[Bug] [zos_job_submit] Incorrect display of DD byte_count

[gngrossi]

Is there an existing issue for this?

• [X] There are no existing issues.

Bug description

The job is submitted, runs and the the output is proper except for the DD byte_counts are not matching with what's on JES spool.

IBM z/OS Ansible core Version

v1.9.0

IBM Z Open Automation Utilities

v1.2.5 (default)

IBM Enterprise Python

v3.11.x (default)

ansible-version

v2.16.x (default)

z/OS version

v2.5 (default)

Ansible module

zos_job_submit

Playbook verbosity output.

• name: Print DDs from Job {{jobname}} stats debug: msg:
  • "{{jobdds[item | int].ddname}}"
  • "{{jobdds[item | int].id}}"
  • "{{jobdds[item | int].record_count}}"
  • "{{jobdds[item | int].stepname}}"
  • "{{jobdds[item | int].byte_count}}" with_sequence: start=0 end="{{jobddl | int-1}}"

Ansible output

JES spool

Ansible configuration.

No response

Contents of the inventory

No response

Contents of group_vars or host_vars

No response

[gngrossi]

Still an issue with ibm.ibm_zos_core:==1.10.0-beta.1 and zoau 1311 and python 3.12.1

[ddimatos]

Looks like we used to access it via ISFEXEC ST , after the migration to ZOAU , someone calculated it incorrectly with dd["byte_count"] = single_dd["record_length"].

[gngrossi]

@ddimatos Any additional documentation needed? thanks

[ddimatos]

@gngrossi - not at the moment, I am currently looking for an efficient way to calculate this and most everything I am coming up with has some sort of corner case and of course performance or introduces a new dependency if we do it in Ansible.

The solution I like the most is to open a enhancement request on ZOAU and have them using the z/OS MVS Subsystem Interface because then they are reading a control block, its efficient and removes the corner cases when trying to calculate this but I would not expect this to be available till the end of the year. Thus this would have to be handled as a known issue.

Could you tell me the use case for the Byte-cnt, I am curious how others use it.

[gngrossi]

@ddimatos Sorry, I don't have a real use case for the byte count data. IMO, if it's going to be available as a statistic, then it should be accurate otherwise it leads to asking the question "what else isn't right?"

In this example, I was interested in seeing the results of using JES spool data encryption which includes compression.

thanks

[ddimatos]

Understood, we should include a few statistical test cases ensuring accuracy, I'll continue to look into this, this helps, we also want the same.

[richp405]

Related to Jira 10606 - ZOAU estimates a Q3/Q4 enhancement to support byte-cnt for DDs.

[ddimatos]

I went through various options, to handle this in Ansible none of those are efficient and without possible corner case issues, I will add this as a known issue to our release notes and wait for ZOAUs change that can access the control block in an efficient and accurate way.