Automic Workload Automation

 View Only
Expand all | Collapse all

Job Failure doesn't trigger notification when job run crosses to next day

  • 1.  Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 19, 2019 02:01 PM
    Hi All,

    I have observed a strange thing that a call alert / notification isn't getting generated and triggered when the job activated on a day runs long and fails on another day. The job is configured to send a failure notification and it has sent the notification when the failure occurs on the same day. But when the job failure occurs on the next day of job activation it doesn't send an alert. Has anyone encountered such an issue? Any pointers to rectify such this issue will be helpful. 

    Regards,
    Shashank Singh


  • 2.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 21, 2019 12:41 AM
    Hi,

    there are several ways to send a notification in case of an error, e.g. post script, post conditions, successor in workflow, attributes tab of workflow, etc.

    Can you please specify how it is implemented in your system?

    Best regards
    Tim

    ------------------------------
    Automation Evangelist
    Fiducia & GAD IT AG
    ---
    Mitglied des deutschsprachigen Automic-Anwendervereins FOKUS e.V.
    Member of the German speaking Automic user association FOKUS e.V.
    ------------------------------



  • 3.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 21, 2019 01:04 AM
    Hi Tim,

    Alert mechanism is implemented in Post-Conditions. 
    Job's task properties --> Post Conditions.

    It is a normal standalone job (JOBS) directly scheduled via Schedules (JSCH).

    Regards,
    Shashank Singh


  • 4.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 21, 2019 05:18 AM
    Edited by PROSENJIT MAJI Oct 21, 2019 05:18 AM
    Hi Shashank, 

    So far with the available information, I believe, you are not hitting into some 'known' generic issues as such.

    Would you be able to post a 'Postcondition' properties screen for reference ?

    ------------------------------
    Prosenjit
    ------------------------------



  • 5.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 21, 2019 05:29 AM
    Here is the screenshot. 



  • 6.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 21, 2019 06:26 AM
    I would try to remove the checkmark at "Repeat" because it seems at least to be unnecessary. Perhaps it interfers with the following "Execute".

    ------------------------------
    Automation Evangelist
    Fiducia & GAD IT AG
    ---
    Mitglied des deutschsprachigen Automic-Anwendervereins FOKUS e.V.
    Member of the German speaking Automic user association FOKUS e.V.
    ------------------------------



  • 7.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 21, 2019 07:53 AM
    Hi Tim,

    The checkmark has come by default. I never set that up. Also I do not think it to be messing with the alerting since the alerting works fine for the jobs starting and failing / aborting on same day. The issue is with jobs running longer and aborting on next day wherein the alert fails to generate.

    Regards,
    Shashank Singh


  • 8.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 21, 2019 10:40 AM
    Edited by PROSENJIT MAJI Oct 21, 2019 10:00 PM
    Hi Shashank,

    The instance gets long runner that really gets "Abend" status next day ?
    Or
    It gets any other status like "Ended_empty" etc. ? 

    Did you check the call operator statistics ? Did that get execute by the given property and still you did not receive email ? 

    I am not sure on this because the behavior explained is not expected & so far not aware of such bugs.

    ------------------------------
    Prosenjit
    ------------------------------



  • 9.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 21, 2019 11:58 PM
    Hi Prosenjit,

    Yes the job runs long and actually fails. Status - ENDED_NOT_OK - aborted.
    The Call Operator statistics also do not show the Call Alert generated which means the execution of the alert isn't happening.
    I suspect this to be a bug unless someone can guide / point me to any missing configuration.

    Regards,
    Shashank Singh



  • 10.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 22, 2019 12:31 AM
    Hi Shashank,

    I also believe that this is a bug. This may be due to the fact that the JSCH is assigned a new instance at the period change and the check is therefore not performed.

    The only uncommon setting I can see at the moment is the unnecessary marked checkbox "Repeat" but I also do not believe that this is the cause.

    May I suggest that you put the jobs into workflows and then do the results check there?

    What version are you using?

    Best regards,
    Tim

    ------------------------------
    Automation Evangelist
    Fiducia & GAD IT AG
    ---
    Mitglied des deutschsprachigen Automic-Anwendervereins FOKUS e.V.
    Member of the German speaking Automic user association FOKUS e.V.
    ------------------------------



  • 11.  RE: Job Failure doesn't trigger notification when job run crosses to next day

    Posted Oct 22, 2019 12:54 AM
    Hi Tim,

    I am still on 11.2 (11.2.8 HF2).
    Yes I will try and put the jobs in workflow and test. I do believe it should work when put in workflow as we have several workflows running over to the other day and did not find issues with respect to job alerts in those workflows.
    The issue is only with individually scheduled jobs and as you said it can be due to the reloading of JSCH at period turnaround.

    Regards,
    Shashank Singh


  • 12.  RE: Job Failure doesn't trigger notification when job run crosses to next day
    Best Answer

    Posted Oct 22, 2019 01:19 AM
    Edited by Diane Craddock Oct 23, 2019 09:11 AM

    Thanks Tim. I missed one information. The job is added to the Scheduler & the property is 'Of the schedule' for 'the job'. 

    following might be the story : 
    ---------------------------------------------------
    - The schedule object generally has only one ACTIVE instance at a time as it gets reloaded at given time. 
    - The instance of the job failed next day & the parent where properties are defined is already ended. So no more monitoring. 
    - This is why next day when the job gets failed - No parent exists to monitor the status & report it back. 
    - This is why if you have some time event runs 24X7 from a schedule, if you monitor older instance of scheduler Monitor you might see those file events as - Active / Sleeping. Because when the scheduler got ENDED_OK last time, AE stopped monitoring of it's child tasks. If anything gets completed before the parent is ENDED_OK it reflects the accurate status.  

    ---------------------------------------------------
    I am not sure if we can call it as bug but Automic can improve their product to reflect relevant info.

    Sorry I put too many irrelevant queries earlier. 

    ------------------------------
    Regards,
    Prosenjit
    ------------------------------