Setting up a job manager environment (deprecated)

A job manager environment consists of a job manager and the targets that it manages. The job manager targets can be deployment managers, stand-alone application server nodes that are managed by administrative agents, and host computers. Setting up a job manager environment involves creating a job manager profile and any other profiles that are needed for the environment, synchronizing the clocks on all environment computers, and then registering the targets with the job manager.

Before you begin

Deprecated feature: The job manager is deprecated. Instead of job manager, use Urban Code Deploy to install the product, to deploy applications, and to manage remote hosts. To manage WebSphere® Application Server installations, you also can use swinging profiles. See Swinging profiles between product installations.

Install the WebSphere Application Server product.

1. Determine the topology for your flexible management environment. Flexible management encompasses administrative agents and job managers.

   Determine which machines, targets, and target resources such as servers and applications to be in the flexible management environment.

   To manage stand-alone application servers, use an administrative agent on each computer where the stand-alone application servers reside. For more information, see topics on the administrative agent and Scenarios 5 in the Planning to install WebSphere Application Server topic.

   To collectively manage deployment managers and stand-alone application servers on the same or different computers, use a job manager. The stand-alone application servers must be registered with an administrative agent before you can manage them using a job manager. For more information, see Scenarios 5 and 10 in the Planning to install WebSphere Application Server topic.

2. Determine the security roles needed for your flexible management environment.

   Depending on your environment, you might need profiles for administrative agents, the nodes registered to the administrative agents, deployment managers, the nodes federated with the deployment manager, and job managers. Profiles in the flexible management environment must either all have security enabled or all have security disabled. When you create the profiles, you can specify security options, user names, and passwords.

   You must have security roles that authorize you to work with a job manager and to manage registered targets and resources on those targets. If the environment includes stand-alone application server target nodes, then you must be authorized to work with an administrative agent and its nodes.

   For more information, see the job manager security topic.

3. Create a management profile for the job manager.

   You can use the Profile Management Tool or the manageprofiles command.

   For example, in the Profile Management Tool, select the Management environment and click Next, select the Job manager server type, and select options that create the profile. By default, a job manager has its own administrative console, administrative security is enabled, and the console port is 9960. To disable administrative security, to specify a security certificate, or to change the default ports, use the advanced profile creation option when creating the job manager profile.

   By default, the first administrative agent profile in a product installation is named JobMgr01 and its server name is jobmgr.

   For more information, see the topic on creating management profiles for job managers.

   For manageprofiles examples, see the topic on the manageprofiles command. For -templatePath, specify the management template. For -serverType, specify JOB_MANAGER.

   Note: The job manager configuration includes a datasource named OTiSDataSource. This datasource is used in the implementation of the job manager, and does not need to be configured or otherwise managed by the administrator.
4. Create profiles for any administrative agents and stand-alone application server nodes that you intend to have in your flexible management environment. Then, register the stand-alone application server nodes with the administrative agent.

   Stand-alone nodes are also called unfederated or base application servers. They are not managed by a deployment manager. Stand-alone application servers typically have a profile name such as AppSvr01. An administrative agent must be on the same computer as its stand-alone nodes. Registering the stand-alone nodes with the administrative agent enables the administrative agent to manage the nodes.

   Avoid trouble: You must register stand-alone application servers with an administrative agent before you can register the stand-alone application servers with the job manager.

   For details on creating the profiles and registering with an administrative agent, see the topic on setting up the administrative agent environment.

5. Create profiles for any deployment managers and federated nodes that you intend to have in your flexible management environment.

   Federated nodes are managed by a deployment manager. Federated application servers typically have a profile name such as AppSvr01, however you cannot administer them individually. You must administer federated nodes using the deployment manager.

   See topics on creating cell profiles, management profiles for deployment managers, or the manageprofiles command.

6. Synchronize the clocks on all involved systems.

   If you plan to change the system clock, stop all the application servers, the node agent servers, the deployment manager server, the administrative agent server, and the job manager server first. After you stop the servers, change the system clock, and then restart the servers.

   For z/OS, also stop and then restart the location service daemon.

