From fa90a9dbe0d96bc0b0ca632b477227d39030c544 Mon Sep 17 00:00:00 2001
From: Giambattista Bloisi <giambattista.bloisi@openaire.eu>
Date: Mon, 21 Oct 2024 09:36:55 +0200
Subject: [PATCH] Update spark operator version

---
 docker-images/spark-operator/Dockerfile |   7 ++
 modules/airflow/airflow.tf              | 126 +++++++++++-------------
 modules/airflow/providers.tf            |  17 ++--
 modules/minio/providers.tf              |  17 ++--
 providers.tf                            |  12 +++
 spark-image/Dockerfile                  |   5 +-
 6 files changed, 96 insertions(+), 88 deletions(-)
 create mode 100644 docker-images/spark-operator/Dockerfile
 create mode 100644 providers.tf

diff --git a/docker-images/spark-operator/Dockerfile b/docker-images/spark-operator/Dockerfile
new file mode 100644
index 0000000..4e58fa3
--- /dev/null
+++ b/docker-images/spark-operator/Dockerfile
@@ -0,0 +1,7 @@
+# docker build -t spark-operator:2.0.2 . && kind load docker-image -n dnet-data-platform spark-operator:2.0.2
+FROM kubeflow/spark-operator:2.0.2
+
+ENV SPARK_HOME /opt/spark
+USER root
+RUN curl https://repo1.maven.org/maven2/org/apache/hadoop/hadoop-aws/3.3.4/hadoop-aws-3.3.4.jar -o ${SPARK_HOME}/jars/hadoop-aws-3.3.4.jar
+RUN curl https://repo1.maven.org/maven2/com/amazonaws/aws-java-sdk-bundle/1.12.262/aws-java-sdk-bundle-1.12.262.jar -o ${SPARK_HOME}/jars/aws-java-sdk-bundle-1.12.262.jar
diff --git a/modules/airflow/airflow.tf b/modules/airflow/airflow.tf
index b5000d4..6b0c267 100644
--- a/modules/airflow/airflow.tf
+++ b/modules/airflow/airflow.tf
@@ -22,15 +22,17 @@ resource "kubernetes_role" "airflow_spark_role" {
 
   rule {
     api_groups = ["sparkoperator.k8s.io"]
-    resources  = ["sparkapplications", "sparkapplications/status",
-        "scheduledsparkapplications", "scheduledsparkapplications/status"]
-    verbs      = ["*"]
+    resources = [
+      "sparkapplications", "sparkapplications/status",
+      "scheduledsparkapplications", "scheduledsparkapplications/status"
+    ]
+    verbs = ["*"]
   }
 
   rule {
     api_groups = [""]
-    resources  = ["pods/log"]
-    verbs      = ["*"]
+    resources = ["pods", "pods/log"]
+    verbs = ["*"]
   }
 }
 
@@ -55,49 +57,27 @@ resource "kubernetes_role_binding_v1" "airflow_spark_role_binding" {
 }
 
 resource "kubernetes_role_binding_v1" "airflow_spark_role_binding2" {
- depends_on = [kubernetes_namespace.spark_jobs_namespace]
- metadata {
-   name      = "airflow-spark-role-binding2"
-   namespace = "${var.namespace_prefix}spark-jobs"
- }
+  depends_on = [kubernetes_namespace.spark_jobs_namespace]
+  metadata {
+    name      = "airflow-spark-role-binding2"
+    namespace = "${var.namespace_prefix}spark-jobs"
+  }
 
- subject {
-   kind      = "ServiceAccount"
-   name      = "airflow-worker"
-   namespace = "${var.namespace_prefix}airflow"
- }
+  subject {
+    kind      = "ServiceAccount"
+    name      = "airflow-worker"
+    namespace = "${var.namespace_prefix}airflow"
+  }
 
