Is there a way to determine Clist usage in an LPAR?

mnaidoo2 · New User Joined: 02 Mar 2021 Posts: 3 Location: Australia

Hi All,
Is there a way to determine CLIST usage on an LPAR:
- CLIST name
- extent of usage (x number of time per month)
- dataset the CLIST resides in

I have looked at the SMF type 32 records but it does not capture the CLIST name.

Pedro · Posted: Thu Mar 04, 2021 11:02 pm

May I ask what you are going to do with the information?

mnaidoo2 · New User Joined: 02 Mar 2021 Posts: 3 Location: Australia

Busy with a transition of a mainframe system to Microfocus ES on AWS. Need Clist usage to determine conversion effort. (CLIST is not supported on Microfucus ES)

Willy Jensen · Posted: Fri Mar 05, 2021 3:47 am

Only commercial products as far as I am aware. CA-TSOMON springs to mind, otherwise one or two products that monitors PDS usage could be used if you can identify CLIST libraries. Which may be another problem as CLIST and REXX can and often do coexist in the same library.

Pedro · Posted: Fri Mar 05, 2021 9:54 am

mnaidoo2 · New User Joined: 02 Mar 2021 Posts: 3 Location: Australia

Hi Pedro, We have the list of CLIST libraries on the LPAR/s. Some of them are huge with over 800 CLIST members. Therefore we need to determine which of the CLISTs are used and frequency of use as well. Failing any easy way, we will have to insert a stub in each CLIST. I liked the suggestion to use CA-TSOMON but we don't have the software.

sergeyken · Posted: Fri Mar 05, 2021 7:41 pm

This is a common problem with any initially bad configuration design of the whole system/environment. And the problem is not only with CLIST, but - even more important - with LOADLIB/LINKLIB libraries... In general, such situations cannot be resolved without full system re-design.

I remember one huge company where the total number of common load libraries in LINKLIST was... 97 DSNAMEs

This happened because nobody in this world is able to remember: which library was added when, and by whom, and what was the purpose of particular libraries. Nobody tries to resolve this issue since many years, and it’s getting worse, and worse, and worse...

Pedro · Posted: Fri Mar 05, 2021 10:41 pm

Do you have a test system where you can run a test workload?

My suggestion is to rename the clist data sets and see which jobs fail. And then later provide empty data sets and see which clists are called (but are not found). And then later provide those missing clists to see which other clists are called as subroutines. You might need a few iterations of this.

And then add your stub so that you can count how often it is used.

enrico-sorichetti · Posted: Sat Mar 06, 2021 3:09 am

in a well planned conversion project
usually some financial resources are allocated for the acquisition of the appropriate tracking tools

expat · Posted: Mon Mar 08, 2021 9:42 am

Pedro · Posted: Mon Mar 08, 2021 11:22 pm