ansible-collections / ibm_zos_core

Red Hat Ansible Certified Content for IBM Z
78 stars 44 forks source link

[Enhancement] Enhancmeent for zos_operator to consider adopting shorter increments than a second #1753

Open sjwar455 opened 1 month ago

sjwar455 commented 1 month ago

Is there an existing issue for this?

Enhancement or feature description

An alternative/additional solution/feature request for the following issue: https://github.com/ansible-collections/ibm_zos_core/issues/1752

Originally asked a question about this on the IBM Z Enthusiast Discord - you can see the full discussion here: https://discord.com/channels/880322471608344597/1291085161462501469

Ansible module

zos_operator

ketankelkar commented 5 days ago

For the singly-threaded Ansible automation, introducing such an option would have little impact because the overhead to form and close the SSH connection between the controller and the managed node would take much longer than the fractional seconds saved by allowing for smaller units of wait_time_s.

Where things could get interesting is if there is a multi-threaded approach, where the overhead overlaps each other, so even though each 'operation' takes second(s) to form the connection, run the command, gather the output and close the connection, each item in the 'overhead' can overlap and take up the "real-time" same few second(s). In such a case, any time savings on the target system would add up and introducing a smaller unit of time for wait_time_s would yield more significant performance boost.

Another idea where the feature request in this ticket makes sense is if an adjacent feature were introduced to "stack" operator commands such that in a single connection, multiple operator commands are run sequentially before returning the results of the full payload in one go.

This item warrants additional research