Help Center/ MapReduce Service/ Component Operation Guide (LTS) (Ankara Region)/ Using Elasticsearch/ Elasticsearch Performance Tuning/ Performance Optimization for Large Clusters
Updated on 2024-11-29 GMT+08:00
View PDF
Share

Performance Optimization for Large Clusters

Scenario

Reconfigure the parameters listed below to improve the stability of the Elasticsearch cluster if you have planned over 200 EsNode instances in your cluster and there are petabytes of data.

Procedure

  1. Log in to FusionInsight Manager.
  2. Choose Cluster > Services > Elasticsearch. Click Configurations then All Configurations. Click Elasticsearch(Service) and choose Customization.
  3. Add the following custom parameters to the elasticsearch.customized.configs parameter:

    Parameter

    Description

    Default Value

    Recommended Value

    cluster.fault_detection.follower_check.interval

    Interval for heartbeat detection between master nodes and other nodes

    10s

    60s

    cluster.fault_detection.leader_check.interval

    Interval for other nodes to check whether master nodes are alive

    10s

    60s

    cluster.fault_detection.follower_check.timeout

    Timeout interval for heartbeat detection between master nodes and other nodes

    60s

    360s

    cluster.fault_detection.leader_check.timeout

    Timeout period for other nodes to check whether master nodes are alive

    60s

    360s

    cluster.follower_lag.timeout

    Timeout interval for a node to update the cluster state

    300s

    600s

    cluster.publish.timeout

    Timeout interval for master nodes to wait for the completion of publishing the cluster state

    90s

    120s

  4. Click Save to save the configuration.
  5. Click Dashboard to go to the Elasticsearch service page. Choose More > Restart Service. Verify the identity and click OK. Wait until the restart is successful.