chore: Do not reconcile CephCluster spec if managed by Helm

[rrpolanco]

When the Ceph Cluster is created via helm.InstallChartArchive() when rookMinimumNodeCount >= 3, no modifications should be made to the CephCluster CR. Namely for two reasons, 1) CephCluster spec in the cluster should not deviate from what was configured in the chart 2) modifying the CR delays the Rook reconciliation process

[laverya]

What is being modified here? Is it just labels? Is it more? Does this exclude 'what nodes should I put storage on' updates, and other things of that nature?

If it is more, we need to keep it. If it is just labels, I would prefer to fix the process to not change the labels.

If it's a true noop apply, I would love a way to skip all noops, and not exclude helm-created clusters from management entirely.

The goal of creating a cephcluster with helm via the operator was not to create two paths in perpetuity (script-created and helm-created cephclusters) it was instead to add an additional way to join the script path. This PR seems to split the paths in perpetuity, and will cause problems in the future.

[rrpolanco]

This is change is not needed. Closing.