Skip to main content
Version: 5.1

Restarting OpenSearch Nodes

Restarting OpenSearch nodes is the final step when making cluster configuration changes such as certificate updates, security configurations, or operational parameter modifications. This process is critical for applying new settings and ensuring system stability.

Please note!

Proper sequence must be followed as incorrect restart procedures may lead to temporary data unavailability or cluster integrity issues.

Verifying Cluster Status​

Before restarting nodes, verify the cluster is in stable condition:

GET _cluster/health
Note!

If status is not green, resolve existing issues before proceeding.

Disabling Shard Allocation​

To prevent data redistribution during restart, temporarily disable shard allocation via Dev Console (Main Menu - System Settings - Dev Console) by executing:

PUT _cluster/settings
{
  "persistent": {
    "cluster.routing.allocation.enable": "none"
  }
}

Restarting Master Nodes​

Master nodes control cluster operations and must be restarted one at a time to maintain quorum.

  • Stop the master node and wait for process completion:
systemctl stop opensearch
  • Start the node:
systemctl start opensearch
  • Verify successful cluster rejoining:
GET _cat/nodes?v

Restarting Data Nodes​

Restarting SA Data Storage should be performed in stages: first, it is necessary to perform a sequential reboot of nodes with the routing_mode parameter cold, then - warm, after - `hot'.

Re-enabling Shard Allocation​

After updating all cluster nodes, re-enable shard allocation via Dev Console (Main Menu - System Settings - Dev Console) by executing:

PUT _cluster/settings
{
  "persistent": {
    "cluster.routing.allocation.enable": "all"
  }
}