1 2 Previous Next 25 Replies Latest reply: Mar 27, 2008 4:46 PM by Rnr-Oracle Go to original post RSS
      • 15. Re: limit runtime of a JOB
        592384
        Hi,
        I have also noticed that for some reason the job has been stopped even though i haven't phsyically called the job.stop function. Can this happen and why?
        • 16. Re: limit runtime of a JOB
          Rnr-Oracle
          Hi,

          The only reason a job could say stopped if you didn't call stop_job is if the database went down during the job run or if the process (or oracle session) running the job was abruptly terminated.

          -Ravi
          • 17. Re: limit runtime of a JOB
            600257
            Hi
            I ran into the same problem: A chain job stopped at chain step1.
            select * from dba_scheduler_job_run_details;

            The query results show that the chain failed at step1 with additional_info =
            CHAIN_LOG_ID="12264", STEP_NAME="STEP1", REASON="Stop job with force called by user: 'SYS'"

            I'm sure that nobody was running dbms_scheduler.stop_job (nor drop_job with 'FORCE') and the database was not shutdown. I also verified that the chain was enabled. I've checked the alert file and found no error recorded there. How do I do the further trouble-shooting?

            Thanks!

            Joan
            • 18. Re: limit runtime of a JOB
              600257
              Hi
              I ran into the same problem: A chain job stopped at chain step1.
              select * from dba_scheduler_job_run_details;

              The query results show that the chain failed at step1 with additional_info =
              CHAIN_LOG_ID="12264", STEP_NAME="STEP1", REASON="Stop job with force called by user: 'SYS'"

              I'm sure that nobody was running dbms_scheduler.stop_job (nor drop_job with 'FORCE') and the database was not shutdown. I also verified that the chain was enabled. I've checked the alert file and found no error recorded there. How do I do the further trouble-shooting?

              Thanks!

              Joan
              • 19. Re: limit runtime of a JOB
                600257
                Hi
                I ran into the same problem: A chain job stopped at chain step1.
                select * from dba_scheduler_job_run_details;

                The query results show that the chain failed at step1 with additional_info =
                CHAIN_LOG_ID="12264", STEP_NAME="STEP1", REASON="Stop job with force called by user: 'SYS'"

                I'm sure that nobody was running dbms_scheduler.stop_job (nor drop_job with 'FORCE') and the database was not shutdown. I also verified that the chain was enabled. I've checked the alert file and found no error recorded there. How do I do the further trouble-shooting?

                Thanks!

                Joan
                • 20. Re: limit runtime of a JOB
                  Rnr-Oracle
                  Hi,

                  For a chain it is possible that one of the chain rules has a "stop" action that causes a step to be stopped so you could check this too. Also if the chain is ended (via an end action) then still running steps will be stopped, so that is another possibility.

                  Is this reproducible, do you have a test case ?

                  Thanks,
                  Ravi.
                  • 21. Re: limit runtime of a JOB
                    605203
                    Hi Ravi,

                    I created the same sniper jobs to kill the jobs that run overdue(job_over_max).

                    I am using dbms_lock.sleep to suspend the execution in jobs for a 5 mins but, when 2 or more jobs happen to run continuously. Even though i have set the max_run_duration to 2min, its not rasing the event and jobs are not stopped.

                    In-case of single jobs running at a point of time it works fine, job stopped when it exceeds the max_run_duration.

                    Question:-

                    1. Is it because of dbms_lock in every jobs ?
                    2. Is it because of problem in event generation on multiple jobs running at a point of time?

                    Thanks,
                    Ram
                    • 22. Re: limit runtime of a JOB
                      605203
                      I removed dbms_lock from the jobs but, still facing the same issue.

                      Jobs were stopped if it runs only one job at any point of time.

                      If two or more jobs are running at a point of time even though the start time and jobs max_run_duration are different, jobs are not getting stopped.

                      Requirement:

                      I classifed the jobs based on the priority, if the priority is high it will have a max run duration and low priority will have minimum.

                      i.e., high priority -> max_run_duration = 60 min
                      medium ,, -> ,, = 10 min
                      low priority -> ,, = 5 min


                      Any workaround or how to handle this scenerio....

                      If its a bug pls give me some relative url abt this bug

                      Message was edited by:
                      user602200
                      • 23. Re: limit runtime of a JOB
                        Rnr-Oracle
                        Hi,

                        What method are you using to have the same sniper job wait for both jobs ? You should be using the second piece of code above where a single sniper job waits on all jobs not the first where a single sniper job only waits on a single job.

                        Assuming you are using the second example above, that example includes a test case which shows that it can stop 2 jobs even if they are running at the same time. Are you sure that what you are seeing is only one of the jobs being stopped ?

                        Are both of the jobs finishing at exactly the same time ? If they are, as I point out above the sniper job will ignore the second snipe because it is still processing the first snipe. If you want this to work you will have to either use one sniper job per job or upgrade to 11g where you can set the parallel_instances flag for the sniper job so it can send out multiple snipes simultaneously.

                        Hope this helps,
                        Ravi.
                        • 24. Re: limit runtime of a JOB
                          605203
                          i am using the oracle 10g Enterprise Edition Release 10.2.0.3.0.

                          Any patches available for these....?
                          • 25. Re: limit runtime of a JOB
                            Rnr-Oracle
                            Hi,

                            Unfortunately the PARALLEL_INSTANCES attribute which allows an event-based job to respond to multiple events occurring at the same time is new to 11g.

                            Thanks,
                            Ravi.
                            1 2 Previous Next