06 will have 60G+ and 100,000,000 documents in it.

3我们创建一个新的logstash来接收旧日志，我们在ES日志中发现: ES日志信息

create index for ics_pcs_2017.06.06( didn't get the detail log)

.....
[o.e.c.m.MetaDataCreateIndexservice] create index cause [auto(bulk api)], templates[icstemplate],shards [6]/[1], mappings [icslog]

.....
[o.e.c.m.MetaDataCreateIndexservice] [ics_pcs_2017.06.06/xtNoHLSOS4GBQo01-Nprg] update_mapping [icslog]

...
[o.e.a.a.1.m.p.TransportPutMappingAction] [ics-master1] failed to put mappings on indices[[[ics_pcs_2017.06.06/xtNoHLSOS4GBQo01-Nprg]]]
org.elasticsearch.cluster.metada.ProcessClusterEventTimeoutException:failed to process cluster event(put-mapping) within 30s

....
-----------
then there is many logs like  failed to put mappings on indices[[[ics_pcs_2017.06.06/xtNoHLSOS4GBQo01-Nprg]]]
... failed to put mappings on indices[[[ics_pcs_2017.06.06/xtNoHLSOS4GBQo01-Nprg]]]
...failed to put mappings on indices[[[ics_pcs_2017.06.06/xtNoHLSOS4GBQo01-Nprg]]]

and cause the elasticsearch master down.

请各位帮忙解释一下原因，非常感谢。

Answer 1

当映射到给定文档的字段时出现错误时，就会发生此类错误。就像 ES-5.4 的例子之一:-当您尝试映射“字符串”时会发生这种情况，因为在 ES 词汇中“字符串”被称为“关键字”。关于本文的更多信息

Mapping change

在您的情况下，服务器首先尝试进行映射，在尝试过程中，它也给出超时消息。

我的建议:修复映射和分析器问题。