Reconcile loop called every second on OpenShift, but not on Kubernetes

[Juraci Paixão Kröhling]

Hello,

I'm currently experiencing a difference in behavior that I cannot
account for, and perhaps it's related to the Operator Framework: when
the my operator is running in Kubernetes, the reconciliation loop is
called only when the CR is changed. When it's running in OpenShift, the
reconciliation loops is called every second.

Which one is the correct behavior?

- Juca.

[Juraci Paixão Kröhling]

As an update: looks like this was happening because I store back the CR
at the end of the reconcile loop. For Kubernetes, it's a noop, as it
detects that the CR hasn't changed.

Doing a `reflect.DeepEqual(original, reconciled)` before issuing an
update did the trick.

- Juca.