Problem: the current fluence build is broken.
Solution: the issue is two fold, and unfortunately depends on the docker version. For newer docker that requires --load, that argument would be needed in the build script, and we will need to add it (and then add the script to our assets) if a newer version is used in the CI. The other breaking change was changing the location of the kube- scheduler binary in the controller image. Both of these should now be fixed, and I will update the PR further pending results in CI.
Problem: the current fluence build is broken. Solution: the issue is two fold, and unfortunately depends on the docker version. For newer docker that requires --load, that argument would be needed in the build script, and we will need to add it (and then add the script to our assets) if a newer version is used in the CI. The other breaking change was changing the location of the kube- scheduler binary in the controller image. Both of these should now be fixed, and I will update the PR further pending results in CI.