Closed shorowit closed 4 months ago
I think this should be a top priority @wenyikuang and @DavidGoldwasser. It is another thing that is good to do now, before we get into another high pressure release cycle. It seems like we should be able to move this over to Jenkins without too much issue, right?
We could start by stopping running the workflow for all pushes to develop right now, until we figure out a better plan of action.
And yes, this is movable to Jenkins. With a couple of notes:
Hi @jmarrec @kbenne
I believe we could re-used the script we have here with some debugging and configuration update in the Jenkins system to build the CSharp binding release for each time we merge PR to the CSharp
branch. After that we may could deprecate the github action.
Or that's not the same case?
I notice it(the os_full_build/csharpbinding) is consuming Mingbo's Openstudio and using the flag of DBUILD_NUGET_PACKAGE
instead of DBUILD_CSHARP_BINDINGS
.
On a related note, GitHub Linux runners now have 4 cpus instead of 2: https://github.blog/2024-01-17-github-hosted-runners-double-the-power-for-open-source/
Should we close this issue? Seems like it has been addressed.
We do now have openstudio-gems running as an action, which is new, but I don't think it has been a big issue.
The other thing is that I believe the C# bindings/nuget package are setup as an action, but I'm not sure. We'll need @wenyikuang to weight and confirm if we can close this for sure.
Yeah, C# binding and pypi package are still in github actions.
https://github.com/NREL/OpenStudio/blob/develop/.github/workflows/buildCSharp.yml
I think I could migrant them to come with windows build if there is no hidden issues.
Enhancement Request
Currently OpenStudio uses 11 runners, totaling about 24 hours worth of runtime, to generate python and C# bindings on the CI. I believe that NREL gets a total of 20 runners for the entire organization. During active OpenStudio development, this has led to severe blockage that causes other repos' actions to queue up for extended periods of time, and can negatively affect other projects getting their work done.
Can we move these over to the Jenkings cluster? Or otherwise reduce/offload them?
@kbenne @wenyikuang