Skip to content

Commit

Permalink
[develop] Fix Jenkins Nightly Build (#1161)
Browse files Browse the repository at this point in the history
The Jenkins nightly builds have been inconsistent or not working at all on the parallel works (PW) platforms. Some of the issues have been related to the instance's infrastructure on Azure or a conda conflict between the host machine and the conda built by the SRW App (originally seeing on GCP is now being observed for all PW platforms). This PR resolves the conda conflict by deactivating the host conda before activating the srw_app environment for all PW platforms. The solution for Azure requires configurations changes, which were done on the backend.

---------

Co-authored-by: Parallel Works app-run user <Edward.Snyder@mgmt-edwardsnyder-awsrocky8test-00006.pw-noaa-us-east-1.pw.local>
Co-authored-by: michael.lueken <[email protected]>
  • Loading branch information
3 people authored Dec 2, 2024
1 parent dfb8fd3 commit 4fbdf7f
Show file tree
Hide file tree
Showing 5 changed files with 3 additions and 12 deletions.
2 changes: 1 addition & 1 deletion .cicd/scripts/srw_ftest.sh
Original file line number Diff line number Diff line change
Expand Up @@ -89,7 +89,7 @@ module use modulefiles
module load build_${platform,,}_${SRW_COMPILER}
module load wflow_${platform,,}
# Deactivate conflicting conda env on GCP
[[ "${SRW_PLATFORM}" =~ "gclusternoaa" ]] && conda deactivate
[[ "${platform}" == "noaacloud" ]] && conda deactivate

[[ ${FORGIVE_CONDA} == true ]] && set +e +u # Some platforms have incomplete python3 or conda support, but wouldn't necessarily block workflow tests
conda activate srw_app
Expand Down
2 changes: 1 addition & 1 deletion .cicd/scripts/srw_metric.sh
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,7 @@ cd ${workspace}
# Activate workflow environment
module load wflow_${platform,,}
# Deactivate conflicting conda env on GCP
[[ "${SRW_PLATFORM}" =~ "gclusternoaa" ]] && conda deactivate
[[ "${platform}" == "noaacloud" ]] && conda deactivate

[[ ${FORGIVE_CONDA} == true ]] && set +e +u # Some platforms have incomplete python3 or conda support, but would not necessarily block workflow tests
conda activate srw_app
Expand Down
1 change: 1 addition & 0 deletions doc/conf.py
Original file line number Diff line number Diff line change
Expand Up @@ -103,6 +103,7 @@
linkcheck_ignore = [r'https://www\.intel\.com/content/www/us/en/docs/cpp\-compiler/developer\-guide\-reference/2021\-10/thread\-affinity\-interface\.html',
r'https://www\.intel\.com/content/www/us/en/developer/tools/oneapi/hpc\-toolkit\-download\.html',
r'https://glossary.ametsoc.org/.*',
r'https://readthedocs.org/projects/ufs-srweather-app/',
]

# Ignore anchor tags for SRW App data bucket. Shows Not Found even when they exist.
Expand Down
5 changes: 0 additions & 5 deletions modulefiles/tasks/noaacloud/python_srw.lua
Original file line number Diff line number Diff line change
@@ -1,7 +1,2 @@
load("conda")
setenv("SRW_ENV", "srw_app")

-- Add missing libstdc binary for Azure
if os.getenv("PW_CSP") == "azure" then
setenv("LD_PRELOAD","/opt/nvidia/nsight-systems/2023.1.2/host-linux-x64/libstdc++.so.6")
end
5 changes: 0 additions & 5 deletions modulefiles/tasks/noaacloud/run_vx.local.lua
Original file line number Diff line number Diff line change
Expand Up @@ -25,8 +25,3 @@ end
load("ufs-pyenv")
load("conda")
setenv("SRW_ENV", "srw_app")

-- Add missing libstdc binary for Azure
if os.getenv("PW_CSP") == "azure" then
setenv("LD_PRELOAD","/opt/nvidia/nsight-systems/2023.1.2/host-linux-x64/libstdc++.so.6")
end

0 comments on commit 4fbdf7f

Please sign in to comment.