0

I am trying to run an Elastic Cluster on Kubernetes (Minikube). It basically works, but only with one server per node. As soon as I try to set up > 1 node per node set, the cpu usage of

kube-apiserver kube-controller kube-scheduler kubelet cri-dockerd etcd increases and the cluster is not reachable via kubectl anymore.

This is my ECK-configuration:

apiVersion: elasticsearch.k8s.elastic.co/v1
kind: Elasticsearch
metadata:
  name: reporting
spec:
  version: 7.16.1
  volumeClaimDeletePolicy: DeleteOnScaledownOnly
  nodeSets:
  - name: master
    count: 1
    volumeClaimTemplates:
    - metadata:
        name: elasticsearch-data # Do not change this name unless you set up a volume mount for the data path.
      spec:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 5Gi
        storageClassName: standard
    config:
      node.roles: ["master"]
    podTemplate:
      spec:
        initContainers:
        - name: sysctl
          securityContext:
            privileged: true
            runAsUser: 0
          command: ['sh', '-c', 'sysctl -w vm.max_map_count=262144']
  - name: data
    count: 1
    volumeClaimTemplates:
    - metadata:
        name: elasticsearch-data # Do not change this name unless you set up a volume mount for the data path.
      spec:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 5Gi
        storageClassName: standard
    config:
      node.roles: ["data"]
    podTemplate:
      spec:
        initContainers:
        - name: sysctl
          securityContext:
            privileged: true
            runAsUser: 0
          command: ['sh', '-c', 'sysctl -w vm.max_map_count=262144']

With count = 1 on the nodeset, everything works fine. As soon as I increase it on one of the node sets the described problem arises.

EDIT: Since I don't have access to kubectl, I have no access to the logs, but maybe someone has an idea regarding that.

EDIT: I asked that question on stack overflow before , but got the recommendation to post it at serverfault. I deleted the question on stack overflow so no double posting.

xm22
  • 1

0 Answers0