Autosys job showing multiple running instances

[Keith Alderson]

Hi All, 

Have been searching for an answer to this as we have been asked by our client to kill off the 2 running instances of 1 Autosys job.
There was a current running instance which I have killed off but there appears to be a prior instance from a week ago that is still showing as running as well. Is there any way to kill this off or does it even need to be killed off if the job has run 28 times since then successfully and with the most recent run killed off?

Thanks

------------------------------
2IC Midrange Applications
DXC Technology
------------------------------

[Reginold DAlmeida]

Hi Keith,

You can use the parameter term_run_time: mins

The term_run_time attribute specifies the maximum run time (in minutes) that the job you are defining should require to finish normally. If the job runs longer than the specified time, CA Workload Automation AE terminates it.

https://docops.ca.com/ca-workload-automation-ae/11-4-2/en/reference/ae-job-information-language/jil-job-definitions/term_run_time-attribute-specify-the-maximum-runtime

[Doug Hebard]

Reginold -  We use term_run_time  and recently upgraded to 11.3.6 SP8  ( AIX master / Linux client ) and we are seeing issues with the jobs not terminating.

we see this in the logs -  
​

autosys@autosysp01:/export/ha/home/autosys>grep dsbospltf_1 /opt/CA/WorkloadAutomationAE/autouser.C01/out/event_demon.C01
[08/02/2019 11:00:01]      CAUAJM_I_40245 EVENT: CHK_TERM_RUNTIME  JOB: dsbospltf_1
[08/02/2019 11:00:01]      CAUAJM_E_10373 Job <dsbospltf_1> has no extended runtime information.
[08/02/2019 11:00:01]      CAUAJM_I_40245 EVENT: ALARM            ALARM: KILLJOBFAIL      JOB: dsbospltf_1
[08/02/2019 11:00:01]      <Job <dsbospltf_1> is missing runtime information required for the KILLJOB operation.>

autosys@autosysp01:/export/ha/home/autosys>jr dsbospltf_1

Job Name                                                         Last Start           Last End             ST/Ex Run/Ntry Pri/Xit
________________________________________________________________ ____________________ ____________________ _____ ________ _______
dsbospltf_1                                                      08/01/2019 19:00:00  -----                RU    34059390/1

Any thoughts?  I have a inc/ticket open but have not had a response.

[Leroy Zine]

Hi Doug,
  Did you get an answer on this issue?  If so, can you share it as I'm seeing this in one of our environments?

------------------------------
AutoSys Engineer
Bank of America
------------------------------

[Doug Hebard]

No we have not resolved this issue..  We recently upgraded to  sp8 on master and clients, separated our database to our private cloud and are on oracle 18c.. 

do you see this in your logs? 
 ​[08/02/2019 11:00:01]      CAUAJM_I_40245 EVENT: CHK_TERM_RUNTIME  JOB: dsbospltf_1
[08/02/2019 11:00:01]      CAUAJM_E_10373 Job <dsbospltf_1> has no extended runtime information.
[08/02/2019 11:00:01]      CAUAJM_I_40245 EVENT: ALARM            ALARM: KILLJOBFAIL      JOB: dsbospltf_1
[08/02/2019 11:00:01]      <Job <dsbospltf_1> is missing runtime information required for the KILLJOB operation.>

[Leroy Zine]

​Ours wasn't using a term_run_time as the job was just stuck in a running state for 3 days.  When they tried to kill it, that's when those same messages appeared (excluding your CHK_TERM_RUN_TIME).

------------------------------
AutoSys Engineer
Bank of America
------------------------------

[Reginold DAlmeida]

Hi Doug,

You can run a chase command to verify if the job is still active in the agent.

[Steve Carrobis]

Once a new run number is created you cannot kill the "other running process. if it indeed is running. you need to go on to the agent server and physically kill the process.
is this running on an F5 loadbalancer or netscaler device? perhaps that is why autosys is getting confused. if people forcestart. those are the hazards. 
I am unsure how new you are to autosys so allow this old dog to explain.
AutoSys is WYSIWYG. there is no yesterday run today's run or tomorrow's run. there's only the current run.
If someone forcestarts a currently running process. that old run is now off on it's own. will it come back? maybe and when it does will cause issues. because that will come back with it;s completion, possibly trigger something then the force run may come back with its and possible restart stuff already run . 
This is why people need to be careful. Autosys has always had issues with DNS aliases / Netscaler/loadbal;ancers. in knowing which machine it is being sent to so it knows who to "chase" down. so chase on running prcoesses if you have loadbalancer may be wrong more times than correct.
AutoSys Operations should be done carefully. and knowing how the product behaves.,
Good Luck. 
Steve C.