refactor!: Change CRD for rack awareness according to decision (#495)

* reafctor!: Change CRD for rack awareness according to decision

* changelog

* charts

* Update rust/crd/src/lib.rs

Co-authored-by: Nick <NickLarsenNZ@users.noreply.github.com>

* fix: whitespace

---------

Co-authored-by: Nick <NickLarsenNZ@users.noreply.github.com>
Co-authored-by: Nick Larsen <nick.larsen@stackable.tech>

Author: sbernauer

CHANGELOG.md

@@ -6,7 +6,7 @@ All notable changes to this project will be documented in this file.
 
 ### Added
 
-- Added rack awareness support via topology provider implementation ([#429]).
+- Added rack awareness support via topology provider implementation ([#429], [#495]).
 - More CRD documentation ([#433]).
 - Support for exposing HDFS clusters to clients outside of Kubernetes ([#450]).
 - Helm: support labels in values.yaml ([#460]).
@@ -42,6 +42,7 @@ All notable changes to this project will be documented in this file.
 [#475]: https://github.com/stackabletech/hdfs-operator/pull/475
 [#491]: https://github.com/stackabletech/hdfs-operator/pull/491
 [#492]: https://github.com/stackabletech/hdfs-operator/pull/492
+[#495]: https://github.com/stackabletech/hdfs-operator/pull/495
 
 ## [23.11.0] - 2023-11-24
 

deploy/helm/hdfs-operator/crds/crds.yaml

@@ -86,19 +86,18 @@ spec:
                     rackAwareness:
                       description: Configuration to control HDFS topology (rack) awareness feature
                       items:
+                        oneOf:
+                          - required:
+                              - nodeLabel
+                          - required:
+                              - podLabel
                         properties:
-                          labelName:
-                            description: Name of the label that will be used to resolve a datanode to a topology zone.
+                          nodeLabel:
+                            description: Name of the label on the Kubernetes Node (where the Pod is placed on) used to resolve a datanode to a topology zone.
                             type: string
-                          labelType:
-                            description: Name of the label type that will be typically either `node` or `pod`, used to create a topology out of datanodes.
-                            enum:
-                              - node
-                              - pod
+                          podLabel:
+                            description: Name of the label on the Kubernetes Pod used to resolve a datanode to a topology zone.
                             type: string
-                        required:
-                          - labelName
-                          - labelType
                         type: object
                       nullable: true
                       type: array

docs/modules/hdfs/pages/usage-guide/operations/rack-awareness.adoc

@@ -23,10 +23,8 @@ metadata:
 spec:
   clusterConfig:
     rackAwareness:
-      - labelType: node
-        labelName: topology.kubernetes.io/zone
-      - labelType: pod
-        labelName: app.kubernetes.io/role-group
+      - nodeLabel: topology.kubernetes.io/zone
+      - podLabel: app.kubernetes.io/role-group
   nameNodes:
     ...
 ----

rust/crd/src/lib.rs

@@ -184,33 +184,26 @@ pub struct HdfsClusterConfig {
     pub rack_awareness: Option<Vec<TopologyLabel>>,
 }
 
-#[derive(Clone, Debug, Default, Deserialize, Eq, Hash, JsonSchema, PartialEq, Serialize)]
+#[derive(Clone, Debug, Deserialize, Eq, Hash, JsonSchema, PartialEq, Serialize)]
 #[serde(rename_all = "camelCase")]
-pub struct TopologyLabel {
-    /// Name of the label type that will be typically either `node` or `pod`, used to create a
-    /// topology out of datanodes.
-    label_type: TopologyLabelType,
+pub enum TopologyLabel {
+    /// Name of the label on the Kubernetes Node (where the Pod is placed on) used to resolve a datanode to a topology
+    /// zone.
+    NodeLabel(String),
 
-    /// Name of the label that will be used to resolve a datanode to a topology zone.
-    label_name: String,
+    /// Name of the label on the Kubernetes Pod used to resolve a datanode to a topology zone.
+    PodLabel(String),
 }
 
 impl TopologyLabel {
     pub fn to_config(&self) -> String {
-        format!("{}:{}", self.label_type, self.label_name)
+        match &self {
+            TopologyLabel::NodeLabel(l) => format!("Node:{l}"),
+            TopologyLabel::PodLabel(l) => format!("Pod:{l}"),
+        }
     }
 }
 
-#[derive(
-    Clone, Debug, Default, Display, Deserialize, Eq, Hash, JsonSchema, PartialEq, Serialize,
-)]
-#[serde(rename_all = "lowercase")]
-pub enum TopologyLabelType {
-    #[default]
-    Node,
-    Pod,
-}
-
 fn default_dfs_replication_factor() -> u8 {
     DEFAULT_DFS_REPLICATION_FACTOR
 }
@@ -833,8 +826,7 @@ impl HdfsCluster {
         self.spec
             .cluster_config
             .rack_awareness
-            .clone()
-            .filter(|label_list| !label_list.is_empty())
+            .as_ref()
             .map(|label_list| {
                 label_list
                     .iter()
@@ -1483,8 +1475,7 @@ spec:
   clusterConfig:
     zookeeperConfigMapName: hdfs-zk
     rackAwareness:
-      - labelType: node
-        labelName: kubernetes.io/zone
+      - nodeLabel: kubernetes.io/zone
   nameNodes:
     roleGroups:
       default:
@@ -1512,7 +1503,9 @@ spec:
       default:
         replicas: 1";
 
-        let hdfs: HdfsCluster = serde_yaml::from_str(cr).unwrap();
+        let deserializer = serde_yaml::Deserializer::from_str(cr);
+        let hdfs: HdfsCluster =
+            serde_yaml::with::singleton_map_recursive::deserialize(deserializer).unwrap();
         let role = HdfsRole::DataNode;
         let config = &role.merged_config(&hdfs, "default").unwrap();
         let resources = &config.as_datanode().unwrap().resources;
@@ -1686,10 +1679,8 @@ spec:
   clusterConfig:
     zookeeperConfigMapName: hdfs-zk
     rackAwareness:
-      - labelType: node
-        labelName: kubernetes.io/zone
-      - labelType: pod
-        labelName: app.kubernetes.io/role-group
+      - nodeLabel: kubernetes.io/zone
+      - podLabel: app.kubernetes.io/role-group
   nameNodes:
     roleGroups:
       default:
@@ -1703,7 +1694,10 @@ spec:
       default:
         replicas: 1";
 
-        let hdfs: HdfsCluster = serde_yaml::from_str(cr).unwrap();
+        let deserializer = serde_yaml::Deserializer::from_str(cr);
+        let hdfs: HdfsCluster =
+            serde_yaml::with::singleton_map_recursive::deserialize(deserializer).unwrap();
+
         let rack_awareness = hdfs.rackawareness_config();
         // test the expected value to be used as an env-var: the mapper will use this to
         // convert to an HDFS-internal label

tests/templates/kuttl/topology-provider/11-install-hdfs.yaml.j2

@@ -16,10 +16,8 @@ commands:
         clusterConfig:
           zookeeperConfigMapName: hdfs-zk
           rackAwareness:
-            - labelType: node
-              labelName: kubernetes.io/hostname
-            - labelType: pod
-              labelName: app.kubernetes.io/role-group
+            - nodeLabel: kubernetes.io/hostname
+            - podLabel: app.kubernetes.io/role-group
           dfsReplication: 1
           authentication:
             tlsSecretClass: tls