Hi Loren,
Would think that resource is a great way to address it.
Here is a test I did:
- My Proc:
APPL CAL
JCLLIB ZHALI13.PSS.JCLLIB
SYMBOL %
JOB CYBLZ.1 TASK SELFCOMPLETING
RELDELAY 1
RUN TODAY
ESPNOMSG AJ CYBLZ.2 RESET DELAYSUB('REALNOW +
PLUS 20 SECONDS') APPL(%ESPAPPL..%ESPAPGEN)
REL CYBLZ.2
ENDJOB
JOB CYBLZ.2
CCCHK RC(8) FAIL STOP
RUN TODAY
ENDJOB
In ESP auditlog:
- Application was created:
16.02.47 FRI 12JUL19 ESP6030I APPLGEN: Appl CAL.9 created, slot 0000051B
- CYBLZ.1 was readied one minute later, because of RELDELAY 1, and at the same time it reset the Delaysub time for CYBLZ.2:
16.03.47 FRI 12JUL19 ESP6633I APPLMGR: APPL CAL.9 JOB CYBLZ.1 READIED
16.03.47 FRI 12JUL19 ESP6619I APPLMGR: APPL CAL.9 JOB CYBLZ.2 TIMES RESET BY USERZ
- CYBLZ.2 was readied 21 seconds after the time was reset, NOT 20 seconds after the application was created:
16.04.08 FRI 12JUL19 ESP6633I APPLMGR: APPL CAL.9 JOB CYBLZ.2 READIED
It proved that the REALNOW in the AJ RESET command was the current time when the command was issued, NOT when appl was created (same time when event fired).
Hope this helps,
Lucy
Original Message:
Sent: 07-12-2019 11:01 AM
From: Loren Watts
Subject: NT_JOB job delay smaller than the 1 minute possible with RELDAY 1
Thank you for your help Lucy. For my current need, I was able to use a RENEWABLE Resource at the APPL level. I have about 200 jobs in one APPL which were running concurrently. My customer needed them staggered; they had asked for a timing solution, but a resource meets their need.
I would appreciate your coaching regarding REALNOW. If it is okay with you, I will do that in a separate conversation, to keep it focused. I will first do a review on my own, so I make the best use of your time and brainpower.
------------------------------
Cigna
Original Message:
Sent: 07-11-2019 03:45 PM
From: LIXIN ZHANG
Subject: NT_JOB job delay smaller than the 1 minute possible with RELDAY 1
Hi Loren,
Firstly the REALNOW is the current local time when it's used. With example provided by Mike, it's the current local time when TASK DELAYJOB runs and AJ RESET command is issued. So it should work for jobs that run hours later after the event is triggered.
Secondly would you mind to give more detail on the reason of the delay? Maybe we can figure out a way without the delay.
Lastly if this feature is important, would you post an idea to raise a enhancement request again?
Thank you,
Lucy
Original Message:
Sent: 07-09-2019 11:28 AM
From: Loren Watts
Subject: NT_JOB job delay smaller than the 1 minute possible with RELDAY 1
Thank you for the idea Michael. That is a possibility.
The shortcoming for me of REALNOW is that REALNOW reflects the actual time of event trigger. Much of my organization's setup has the event being triggered to bring in an appl generation, but by design our appls do not begin until hours later. This allows us to review and adjust during our business day all the batch processing which won't begin until the evening.
I have wanted for years for CA to provide a time delay relative to a job's predecessor job - not a time reference bound to when the event was triggered. JobA runs - wait twenty seconds - JobB runs - wait twenty seconds - JobC runs.
For this simple design, the length of time is in relation to the predecessor job
Don't worry about the triple post. It feels like Broadcom has bugs to work out in their transition. I have a a great deal of trouble using Communities since the CA transition.
------------------------------
Loren Watts - Cigna
Original Message:
Sent: 07-05-2019 04:19 PM
From: Loren Watts
Subject: NT_JOB job delay smaller than the 1 minute possible with RELDAY 1
I'm looking for ideas on having a a delay of seconds between NT_JOB jobs. I use RELDELAY 1 for a one minute delay before a successor job can begin, but I'm looking for a shorter delay. I have thought of a separate job using LINUX_JOB just to do sleep - CMDNAME sleep
ARGS 10s, but want to be sure there is not a built-in ESP feature for a short delay.
Thank you for your thoughts and time.
------------------------------
Loren Watts
Cigna
------------------------------