   If you change the system clock on one system, you must ensure the clocks on all systems that communicate with each other and have WebSphere Application Server installed are synchronized. Otherwise, you might experience errors, such as security tokens no longer being valid.

7. Start the job manager server.
   • Run the startServer command.

     For example, suppose the JobMgr01 profile has the server name jobmgr. Run the following command from the bin directory of the JobMgr01 profile:

     startServer jobmgr

   • Use the START command to start the job manager:

     START job_manager_proc_name,JOBNAME=server_short_name,
           ENV=cell_short_name.node_short_name.server_short_name

   If the job manager starts successfully, the message open for e-business displays and is written to the job manager startServer.log file:

   Server launched.  Waiting for initialization status.
   Server jobmgr open for e-business; process id is 1932.

   For more information, see the topic on starting and stopping the job manager.

8. Register stand-alone application server target nodes with a job manager.

   Registering stand-alone nodes with a job manager enables the job manager to administer stand-alone application server nodes.

9. Register deployment managers with the job manager.

   Registering a deployment manager with a job manager enables you to run job manager jobs from a deployment manager console and enables the job manager to administer federated nodes of the deployment manager and their resources.

10. Register host computers with the job manager.

    A remote host target is not required to have any WebSphere Application Server products installed. There are no software requirements for this host beyond its operating system. Registering a remote host with a job manager enables the job manager to access applications, command files, and other resources on the host computer.

    To register Liberty with a job manager, use a procedure for registering a target with a host.

11. Verify that the targets are registered with the job manager.

    You can use an administrative console or wsadmin scripting commands to see a list of targets that are registered with the job manager.

    • In the job manager console or deployment manager console, click Jobs > Targets. The Targets page lists targets that are registered with the job manager.
    • Run the AdminConfig list command to see a list of managed targets. Run the following wsadmin scripting commands from the administrative agent bin directory to list stand-alone application server targets or from the deployment manager bin directory to list other targets.
      • To use the Jython scripting language, enter the following two commands in succession:

        wsadmin -lang jython
        
        print AdminConfig.list('JobManagerRegistration')

      • To use the Jacl scripting language, enter the following two commands in succession:

        wsadmin
        
        $AdminConfig list JobManagerRegistration

      After you verify that the targets are registered with the job manager, enter quit to exit the wsadmin scripting tool.

12. Ensure that the servers in your flexible management environment are running.

    In the job manager console or deployment manager console, click Jobs > Target resources > server_name. On the Target resource page, a server status of Started shows that the server is running.

13. Optional: Disable the com.ibm.otis.Audit_mm_dd_yyyy.log log file by specifying the JVM property otis.audit.location with a value of OFF. or move the log file to a new directory location.
    Job manager maintains an additional log which is in the profile logs directory by default. The purpose of the log is to record job manager internal activity.

    The name of the log file is com.ibm.otis.Audit_mm_dd_yyyy.log, where mm, dd and yyyy are month, date and year respectively. A new file is created for each day if activity occurs that can be logged.

    This log file can be moved to another directory by specifying the JVM property otis.audit.location with a value of the new directory location.

14. Back up the job manager database.
    Periodically use the backupJobManager AdminTask in wsadmin to back up the job manager database.
15. Delete old jobs.
    Periodically use the administrative console or the deleteJob AdminTask to delete old jobs from the job manager database.
16. To recover from a damaged Derby database, do the following:
    1. Stop the dmgr or jobmgr server.
    2. Stop all registered nodes.
    3. If the OTiS database is backed up, restore it from the backup using the restoreJobManager command.
    4. If no backup exists, you can replace the damaged one with the empty OTiS database from the newly created dmgr or jobmgr profile by copying the OTiS directory from the newly created profile. You can remove the newly created profile when finished.
    5. For any nodes previously registered, do the following:
       1. Go to the registered node profile_root/config/cells/cell_name/managednodes directory and delete the directory.
       2. Go to the profile_root/devicestores directory and delete any files under this directory.
       3. Register the node again.
    6. Start the dmgr or jobmgr server.
    7. Restart the registered nodes.