Stopping and starting services running on WML for z/OS base

The services and processes on WML for z/OS base generally run independent of each other. When needed, you can stop and restart a specific service or process without having to shut down the entire solution.

Procedure

  • Restart Spark runtime processes.

    You can issue the following commands to stop and then start Spark from anywhere on your system if you created your Spark runtime environment by running the create-spark-runtime.sh script.

    1. Issue the following command to stop the Spark worker:
      stop_slave
    2. Issue the following command to stop the Spark master:
      stop_master
    3. Issue the following command to start the master again:
      start_master
    4. Issue the following command to start the worker again:
      start_slave
  • Restart Jupyter Kernel Gateway processes.
    1. Locate the kg2at-stop.sh and kg2at-start.sh scripts in the $IML_INSTALL_DIR/imlpython/bin directory.
    2. Issue the following commands to stop the kernel gateway:
      ./kg2at-stop.sh
    3. Issue the following commands to start the gateway again:
      ./kg2at-start.sh -e <conda_env_name>

      where <conda_env_name> is the Anaconda environment name you specified or the default mlzenv that you accepted.

  • Restart WMLz base UI service.
    1. Locate the MLzServer.sh script in the $IML_INSTALL_DIR/iml-portal directory.
    2. Issue the following command to stop the UI service:
      ./MLzServer.sh stop
    3. Issue the following command to start the service again:
      ./MLzServer.sh start
  • Restart WMLz base core services, including the repository, deployment, bath scoring, ingestion, feedback, and training services.
    1. Locate the MLzService.sh script in the $IML_INSTALL_DIR/iml-services directory.
    2. Issue the following command to stop the repository, deployment, bath scoring, ingestion, and associated services:
      ./MLzService.sh stop
    3. Issue the following command to start the services again:
      ./MLzService.sh start
  • Restart the Spark-integration service.
    1. Locate the sparkaas.sh script in the $IML_INSTALL_DIR/bin directory.
    2. Issue the following command to stop the Spark-integration service service:
      ./sparkaas.sh stop
    3. Issue the following command to start the service again:
      ./sparkaas.sh start
  • Restart WML for z/OS scoring service and the Python scoring service.
    1. Locate the server.sh script in the $IML_INSTALL_DIR/bin directory.
    2. Issue the following command to stop the WML for z/OS scoring service:
      ./server.sh stop <serverName>

      If a uWSGI server is attached, issue the following command to stop the server:

      ./server.sh stop-python <serverName>
    3. Issue the following command to start the scoring service again:
      ./server.sh start <serverName>

      If a uWSGI server is attached, issue the following command to restart the server:

      ./server.sh start-python <serverName>

      See Configuring WML for z/OS scoring services for details.

  • If applicable, restart the WML for z/OS scoring service in a CICS® region.
    1. Issue the following CICS region command to stop the ALNSCSER server and discard the scoring service bundle:
      
      CEMT SET BUNDLE(ALNSCBDL) DISABLED
      CEMT SET JVMSERVER(ALNSCSER) DISABLED
      CEMT DISCARD BUNDLE(ALNSCBDL)
      CEMT DISCARD JVMSERVER(ALNSCSER)
    2. Issue the following command to start the ALNSCSER server and the scoring service bundle again:
      
      CEMT SET JVMSERVER(ALNSCSER) ENABLED
      CEMT SET BUNDLE(ALNSCBDL) ENABLED

      See Configuring WML for z/OS scoring services in a CICS region for details.