[Merged by Bors] - Fix PVC-related tests on managed k8s

CHANGELOG.md

@@ -9,8 +9,10 @@ All notable changes to this project will be documented in this file.
 ### Changed
 
 - BREAKING: Use current S3 connection/bucket structs ([#86])
+- Add node selector to top-level job and specify node selection in PVC-relevant tests ([#90])
 
 [#86]: https://github.com/stackabletech/spark-k8s-operator/pull/86
+[#90]: https://github.com/stackabletech/spark-k8s-operator/pull/90
 
 ## [0.2.0] - 2022-06-21
 

deploy/crd/sparkapplication.crd.yaml

@@ -272,6 +272,11 @@ spec:
                 mode:
                   nullable: true
                   type: string
+                nodeSelector:
+                  additionalProperties:
+                    type: string
+                  nullable: true
+                  type: object
                 s3bucket:
                   description: Operators are expected to define fields for this type in order to work with S3 buckets.
                   nullable: true

deploy/helm/spark-k8s-operator/crds/crds.yaml

@@ -274,6 +274,11 @@ spec:
                 mode:
                   nullable: true
                   type: string
+                nodeSelector:
+                  additionalProperties:
+                    type: string
+                  nullable: true
+                  type: object
                 s3bucket:
                   description: Operators are expected to define fields for this type in order to work with S3 buckets.
                   nullable: true

deploy/manifests/crds.yaml

@@ -275,6 +275,11 @@ spec:
                 mode:
                   nullable: true
                   type: string
+                nodeSelector:
+                  additionalProperties:
+                    type: string
+                  nullable: true
+                  type: object
                 s3bucket:
                   description: Operators are expected to define fields for this type in order to work with S3 buckets.
                   nullable: true

rust/crd/src/lib.rs

@@ -100,6 +100,8 @@ pub struct SparkApplicationSpec {
     pub volumes: Option<Vec<Volume>>,
     #[serde(default, skip_serializing_if = "Option::is_none")]
     pub env: Option<Vec<EnvVar>>,
+    #[serde(default, skip_serializing_if = "Option::is_none")]
+    pub node_selector: Option<std::collections::BTreeMap<String, String>>,
 }
 
 #[derive(Clone, Debug, Deserialize, Eq, JsonSchema, PartialEq, Serialize, Display, EnumString)]
@@ -411,6 +413,10 @@ impl SparkApplication {
             .as_ref()
             .and_then(|executor_config| executor_config.node_selector.clone())
     }
+
+    pub fn job_node_selector(&self) -> Option<std::collections::BTreeMap<String, String>> {
+        self.spec.node_selector.clone()
+    }
 }
 
 #[derive(Clone, Debug, Default, Deserialize, Eq, JsonSchema, PartialEq, Serialize)]

rust/operator-binary/src/spark_k8s_controller.rs

@@ -388,6 +388,7 @@ fn spark_job(
                 .fs_group(1000)
                 .build()
                 .into(), // Needed for secret-operator
+            node_selector: spark_application.job_node_selector(),
             ..PodSpec::default()
         }),
     };

tests/templates/kuttl/spark-ny-public-s3/02-deps-volume.yaml

@@ -17,6 +17,8 @@ metadata:
 spec:
   template:
     spec:
+      nodeSelector:
+        node: "1"
       restartPolicy: Never
       volumes:
         - name: job-deps

tests/templates/kuttl/spark-ny-public-s3/10-deploy-spark-app.yaml.j2

@@ -18,6 +18,8 @@ spec:
   mode: cluster
   mainClass: tech.stackable.demo.spark.NYTLCReport
   mainApplicationFile: s3a://my-bucket/ny-tlc-report-1.1.0.jar
+  nodeSelector:
+    node: "1"
   volumes:
     - name: spark-ny-deps
       persistentVolumeClaim:
@@ -48,6 +50,8 @@ spec:
         mountPath: /dependencies
       - name: cm-job-arguments
         mountPath: /arguments
+    nodeSelector:
+      node: "1"
   executor:
     cores: 1
     instances: 3
@@ -57,3 +61,5 @@ spec:
         mountPath: /dependencies
       - name: cm-job-arguments
         mountPath: /arguments
+    nodeSelector:
+      node: "1"

