Don't default roleGroup replicas to zero when not specified

[sbernauer]

Affected version

0.0.0-dev

Current and expected behavior

In

hdfs-operator/rust/crd/src/lib.rs

Lines 373 to 392 in f55c8ff

	/// Return replicas for a certain rolegroup.
	pub fn role_group_replicas(&self, hdfs: &HdfsCluster, role_group: &str) -> i32 {
	match self {
	HdfsRole::NameNode => hdfs
	.namenode_rolegroup(role_group)
	.and_then(|rg| rg.replicas)
	.unwrap_or_default()
	.into(),
	HdfsRole::DataNode => hdfs
	.datanode_rolegroup(role_group)
	.and_then(|rg| rg.replicas)
	.unwrap_or_default()
	.into(),
	HdfsRole::JournalNode => hdfs
	.journalnode_rolegroup(role_group)
	.and_then(|rg| rg.replicas)
	.unwrap_or_default()
	.into(),
	}
	}

we set the sts replicas to 0, when the user does not configure a replicas number on a rolegroup.
Instead we should change the fn to pub fn role_group_replicas(&self, hdfs: &HdfsCluster, role_group: &str) -> Option<i32> and pass the Option<i32> directly into the sts.
This way when the user does not specify replicas, the sts will have not set set replicas and k8s will default to 1 pod.
Users can than use HPA to scale their Pods based on e.g. resource usage.
This is how other operators e.g. trino are doing it (the right way)

Possible solution

No response

Additional context

No response

Environment

No response

Would you like to work on fixing this bug?

None