https://github.com/elastic/ansible-elasticsearch
ansible快速部署elastic集群
"index.refresh_interval" : "20s"
(這一步是memory buffer 刷到文件系統(tǒng)緩存的間隔時(shí)間)
好吧~我覺得優(yōu)化elastic還是得從原理去研究了優(yōu)化(重寫一張好了)
reroute接口中allocate指令 當(dāng)主節(jié)點(diǎn)宕機(jī)或者無響應(yīng)可以加個(gè)參數(shù)allow_primary:true來解決
reroute接口中move指令可以讓負(fù)載過高的答捕,磁盤使用高的節(jié)點(diǎn)上的分片移走
ES讀寫分離
https://www.elastic.co/guide/en/elasticsearch/reference/6.3/index-modules-allocation.html
https://blog.csdn.net/mlljava1111/article/details/51363230