|
View previous topic :: View next topic
|
| Author |
Message |
jzhardy
Active User
Joined: 31 Oct 2006
Posts: 131
Location: brisbane
|
|
|
|
I have an need to synchronize jobs running across DEV and PROD LPARs.
Multiple instances of job P needs to run in Production in sequence 1... N, where N is sometimes large (1000 +)
Job P(1) creates file X on a shared disk pack. Job P(1) (in the final step) submits Job D(1) which runs in the Development LPAR
D(1) requires SHR access to X
the problem is that D(1) needs to en-queue X before job P(2) en-queues X (to overwrite), otherwise job D(1) will fail and file X is lost.
so ... I added a job delay (5 seconds) as a final step in job P1. This works about 95% of the time, as there are some conditions that will delay jobs D(n) outside of my control. Even setting it to 1 minute won't help.
so .. looking for suggestions. The constraints are such that I can't identify the output file X with individual jobs (ie, X(1), X(2).. X(n) ) , so X is the only file available for output.
is there a shared resource between LPARs, something that I can use as a semaphore ? |
|
| Back to top |
|
 |
prino
Senior Member
Joined: 07 Feb 2009
Posts: 1306
Location: Vilnius, Lithuania
|
|
|
|
| That's why they invented schedulers... |
|
| Back to top |
|
|
Willy Jensen
Active Member
Joined: 01 Sep 2015
Posts: 712
Location: Denmark
|
|
|
|
Do you have dasd shared between the sysplexes? Then you can have a program (REXX?) reading and updating a shared dataset.
I also prefer a proper scheduler. |
|
| Back to top |
|
|
jzhardy
Active User
Joined: 31 Oct 2006
Posts: 131
Location: brisbane
|
|
|
|
| Agree. We have OPC which I have used extensively in the past, and this woudl be the ideal tool, but there is a process for getting new applications added which takes considerable time in our shop. This piece of work is to meet an ad-hoc request from management. |
|
| Back to top |
|
|
jzhardy
Active User
Joined: 31 Oct 2006
Posts: 131
Location: brisbane
|
|
|
|
| but yes ... the REXX option is probably the most expedient .. |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10873
Location: italy
|
|
|
|
| frankly I do not see how using rexx to process a dataset would be different from doing the same with a <any other language> program |
|
| Back to top |
|
|
prino
Senior Member
Joined: 07 Feb 2009
Posts: 1306
Location: Vilnius, Lithuania
|
|
| Back to top |
|
|
Willy Jensen
Active Member
Joined: 01 Sep 2015
Posts: 712
Location: Denmark
|
|
|
|
| Quote: |
| frankly I do not see how using rexx to process a dataset would be different from doing the same with a <any other language> program |
It isn't. It's just that I have done something similar using REXX because it was easy to do and this part of the application was not performance sensitive. |
|
| Back to top |
|
|
|
|
