After you modify the configuration of an Elasticsearch cluster or node or perform other operations on them, you may need to manually restart the cluster for the changes to take effect. This topic describes how to restart an Elasticsearch cluster or node in the console.

Prerequisites

Before you restart a cluster, make sure that the status of the cluster is Active (green), the index has at least one replica, and the resource usage is not high.
Note You can view the resource usage on the Cluster Monitoring page. For example, the value of NodeCPUUtilization(%) is about 80%, the value of NodeHeapMemoryUtilization is about 50%, and the value of NodeLoad_1m is lower than the number of vCPUs of the current data node. For more information, see Cluster monitoring.

Restart a cluster

  1. Log on to the Alibaba Cloud Elasticsearch console.
  2. In the top navigation bar, select the region where your Alibaba Cloud Elasticsearch cluster resides.
  3. Find the target cluster and click its ID.
  4. In the upper-right corner of the Basic Information page, click Restart.
  5. In the Restart dialog box, set parameters to restart a cluster.
    Restart an Elasticsearch cluster
    Parameter Description
    Object Cluster and Node are supported.
    • Cluster: Restarts all nodes in a cluster.
    • Node: Restarts a single node. For more information, see Restart a node.
    Restart mode Alibaba Cloud Elasticsearch provides two restart modes: Restart and Force Restart.
    • Restart: You can only restart a cluster whose status is Active (green). Otherwise, you need to forcibly restart the cluster. This restart mode does not affect the service of your cluster but is time-consuming.
      Notice
      • During the restart, the CPU utilization and memory usage of the nodes in your Elasticsearch cluster will increase sharply. This may affect the stability of your services for a short period of time.
      • The specific amount of time that is used to restart a cluster depends on the volume of data stored on the cluster and the numbers of nodes, indexes, and shards in the cluster. You can view the progress of a restart task in the Tasks dialog box.
    • Force Restart: If your Elasticsearch cluster is unhealthy (yellow or red), you can only use the Force Restart mode to restart the cluster.
      Notice If the disk usage exceeds the value of cluster.routing.allocation.disk.watermark.low, the Status of your Elasticsearch cluster may be displayed yellow or red. If your Elasticsearch cluster is abnormal, do not perform the following operations on the cluster: node addition, node capacity expansion, disk space expansion, restart, password reset, or other operations that may change the configuration of the cluster. Perform the operations only after the Status of the cluster is Active (green).
    Concurrency You can set the concurrency of the cluster to improve the restart speed. The higher the concurrency, the faster the force restart. The default concurrency is calculated by dividing one by the total number of nodes in the cluster.
    Estimated to Take This value is calculated by multiplying the average time of the previous restart tasks by the total number of nodes. The actual restart time takes precedence.
  6. Click OK.
    Note If you choose Force Restart, select Restart Cluster Forcibly to confirm the restart.

    During the restart, the Status field is Initializing (yellow). You can view details in the Tasks dialog box. After the cluster is restarted, the Status field is Active (green).

    Restart progress

Restart a node

Restart a single node. The procedure and precautions for restarting a node are similar to those in Restart a cluster. The differences are as follows:

Restart a node
  • In the Restart dialog box, set Object to Node.
  • Select the node you want to restart.
    Notice If your Elasticsearch cluster is abnormal, you need to forcibly restart nodes in the cluster.
  • The system provides the Change Blue-green Release feature. If you select it, Elasticsearch adds a node to your Elasticsearch cluster, migrates the data on the original node to the new node, and removes the original node. If a node experiences a hardware failure, you can use the Change Blue-green Release feature to remove the node.
    Warning
    • If you use the Change Blue-green Release feature, make sure that your Elasticsearch cluster is in the Active state (green) and do not set Restart Mode to Force Restart.
    • If you select Change Blue-green Release, the IP address of the node changes after the restart.