Restarting an Elasticsearch Cluster

Constraints

• Avoid restarting a cluster during peak hours, as doing so may impact data integrity and service availability.
• During a quick restart, the cluster is unavailable for the entire duration of the restart.
• Rolling restart is supported only for clusters with three or more nodes, including the master node, client node, and cold data nodes. During a rolling restart, only the nodes that are being restarted are unavailable. When the cluster stores a large amount of data, the restart may take a long time.

Prerequisites

• The cluster you want to restart is not frozen and has no task in progress.
• The cluster is available and is not processing any service requests, such as importing and searching for data. Otherwise, data loss may occur during a cluster restart.

Quick Restart

1. Log in to the CSS management console.
2. In the navigation pane on the left, choose Clusters > Elasticsearch.
3. In the cluster list, find the target cluster, and choose More > Restart in the Operation column.
4. On the Restart Cluster page, set parameters.

   Table 1 Configuring a quick restart

   Parameter	Description
   Restart Mode	Select Quick Restart. WARNING: The cluster will be unavailable for the entire duration of a quick restart.
   Operation Role	Select Node type or Node name. Node type: You can select multiple node types to restart them at the same time. Node name: You can restart only one node at a time.
   Operation Value	Select one or more node types or a node name, depending on the setting of Operation Role. Values for Node type: ess: data node ess-master: master node ess-client: client node ess-cold: cold data node

5. Click OK to start the restart task.

   Refresh the page and check the cluster status. During the restart, the cluster status is Processing, and the task status is Restarting. When the cluster status changes to Available, the cluster has restarted successfully.

Rolling Restart

1. Log in to the CSS management console.
2. In the navigation pane on the left, choose Clusters > Elasticsearch.
3. In the cluster list, find the target cluster, and choose More > Restart in the Operation column.
4. On the Restart Cluster page, set parameters.

   Table 2 Configuring a rolling restart

   Parameter	Description
   Restart Mode	Select Rolling Restart. Rolling restart is supported only for clusters with three or more nodes, including the master node, client node, and cold data nodes. During a rolling restart, the nodes of a cluster are restarted one at a time. This may take a long time if the cluster has a large number of indexes.
   Operation Role	Only Node type is supported. You can select multiple node types to restart them at the same time.
   Operation Value	Select node types for the restart. Values for Node type: ess: data node ess-master: master node ess-client: client node ess-cold: cold data node

5. Click OK to start the restart task.

   Refresh the page and check the cluster status. During the restart, the cluster status is Processing, and the task status is Restarting. When the cluster status changes to Available, the cluster has restarted successfully.