Starting a Cluster

If all the nodes in a cluster are stopped, you must follow this procedure to start the cluster safely and avoid data loss.

  1. On the last node that was stopped last, remove the seed nodes and start the server.
    Warning: The first node that you start must be the last node that was stopped, and that node must be a seed node. Starting another node first may result in loss of data such as job history and configuration settings. If you do not know which node was stopped last, look in each node's wrapper log for the time stamp of the shutdown message. You can find the wrapper log in: Spectrum Location\server\app\repository\logs\wrapper.log.
    1. Open this file in a text editor:

      server/app/conf/spectrum-container.properties

    2. In the spectrum.cluster.seeds property, remove all host names and IP addresses except for the one for this server. Save the host names and IP addresses so that you can re-add them later.
    3. Save the file.
    4. Start the server.
    5. Wait for the Spectrumâ„¢ Technology Platform server to completely start.

      You can tell when the Spectrumâ„¢ Technology Platform server has completely started by looking in the wrapper log: Spectrum Location\server\app\repository\logs\wrapper.log. This message is displayed when the server is completely started:

      Pitney Bowes Spectrum(TM) Technology Platform (Version Version Number) Started.
    6. In the properties file spectrum-container.properties, in the spectrum.cluster.seeds property, add the host names or IP addresses that you had removed, separating each with a comma.
    7. Save and close the file. You do not need to restart the server.
  2. Start the other nodes in the cluster.
    Warning: Be sure to wait for the first node to start completely before starting additional nodes. Starting additional nodes before the first one is started may result in loss of data.