Update gdas.cd #2978
Update gdas.cd #2978
Conversation
|
@guillaumevernieres , the hash for |
Co-authored-by: Rahul Mahajan <[email protected]>
guillaumevernieres
dismissed stale reviews from aerorahul and CoryMartin-NOAA
via
9538ba5
Thanks @RussTreadon-NOAA , working on it this morning. |
aerorahul
added
CI-Hera-Ready
emcbot
added
CI-Orion-Building
|
interestingly I found this Unidata/netcdf-c#2158 |
No idea @RussTreadon-NOAA for the WCOSS2 problem. I'm finishing up testing a subset of the |
|
Is it complaining about ending with an integer? shouldn't the string be irrelevant? |
|
Why is it referencing SOCA? eg: Is SOCA building FMS separately? Is it linking to two FMS versions somehow? Not sure if this is related: |
|
@RussTreadon-NOAA Can you try again with the latest feature/fms2io-fix . I have a hunch about why that illegal character error is being raised on Cactus |
|
Thank you @DavidNew-NOAA for the update. I can't directly pull your fv3-jedi branch, feature/fms2io-fix, into |
|
@DavidNew-NOAA , unfortunately, jobs gdas_aeroanlgenb and gdas_snowanl fail on Cactus in the same way as before. Note that the traceback for both failures reference libsoca. Why? |
|
g-w CI tests on Hercules All jobs successfully complete in the four tested configurations. This Passed result is the same as Hera. |
|
@DavidNew-NOAA will be on leave until Tuesday. Should we accept that this may fail on WCOSS, and allow this to proceed, but prioritize fixing this next week? Or wait until we have a solution for WCOSS? It's entirely possible it is a WCOSS library issue, I wonder if we can reproduce this with simplified code? |
|
I agree, @CoryMartin-NOAA. Let's move this PR forward as is and fix the WCOSS2 failures as soon as possible. All four g-w CI configurations tested on Hera and Hercules pass. 2 out of the 3 g-w CI tested on Cactus pass. What do you think @WalterKolczynski-NOAA and @aerorahul ? It would be much easier to troubleshoot what's going on if we could replicate the problem in a simple script. I will try to encapsulate the failing portions of the gdas_aeroanlgenb or gdas_snowanl jobs in a simple wrapper script. |
CI still needs to pass on WCOSS to be merged, so if we approve this plan:
Will let @aerorahul chime in first. |
|
Oops, I realize there's a bit more work required for this PR to move forward. All my tests are using GDASApp |
|
Sorry what @WalterKolczynski-NOAA said is what I intended to propose. We turn off the test and fix it and then turn the test back on at a later date. @aerorahul and I discussed yesterday that this would be acceptable in this sort of scenario |
RussTreadon-NOAA
dismissed stale reviews from danholdaway and aerorahul
via
81dcd72
|
Completion of this PR requires updates to files in EIB's |
|
Thank you @CoryMartin-NOAA for catching this. As a test I updated the soca |
|
g-w CI on Hera, Hercules, Orion, and WCOSS2 Install
with results as follows on the indicated machines Hera Hercules Orion WCOSS2 (Cactus) The Cactus failures are The following GDASApp issues have been opened to address these failures
g-w CI C96C48_hybatmaerosnowDA and C48mx500_3DVarAOWCDA are disabled on WCOSS2 |
|
Thank you @aerorahul for outlining what needs to be done before this PR can be closed. I converted your list into tasks and added the fix update.
|
|
@WalterKolczynski-NOAA , this PR is ready for final review. Once g-w issue #3012 is closed, we need to update |
Description
Updates to the
gdas.cd#.@RussTreadon-NOAA will submit a PR in the GDASApp, we'll update the gdas.cd # in this branch after the GDASApp PR is merged.
In the mean time, could somebody review the few simple code changes that are needed to run with the new #?
Type of change
Change characteristics
How has this been tested?
Run subset of g-w CI on Hera, Hercules, Orion, and WCOSS2 (Cactus)
Checklist