Set runtime shell requirements within Jenkins Pipeline

[TerrenceMcGuinness-NOAA]

Description

This PR updates the Jenkins Pipeline script to source gw_setup.sh from within the PR itself for when ever a shell command is issued that needs it.

Currently the runtime bash shell environment requirements for running the global-workflow (set by sourcing $HOMEgfs/workflow/gw_setup.sh) is established by the user's shell that makes the connection from the Java remote agent to the Jenkins Controller. This update simply sources gw_setup.sh directly during run time when ever needed instead.

NOTE: This does not apply for GitHub CLI and pyGitHub since they are not currently supported by Spack Stack and therefore not incorporated in to gw_setup.sh. These remaining dependances are the responsibility of the role account holders that manage the Jenkins Controller to Node Agent connections.

Type of change

• Bug fix (fixes something broken)
• New feature (adds functionality)
• Maintenance (code refactor, clean-up, new CI test, etc.)

[AnilKumar-NOAA]

We have added HOMEGFS to the runtime shell environment for gaea C5/C6 both in our epic role account .bashrc file as follows:

Added for Global Workflow

if [ -n "${HOMEGFS+x}" ] && [ -n "${HOMEGFS}" ]; then
source ${HOMEGFS}/workflow/gw_setup.sh
fi

[TerrenceMcGuinness-NOAA]

@AnilKumar-NOAA Oh I see now that you were referring to the .bashrc in the roll account. In a Jenkins Pipeline the shell commands are not login shells but are issued from the shell environment from which the agent had been made to make the connection to the controller. Having said that, I will have to make some adjustments to the wrapper scripts so the needed modules can be set at runtime. Let me make some updates to the wrapper scripts in the CI framework and run some tests from a generic shell. In the meantime this PR has been reverted to a Draft

[emcbot]

Checkout Failed on Orion in Build# 6: Could not perform submodule update

[emcbot]

CI Passed on Orion in Build# 1
Built and ran in directory /work2/noaa/stmp/CI/ORION/3171

Experiment C48_ATM_8bafe297 Completed 2 Cycles: *SUCCESS* at Thu Dec 26 11:50:34 AM CST 2024
Experiment C96_S2SWA_gefs_replay_ics_8bafe297 Completed 1 Cycles: *SUCCESS* at Thu Dec 26 12:10:03 PM CST 2024
Experiment C96_atm3DVar_8bafe297 Completed 3 Cycles: *SUCCESS* at Thu Dec 26 01:16:04 PM CST 2024
Experiment C96C48_hybatmDA_8bafe297 Completed 3 Cycles: *SUCCESS* at Thu Dec 26 01:28:23 PM CST 2024
Experiment C48_S2SW_8bafe297 Completed 2 Cycles: *SUCCESS* at Thu Dec 26 02:25:26 PM CST 2024
Experiment C48_S2SWA_gefs_8bafe297 Completed 1 Cycles: *SUCCESS* at Thu Dec 26 02:51:50 PM CST 2024

[aerorahul]

changes to source gw_setup.sh are good.
A comment at the end of this PR re. GH might solve the shellnorm error. Please check.
Approve.

[TerrenceMcGuinness-NOAA]

committing update to launch script so exit works correctly when gh is not found

[aerorahul]

looks good.

@TerrenceMcGuinness-NOAA
Just a note here that "launch" is misspelled as "lanuch". Consequently, variables such as JENKINS_AGENT_LAUNCH_DIR is misspelled as JENKINS_AGENT_LANUCH_DIR. Not sure where that variable is defined.

This should probably be corrected in a future PR.

[aerorahul]

@TerrenceMcGuinness-NOAA
I just merged this PR and got a message about a Hercules jenkins-node.env fail. I hope this PR does not need additional work to be taken on the HPCs