Elasticsearch Upgrade from 6.7 to 7.0 on Kubernetes

(Peter Lee) #1

Hi , I have some question about upgrade ELK from 6.x to 7.0.

Nodes: elasticsearch-0, elasticsearch-1

Part of Elasticsearch 6.7.0 yaml

  spec:
  containers:
  - env:
     #- name: cluster.initial_master_nodes
     # value: elasticsearch-0.es
    - name: cluster.name
      value: elasticsearch-cluster
    - name: node.name
      valueFrom:
        fieldRef:
          apiVersion: v1
          fieldPath: metadata.name
    - name: discovery.seed_hosts
      value: elasticsearch-0.es,elasticsearch-1.es
    - name: ES_JAVA_OPTS
      value: -Xms3g -Xmx3g
    - name: cluster.remote.connect
      value: "true"
    image: docker.elastic.co/elasticsearch/elasticsearch:7.0.0
    imagePullPolicy: IfNotPresent
    name: elasticsearch

On Kibana-dev_tools

GET _cat/nodes?v

ip          heap.percent ram.percent cpu load_1m load_5m load_15m node.role master name
10.12.16.12           16          92  25    1.21    1.79     1.39 mdi       -      elasticsearch-1
10.12.16.10           16          92  25    1.21    1.79     1.39 mdi       *      elasticsearch-0

GET _cat/master?v

id                     host        ip          node
rnjEQ3gfQuO50giXbgE8sw 10.12.16.10 10.12.16.10 elasticsearch-0

Part of Elasticsearch 7.0.0 yaml

  spec:
  containers:
  - env:
    - name: cluster.initial_master_nodes
      value: elasticsearch-0.es
    - name: cluster.name
      value: elasticsearch-cluster
    - name: node.name
      valueFrom:
        fieldRef:
          apiVersion: v1
          fieldPath: metadata.name
    - name: discovery.seed_hosts
      value: elasticsearch-0.es,elasticsearch-1.es
    - name: ES_JAVA_OPTS
      value: -Xms3g -Xmx3g
    - name: cluster.remote.connect
      value: "true"
    image: docker.elastic.co/elasticsearch/elasticsearch:7.0.0

Sometimes cat master or nodes return different node info.

On Kibana-dev_tools

GET _cat/master?v

id                     host      ip        node
s9jkuIrTSdeP4mJ-QgzbVQ 10.0.1.25 10.0.1.25 elasticsearch-0

GET _cat/master?v

id                     host      ip        node
R5DcgRMvQH60HjaEY6A-kA 10.0.5.20 10.0.5.20 elasticsearch-1

GET _cat/nodes?v

ip        heap.percent ram.percent cpu load_1m load_5m load_15m node.role master name
10.0.5.20           49          79  33    0.82    0.83     1.09 mdi       *      elasticsearch-1

GET _cat/nodes?v

ip        heap.percent ram.percent cpu load_1m load_5m load_15m node.role master name
10.0.1.25           29          81  33    0.20    0.32     0.52 mdi       *      elasticsearch-0

1. Output cat/nodes return different node every time.

In 6.7.0 it seem normal.
But in 7.0.0, "_cat/nodes?v" return the different node every time.

Reference for setting up ELK 6.7.0 on GKE

2. Logstash didn't use the default index indices

default index: "logstash-%{+YYYY.MM.dd}".

output {
    elasticsearch { hosts => ["10.138.15.211:9200"]}
}

It use "logstash" as index indices.

3. In my case. Should I config "cluster.initial_master_nodes" ?

- name: cluster.initial_master_nodes
  value: elasticsearch-0.es, elasticsearch-1.es

38%20AM

(system) closed #2

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