dbernard
(David Bernard)
March 9, 2021, 1:38am
1
I am a novice with K8S but experiences with Elasticsearch.
I want to run about 8 Elasticsearch pods per server and have each pod claim a NMVe volume.
There will be about 6 servers.
What is the best way to do this?
Do I specifically need to set up zoning to ensure replicas are distributed across servers?
sebgl
(Sebastien Guilloux)
March 15, 2021, 9:45am
2
ECK sets up a default affinity rule so the Kubernetes scheduler avoids scheduling 2 Pods of the same Elasticsearch cluster: Advanced Elasticsearch node scheduling | Elastic Cloud on Kubernetes [1.4] | Elastic .
You can tweak these rules yourself, and I also suggest setting up availability zone awareness: Advanced Elasticsearch node scheduling | Elastic Cloud on Kubernetes [1.4] | Elastic .
1 Like
system
(system)
Closed
April 12, 2021, 9:46am
3
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.