Get SOCA vrfy job working on Hera again#1045
Conversation
|
unit tests fail because EMC RZDM is down... |
guillaumevernieres
left a comment
There was a problem hiding this comment.
Thanks @CoryMartin-NOAA
I am unable to build GDASApp from PR #1033. I no longer think PR #1033 is the issue. I'm beginning to suspect the failure is related to EMC RZDM being offline. Attempts to build GDASApp |
|
I'm not sure what to do to fix this logjam. We have no idea when RZDM will be back. Do we disable all unit tests in the interim? Find an alternate place to store the files? Other options? |
|
Automated Global-Workflow GDASApp Testing Results: |
|
Automated Global-Workflow GDASApp Testing Results: |
should we put the tarball on hpc instead? |
|
Why does our not being able to download test files from EMC RZDM break the GDASApp build? It should only break testing, not the build, right? Does cmake intertwine application testing with the application build in such a way that if we can't build the tests, we can't build GDASApp? |
|
Do we have tarball |
It's because CMake is doing the downloading. It fails, so CMake can't finish. Perhaps we should instead have a test that runs that downloads/links the files. This is how JCSDA does it for JEDI. |
|
Placing the tarball on HPC is an option but when the given machine is offline, we're stuck. We could place the tarball on multiple HPC platforms. Even this approach has potential pitfalls. For example, we place the tarball in |
This is a good option. It separates the application build from the application tests. We should be able to compile GDASApp whether or not data for ctests is available. |
|
Automated Global-Workflow GDASApp Testing Results: |
|
Automated Global-Workflow GDASApp Testing Results: |
|
Automated Global-Workflow GDASApp Testing Results: Build: SUCCESS
|
|
Automated Global-Workflow GDASApp Testing Results: Build: SUCCESS
|
|
Automated Global-Workflow GDASApp Testing Results: |
|
Automated Global-Workflow GDASApp Testing Results: |
guillaumevernieres
left a comment
There was a problem hiding this comment.
The test_gdasapp_util_ghrsst2ioda job is failing too but isn't reported as failed. Can you update the references @apchoiCMD ?
The vrfy task is failing with an issue reading the layer thicknesses, not sure why ...
This PR fixes the previous error, good enough to be merged. We'll comment out the verify test until one of us water people have time to look into it and fix the issue.
My bad but your PR #1050 already includes a modified test reference file https://github.com/NOAA-EMC/GDASApp/pull/1050/files#diff-68603b0771f7acb935fa0d121599ea74718ba8012d0e7dc7c1ee1a192150d93e Do you want me to update before merging your PR? Thanks @guillaumevernieres |
I fixed it!!!!
* upstream/develop: Get SOCA vrfy job working on Hera again (#1045) Get test data from a staged location on supported HPC Part Deux (#1052) Prepare observations for snow DA updates to the ensemble members (#998) Save basic stats in csv at each cycle (#1040) Fix bug for datetime in GHRSST Ioda Converter (#1027)
What the title says. Fixes #994 --------- Co-authored-by: Guillaume Vernieres <guillaume.vernieres@noaa.gov>
What the title says. Fixes #994