Starting and Stopping a Cluster

Starting a cluster would mean starting all the nodes of a cluster. Similarly, stopping a cluster would mean stopping all the nodes of a cluster.

You start or stop a node by starting or stopping the Elasticsearch service on each node.

Important! Before you stop a node, you should ensure that no indexing requests or administration-related tasks are being made on the cluster. If you stop a node during indexing, the cluster meta data might get corrupted and the cluster could become non-operational (red color code). To ensure that no instances of PTSF_GENFEED are running, check the Process Monitor. If all processes are completed, you may stop all the nodes in a cluster and make the required modifications. After completing the modifications, you may start all the nodes of the cluster.

You start or stop a node by starting or stopping the Elasticsearch service on the node. Use the following steps to start or stop a node on Windows and on Linux.

Starting and Stopping the Elasticsearch Service on Windows

On Microsoft Windows, the Elasticsearch service is installed by the DPK setup script.

To start or stop an Elasticsearch service:

  1. Open a command prompt, and change directory to ES_HOME/bin.

  2. To see the usage for the service command: 

    service
service.bat install|remove|start|stop

    • To start the Elasticsearch service:

      service.bat start

    • To stop the Elasticsearch service:

      service.bat stop

Starting and Stopping the Elasticsearch Service on Linux

To start or stop the Elasticsearch service:

  • In a terminal window, change directory to ES_HOME/bin.

  • To start the Elasticsearch process:

    nohup ./elasticsearch &

  • To stop the Elasticsearch process:

    1. Use this command to find the Elasticsearch process ID:

      ps -ef | grep elas

    2. Use this command to stop the process, substituting the process ID for pid:

      kill <pid>