Description of problem:

after upgrading to 2.9 scaling machinepools no longer has any effect on managed clusters

Version-Release number of selected component (if applicable):

ACM 2.9.1

MCE 2.4.2

How reproducible:

customer environment

Steps to Reproduce:

1. deploy 2.8
2. deploy clusters with ACM against openstack
3. set up manual machinepool scaling
4. upgrade to 2.9
5. try scaling a machinepool again

Actual results:

the machinepool does not scale, hive-controller manager logs include errors such as

~~~

2023-12-21T08:23:50.126349775Z time="2023-12-21T08:23:50.126Z" level=info msg="found 1 remote machine sets" controller=machinepool machinePool=52c07a135d/52c07a135d-worker reconcileID=wpggjwz7
2023-12-21T08:23:50.126405909Z time="2023-12-21T08:23:50.126Z" level=warning msg="cannot decode OpenstackProviderSpec from master machine" controller=machinepool error="could not decode OpenStack ProviderSpec: no kind \"OpenstackProviderSpec\" is registered for version \"openstackproviderconfig.openshift.io/v1alpha1\" in scheme \"pkg/runtime/scheme.go:100\"" machinePool=toto/tata-worker reconcileID=titi
2023-12-21T08:23:50.126426895Z time="2023-12-21T08:23:50.126Z" level=error msg="error getting os image from master machine" controller=machinepool error="cannot decode OpenstackProviderSpec from master machine: could not decode OpenStack ProviderSpec: no kind \"OpenstackProviderSpec\" is registered for version \"openstackproviderconfig.openshift.io/v1alpha1\" in scheme \"pkg/runtime/scheme.go:100\"" machinePool=toto/tata-worker reconcileID=titi
2023-12-21T08:23:50.126441123Z time="2023-12-21T08:23:50.126Z" level=error msg="unable to create actuator" controller=machinepool error="cannot decode OpenstackProviderSpec from master machine: could not decode OpenStack ProviderSpec: n
o kind \"OpenstackProviderSpec\" is registered for version \"openstackproviderconfig.openshift.io/v1alpha1\" in scheme \"pkg/runtime/scheme.go:100\"" machinePool=toto/tata-worker reconcileID=titi
2023-12-21T08:23:50.126461474Z time="2023-12-21T08:23:50.126Z" level=error msg="could not generateMachineSets" controller=machinepool error="cannot decode OpenstackProviderSpec from master machine: could not decode OpenStack ProviderSpec: no kind \"OpenstackProviderSpec\" is registered for version \"openstackproviderconfig.openshift.io/v1alpha1\" in scheme \"pkg/runtime/scheme.go:100\"" machinePool=toto/tata-worker reconcileID=titi
2023-12-21T08:23:50.126476883Z time="2023-12-21T08:23:50.126Z" level=info msg="reconcile complete" controller=machinepool elapsedMillis=57 elapsedMillisGT=0 machinePool=toto/tata-worker outcome=unspecified reconcileID=titi

~~~

Expected results:

machinepool scales as expected

Additional info:

the manipulation to scale up machinepools is to go through the ui to go to cluster > machinepool then use the "scale machine pool" option

updating the machineset on the managed cluster directly works but there is concern if this issue is fixed it'll break.