ElasticSearch笔记

优化

性能测试

• 在一个节点的一个分片，不设置副本，测试性能
• 在完全默认设置上记录性能数据，作为测试的基准线
• 确保性能测试持续30分钟以上以确认长时间的性能；短时间的测试可能不会碰到segment合并和GC，无法确认这些因素的影响
• 每次基于默认基准线更改一个参数，如果性能有提升就保留设置，并基于此设置做后续的测试

http://blog.youkuaiyun.com/jamesjxin/article/details/50512106

生产环境elasticsearch的配置建议

以下主要来自官方文档，主要分为几块：

http://blog.youkuaiyun.com/thomas0yang/article/details/55518105

分片大小的建议 每个分片小于50GB

https://discuss.elastic.co/t/too-big-a-shard-vs-too-many-shards/75889/6

Elasticsearch集群搭建实战

http://blog.youkuaiyun.com/ch648966459/article/details/51671315

大规模集群实践

https://elasticsearch.cn/article/110

https://elasticsearch.cn/people/wood

PUT /_cluster/settings

{

    "persistent" : {

        "discovery.zen.minimum_master_nodes" : 8

    }

}

gateway.expected_nodes: 15

gateway.recover_after_nodes: 12

gateway.recover_after_time: 5m

笔记

设置分片数

curl -X PUT -u {username}:{password} "{ip}:9200/{indexName}" -d '

{

    "settings" : { 

        "number_of_shards" : {number} 

    }

}

'

task 观察任务运行状态

https://www.elastic.co/guide/en/elasticsearch/reference/5.1/tasks.html#task-cancellation

GET _tasks

GET _tasks?actions=*reindex&detailed

GET _tasks/uOCOZUwHQ-CWfbeizCDOOg:37931

POST _tasks/uOCOZUwHQ-CWfbeizCDOOg:37931/_cancel

别名alias

https://www.elastic.co/guide/en/elasticsearch/reference/5.1/indices-aliases.html

POST /_aliases

{

  "actions": [

    {

      "add": {

        "index": "probe_datecl2",

        "alias": "probe_datecl4"

      }

    }

  ]

}

问答区

https://discuss.elastic.co/

提高索引速度

https://www.elastic.co/guide/en/elasticsearch/reference/master/tune-for-indexing-speed.html

限制reindex的速度，默认不限制

requests_per_second   can be set to any positive decimal number ( 1.4 ,   6 ,   1000 , etc) and throttles rate at which reindex issues batches of index operations by padding each batch with a wait time. The throttling can be disabled by setting   requests_per_second   to   -1 .

https://www.elastic.co/guide/en/elasticsearch/reference/master/docs-reindex.html#reindex-from-remote

ElasticSearch 的一次非正常master脱离的调查

http://simonlei.iteye.com/blog/1669992

5.1.1存在经常节点失联的bug，导致集群状态不稳定（yellow），不停均衡，大量磁盘io以及cpu使用。影响入库以及查询效率。

https://discuss.elastic.co/t/es-5-1-1-cluster-loses-a-node-randomly-every-few-hours-error-message-not-fully-read-response-for-requestid/71208

es solr对比  https://www.cnblogs.com/chowmin/articles/4629220.html

xpack破解方法  http://blog.youkuaiyun.com/u013066244/article/details/73927756