How to modify the ES configuration and reboot ES cluster properly?


(Keith Tt) #1

ELK version: 5.5.1

Restart nodes one by one is terrible......

Should I refer to this guide? https://www.elastic.co/guide/en/elasticsearch/reference/5.5/rolling-upgrades.html


(Mark Walkom) #2

That is the correct method.

Also we’ve renamed ELK to the Elastic Stack, otherwise Beats feels left out :wink:


(Keith Tt) #3

So...If there are 10 nodes in the cluster, it possibly needs one day to restart all the nodes.


(Mark Walkom) #4

Why?


(Keith Tt) #5

When one node is restarted, should I wait all the shards been allocated and the cluster's status recover to green? This progress spend much time...


(Mark Walkom) #6

If you are on version 5, then delayed allocation and synced shards should negate any delays.
If they aren't, then it'd be great to know more information.


(Keith Tt) #7

Even if I disable shard allocation and perform a synced flush before I restart each node, it still spends much time to recover after I reenable shard allocation...


(system) #8

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.