dba_scheduler_job_log tips 

 

By Kamran Agayev Agamehdi

The dba_scheduler_job_log view can be used to view log entries for previously executed jobs. This view displays log information for all dbms_scheduler jobs in the database. When creating a job, we define log_history parameter and logs in that view are available for number of days as it specified in log_history parameter (default is 30)

One of the main columns in dba_scheduler_job_log is the STATUS column. This column gives us the status of the operation. Possible values for this column are dependent on the value in the OPERATION column. In most cases, STATUS will be NULL. Only for job run operations will it have a value.

• STATUS will be NULL when OPERATION is one of the following:
• CREATE - Job was created
• UPDATE - One or more job attributes have been modified
• ENABLE - Job has been enabled
• DISABLE - Job has been disabled
• COMPLETED - For repeating jobs only, job has reached its end date or maximum number of runs
• BROKEN - Job has reached its maximum number of failures

STATUS can be SUCCEEDED (job run completed successfully), FAILED (job run failed), or STOPPED (job run was stopped) when OPERATION is one of the following:

• run - Regular job run
• retry_run - Job is being retried because the previous run resulted in an error and RESTARTABLE is set to TRUE
• recovery_run - Job is being rerun because the database went down, or the job slave crashed and RESTARTABLE is set to TRUE

Let's test the use of dba_scheduler_job_log step-by-step. In below example, we create a table and a job which will run every minute and insert one row in that table.

We will set logging_level of that schedule to logging_full to catch all changes made to that table and view all operations in dba_scheduler_job_log view

 

SQL> create table job_table1 (id number);

Table created.

 

SQL> BEGIN

sys.dbms_scheduler.create_job(

job_name => 'JOB1',

job_type => 'PLSQL_BLOCK',

job_action => 'INSERT INTO job_table1 values(1);',

repeat_interval => 'FREQ=MINUTELY',

start_date => sysdate,

enabled => true);

END;

/

PL/SQL procedure successfully completed.

 

SQL> select * from job_table1;

 

        ID

----------

         1

 

SQL> begin

  DBMS_SCHEDULER.SET_ATTRIBUTE('JOB1','logging_level',DBMS_SCHEDULER.LOGGING_FULL);

  end;

  /

PL/SQL procedure successfully completed.

 

SQL> begin

  2  DBMS_SCHEDULER.DISABLE('JOB1',TRUE);

  3  end;

  4  /

PL/SQL procedure successfully completed.

 

SQL> begin

  2  DBMS_SCHEDULER.ENABLE('JOB1');

  3  END;

  4  /

PL/SQL procedure successfully completed.

 

SQL> drop table job_table1;

Table dropped.

 

Now we can query dba_scheduler_job_log to see the values:

select
   log_id,
   to_char(log_date,'dd/mm/yyyy hh24:mi:ss') log_date,
   substr(job_name, 1,20)                    job_name,
   substr(status, 1,10)                        status,
   additional_info                          operation
from
   dba_scheduler_job_log
where
   job_name='JOB1'
order by
   log_date ;

 

LOG_ID        LOG_DATE             JOB_NAME    STATUS         OPERATION    

------ ----------          ---------   ----------            -----------

450    31/05/2009 17:27:58 JOB1        SUCCEEDED

455    31/05/2009 17:28:57 JOB1        SUCCEEDED

459    31/05/2009 17:29:49 JOB1                       FORCE="TRUE", REASON="manually disabled"

463    31/05/2009 17:30:24 JOB1                       REASON="manually enabled"

467    31/05/2009 17:30:57 JOB1        SUCCEEDED

472    31/05/2009 17:31:57 JOB1        SUCCEEDED

477    31/05/2009 17:32:57 JOB1        FAILED

 

As it seen above, first two minutes, our job ran successfully. After we disabled the job, the information about it was logged in operation column. Then we enabled it, it also logged in operation column. Later, it ran two times successfully.

At the end, we dropped the table, thus our job failed and was logged in the dba_scheduler_job_log view with failed status.

---