When a windows job with a success status or max return code of -1 is set on the runtime tab is run, it should always kick off an object (i.e. a CALL object) when it runs, no matter the return code of the job. The job often finishes with a return of 0, however the associated CALL object does not always run - one day it runs as expected, the next it does not. The agent logs and automation engine logs do not have any references to the associated CALL object when this happens
This is reproducible on 12.3.1 through 12.3.5
Steps to reproduce
1) Create an email notification object - make sure it has a recipient
2) Create a windows job with:
Runtime tab: Return code <= -1, action Execute an alternative object, execution of object from step 1 above
Process tab: simple dir command
3) Run the windows job many times. For example:
a) Run it in a workflow and run the workflow 10 times
b) Write a script that uses activate_uc_object to kick off the job 10 times
c) In one example, the JOBS was put in a WF with a script that started traces at the beginning and another at the end that stopped traces; A max parallel was on the WF of 1 else wait and ran it 10 times
d) Another example simply had this WF running every day at a specific time
Expected behavior:
Notification object should be run 100% of the time no matter what happens with the job
Current observed behavior:
Notification does not consistently kick off - sometimes it gets activated 80% of the time, sometimes it's 50% of the time
Release : 12.3
Component : AUTOMATION ENGINE
This is caused by a defect in the code
This has been fixed with Automic Automation 12.3.7 and requires an update to the components: Automation Engine, AWI, Initialdata and Utilities