tests/templates/kuttl/spark-pi-private-s3/02-deps-volume.yaml

@@ -2,7 +2,7 @@
 apiVersion: v1
 kind: PersistentVolumeClaim
 metadata:
-  name: spark-pi-pvc
+  name: spark-pi-private-pvc
 spec:
   accessModes:
     - ReadWriteOnce
@@ -17,11 +17,13 @@ metadata:
 spec:
   template:
     spec:
+      nodeSelector:
+        node: "1"
       restartPolicy: Never
       volumes:
         - name: job-deps
           persistentVolumeClaim:
-            claimName: spark-pi-pvc
+            claimName: spark-pi-private-pvc
       containers:
         - name: aws-deps
           image: docker.stackable.tech/stackable/tools:0.2.0-stackable0

tests/templates/kuttl/spark-pi-private-s3/10-deploy-spark-app.yaml.j2

@@ -9,6 +9,8 @@ spec:
   mode: cluster
   mainClass: org.apache.spark.examples.SparkPi
   mainApplicationFile: s3a://my-bucket/spark-examples_2.12-{{ test_scenario['values']['spark'] }}.jar
+  nodeSelector:
+    node: "1"
   s3bucket:
     inline:
       bucketName: my-bucket
@@ -22,7 +24,7 @@ spec:
   volumes:
     - name: spark-pi-deps
       persistentVolumeClaim:
-        claimName: spark-pi-pvc
+        claimName: spark-pi-private-pvc
   sparkConf:
     spark.hadoop.fs.s3a.aws.credentials.provider: "org.apache.hadoop.fs.s3a.SimpleAWSCredentialsProvider"
     spark.driver.extraClassPath: "/dependencies/jars/hadoop-aws-3.2.0.jar:/dependencies/jars/aws-java-sdk-bundle-1.11.375.jar"
@@ -34,10 +36,14 @@ spec:
     volumeMounts:
       - name: spark-pi-deps
         mountPath: /dependencies
+    nodeSelector:
+      node: "1"
   executor:
     cores: 1
     instances: 1
     memory: "512m"
     volumeMounts:
       - name: spark-pi-deps
         mountPath: /dependencies
+    nodeSelector:
+      node: "1"

tests/templates/kuttl/spark-pi-public-s3/02-deps-volume.yaml

@@ -2,7 +2,7 @@
 apiVersion: v1
 kind: PersistentVolumeClaim
 metadata:
-  name: spark-pi-pvc
+  name: spark-pi-public-pvc
 spec:
   accessModes:
     - ReadWriteOnce
@@ -17,11 +17,13 @@ metadata:
 spec:
   template:
     spec:
+      nodeSelector:
+        node: "1"
       restartPolicy: Never
       volumes:
         - name: job-deps
           persistentVolumeClaim:
-            claimName: spark-pi-pvc
+            claimName: spark-pi-public-pvc
       containers:
         - name: aws-deps
           image: docker.stackable.tech/stackable/tools:0.2.0-stackable0

tests/templates/kuttl/spark-pi-public-s3/10-deploy-spark-app.yaml.j2

@@ -10,10 +10,12 @@ spec:
   mode: cluster
   mainClass: org.apache.spark.examples.SparkPi
   mainApplicationFile: s3a://my-bucket/spark-examples_2.12-{{ test_scenario['values']['spark'] }}.jar
+  nodeSelector:
+    node: "1"
   volumes:
     - name: spark-pi-deps
       persistentVolumeClaim:
-        claimName: spark-pi-pvc
+        claimName: spark-pi-public-pvc
   s3bucket:
     inline:
       bucketName: my-bucket
@@ -33,10 +35,14 @@ spec:
     volumeMounts:
       - name: spark-pi-deps
         mountPath: /dependencies
+    nodeSelector:
+      node: "1"
   executor:
     cores: 1
     instances: 1
     memory: "512m"
     volumeMounts:
       - name: spark-pi-deps
         mountPath: /dependencies
+    nodeSelector:
+      node: "1"