Topic
  • 6 replies
  • Latest Post - ‏2012-04-06T16:01:22Z by SystemAdmin
cwalsek
cwalsek
14 Posts

Pinned topic Job Scheduler

‏2012-01-24T16:23:17Z |
Has anyone encountered this? When I start the server with Modern Batch Feature Pack I encounter Job Scheduler errors. This is a sudden change. All was working fine previously. Now when I submit xJCL jobs, they just sit there in a submitted status. When I try changing their status to Resume or Restart, console reports job is in an invalid state. Looks like something broke inadvertently, but I don’t see what. Thanks

1/24/12 10:43:31:064 EST 00000030 ResRefListImp W CNTR0076W: A reference to an ResourceRef could not be found in the deployment descriptor for the WebApp or EnterpriseBean with the display name JobScheduler.
1/24/12 10:56:04:501 EST 0000002f SchedulerSing W Job ReadDB2:00252 is not resumeable: invalid state: submitted

1/24/12 10:56:17:323 EST 00000032 SchedulerSing W Job ReadDB2:00252 is not restartable: invalid state: submitted
Updated on 2012-04-06T16:01:22Z at 2012-04-06T16:01:22Z by SystemAdmin
  • timfanelli
    timfanelli
    2 Posts

    Re: Job Scheduler

    ‏2012-01-24T16:39:24Z  
    Hi Charles -

    A "submitted" job can not be cancelled or restarted - so those messages are normal. It's just not a valid state transition.

    The real problem, as you noted, is that the job is not "dispatched" and therefor never enters the "executing" state. Could you post the entire joblog so we could review it? In particular, the last few lines should give some indication of why the job is not dispatched.

    Common problems include:
    • The job scheduler component believes there are already too many jobs executing on the end-point server, and it's waiting for them to end before dispatching more work. Unless you are doing load-testing, this is not likely the case.
    • The target application is not started. This may have happened if you updated your application EAR, and forgot to start the application.
    • The JobScheduler component did not initialize during server startup. This may happen if your back-end database is offline. This is an unlikely scenario for you -- as I recall you are using a standalone server w/ a derby backend. In this environment, you would not be able to reach the JMC web application to submit your jobs if the JobScheduler were not properly init'd.

    Thanks,
  • cwalsek
    cwalsek
    14 Posts

    Re: Job Scheduler

    ‏2012-01-24T17:26:24Z  
    Hi Charles -

    A "submitted" job can not be cancelled or restarted - so those messages are normal. It's just not a valid state transition.

    The real problem, as you noted, is that the job is not "dispatched" and therefor never enters the "executing" state. Could you post the entire joblog so we could review it? In particular, the last few lines should give some indication of why the job is not dispatched.

    Common problems include:
    • The job scheduler component believes there are already too many jobs executing on the end-point server, and it's waiting for them to end before dispatching more work. Unless you are doing load-testing, this is not likely the case.
    • The target application is not started. This may have happened if you updated your application EAR, and forgot to start the application.
    • The JobScheduler component did not initialize during server startup. This may happen if your back-end database is offline. This is an unlikely scenario for you -- as I recall you are using a standalone server w/ a derby backend. In this environment, you would not be able to reach the JMC web application to submit your jobs if the JobScheduler were not properly init'd.

    Thanks,
    Thanks Tim! I see what you mean. Joblog says job class Default is running at full capacity. I’ve been running lotta jobs. I often just stop their execution in mid-stream. I thought just removing job (log) via console was enough. Looks like I may have left them active. This particular job needs an EJB but still, I’d like to get it dispatched.

    I've attached joblog. Where do we go from here? Thanks
  • timfanelli
    timfanelli
    2 Posts

    Re: Job Scheduler

    ‏2012-01-24T19:10:11Z  
    • cwalsek
    • ‏2012-01-24T17:26:24Z
    Thanks Tim! I see what you mean. Joblog says job class Default is running at full capacity. I’ve been running lotta jobs. I often just stop their execution in mid-stream. I thought just removing job (log) via console was enough. Looks like I may have left them active. This particular job needs an EJB but still, I’d like to get it dispatched.

    I've attached joblog. Where do we go from here? Thanks
    Interesting -- we'll need to do some problem determination to figure out how your system got into that state in the first place. When a job reaches an end-state by any means, it should be removed from it's job class' running-job-count.

    I will send you a note off-line, and then follow up again here as soon as we have a resolution so we can share what we find.

    Out of curiosity - have you restarted the server since this began happening? That should reset the count of jobs running in each job class.

    Thanks,
  • cwalsek
    cwalsek
    14 Posts

    Re: Job Scheduler

    ‏2012-01-24T21:10:49Z  
    Interesting -- we'll need to do some problem determination to figure out how your system got into that state in the first place. When a job reaches an end-state by any means, it should be removed from it's job class' running-job-count.

    I will send you a note off-line, and then follow up again here as soon as we have a resolution so we can share what we find.

    Out of curiosity - have you restarted the server since this began happening? That should reset the count of jobs running in each job class.

    Thanks,
    Thanks for taking a look. This one has me stumped. Yes, I’ve stopped and started server multiple times. I see there is a Job Scheduler link in Admin Console, under System Administration category. I thought this link might give me a view of Job Classes and the jobs WebSphere might think still exist.

    However, I don’t see anything. Maybe it is because I’m running Feature Pack on WAS 7.0 server, rather than WAS 8.0 server. I’ve attached pic of admin console.
  • cwalsek
    cwalsek
    14 Posts

    Re: Job Scheduler

    ‏2012-01-27T21:36:32Z  
    • cwalsek
    • ‏2012-01-24T21:10:49Z
    Thanks for taking a look. This one has me stumped. Yes, I’ve stopped and started server multiple times. I see there is a Job Scheduler link in Admin Console, under System Administration category. I thought this link might give me a view of Job Classes and the jobs WebSphere might think still exist.

    However, I don’t see anything. Maybe it is because I’m running Feature Pack on WAS 7.0 server, rather than WAS 8.0 server. I’ve attached pic of admin console.
    Thought this might be useful for other newbees encountering this same situation:

    Job xxxxx cannot be dispatched at this time as its job class Default is running with full capacity.

    I’m running locally in RAD 8.0.3. I was able to resolve the problem by setting CONCJOBCOUNT to 0 in the JOBCLASSMAXCONCJOBS table. This is a Derby database. Use admin console to get the database location from the datasource (jdbc/lrsched). Stop the server. Using RAD Data Perspective, create a new connection to this database and execute the update SQL. Close the connection when done and restart the server. Should be OK. Additional experience should help identify practices to prevent capacity problem from surfacing again.
  • SystemAdmin
    SystemAdmin
    783 Posts

    Re: Job Scheduler

    ‏2012-04-06T16:01:22Z  
    • cwalsek
    • ‏2012-01-27T21:36:32Z
    Thought this might be useful for other newbees encountering this same situation:

    Job xxxxx cannot be dispatched at this time as its job class Default is running with full capacity.

    I’m running locally in RAD 8.0.3. I was able to resolve the problem by setting CONCJOBCOUNT to 0 in the JOBCLASSMAXCONCJOBS table. This is a Derby database. Use admin console to get the database location from the datasource (jdbc/lrsched). Stop the server. Using RAD Data Perspective, create a new connection to this database and execute the update SQL. Close the connection when done and restart the server. Should be OK. Additional experience should help identify practices to prevent capacity problem from surfacing again.
    i want to submitt a job but i have an error msg
    i don't know what to do here is the Systemerror log

    email:s.daouiji@gmail.com