- role_ref {
-   api_group = "rbac.authorization.k8s.io"
-   kind      = "Role"
-   name      = "spark-role"
- }
+  role_ref {
+    api_group = "rbac.authorization.k8s.io"
+    kind      = "Role"
+    name      = "spark-role"
+  }
 }
-#
-#
-# resource "kubernetes_role_binding_v1" "spark_role_binding" {
-#   depends_on = [kubernetes_namespace.spark_jobs_namespace]
-#   metadata {
-#     name      = "spark-role-binding"
-#     namespace = "${var.namespace_prefix}spark-jobs"
-#   }
-#
-#   subject {
-#     kind      = "ServiceAccount"
-#     name      = "spark"
-#     namespace = "${var.namespace_prefix}spark-jobs"
-#   }
-#
-#   role_ref {
-#     api_group = "rbac.authorization.k8s.io"
-#     kind      = "Role"
-#     name      = "spark-role"
-#   }
-# }
-#
 
 resource "helm_release" "gcp_spark_operator" {
-  depends_on        = [kubernetes_namespace.spark_jobs_namespace]
+  depends_on = [kubernetes_namespace.spark_jobs_namespace]
   name              = "gcp-spark-operator"
   chart             = "spark-operator"
   repository        = "https://kubeflow.github.io/spark-operator"
@@ -106,23 +86,38 @@ resource "helm_release" "gcp_spark_operator" {
   dependency_update = "true"
   version           = "2.0.2"
 
-  # set {
-  #   name  = "image.repository"
-  #   value = "kubeflow/spark-operator"
-  # }
-
-  # set {
-  #   name  = "image.tag"
-  #   value = "v1beta2-1.4.5-3.5.0"
-  # }
+  set {
+    name  = "image.repository"
+    value = "spark-operator"
+  }
 
   set {
-    name  = "sparkJobNamespaces"
+    name  = "image.tag"
+    value = "2.0.2"
+  }
+
+  set {
+    name  = "spark.jobNamespaces"
     value = "{${var.namespace_prefix}spark-jobs}"
   }
 
   set {
-    name  = "serviceAccounts.spark.name"
+    name  = "spark.serviceAccount.create"
+    value = "true"
+  }
+
+  set {
+    name  = "spark.serviceAccount.name"
+    value = "spark"
+  }
+
+  set {
+    name  = "controller.serviceAccount.create"
+    value = "true"
+  }
+
+  set {
+    name  = "controller.serviceAccount.name"
     value = "spark"
   }
 
@@ -147,13 +142,13 @@ resource "kubernetes_namespace" "airflow" {
 resource "kubernetes_secret" "s3_conn_secrets" {
   depends_on = [kubernetes_namespace.airflow]
   metadata {
-    name = "s3-conn-secrets"
+    name      = "s3-conn-secrets"
     namespace = "${var.namespace_prefix}airflow"
   }
 
   data = {
-    username = var.s3_key
-    password = var.s3_secret
+    username             = var.s3_key
+    password             = var.s3_secret
     AIRFLOW_CONN_S3_CONN = <<EOT
 {
   "conn_type": "aws",
@@ -171,7 +166,6 @@ EOT
 }
 
 
-
 resource "helm_release" "airflow" {
   depends_on = [kubernetes_secret.s3_conn_secrets]
 
@@ -197,7 +191,7 @@ resource "helm_release" "airflow" {
   }
 
   set {
-    name  = "spec.values.env"
+    name = "spec.values.env"
     value = yamlencode([
       {
         name  = "AIRFLOW__WEBSERVER__BASE_URL",
@@ -211,17 +205,17 @@ resource "helm_release" "airflow" {
   }
 
   set {
-    name ="dags.gitSync.repo"
+    name  = "dags.gitSync.repo"
     value = var.repo_url
   }
 
   set {
-    name ="dags.gitSync.branch"
+    name  = "dags.gitSync.branch"
     value = var.branch_name
   }
 
   set {
-    name ="dags.gitSync.subPath"
+    name  = "dags.gitSync.subPath"
     value = var.dag_path
   }
 
@@ -230,10 +224,10 @@ resource "helm_release" "airflow" {
   #   value = "gbloisi/airflow"
   # }
 
-  # set {
-  #   name  = "images.airflow.tag"
-  #   value = "2.8.3rc1-python3.11"
-  # }
+  set {
+    name  = "images.airflow.tag"
+    value = "2.9.3-python3.11"
+  }
 
   set {
     name  = "ingress.web.host"
diff --git a/modules/airflow/providers.tf b/modules/airflow/providers.tf
index f04b17c..d88c184 100644
--- a/modules/airflow/providers.tf
+++ b/modules/airflow/providers.tf
@@ -1,12 +1,9 @@
-provider "helm" {
-  # Several Kubernetes authentication methods are possible: https://registry.terraform.io/providers/hashicorp/kubernetes/latest/docs#authentication
-  kubernetes {
-    config_path = pathexpand(var.kube_config)
-    config_context = var.kube_context
+terraform {
+  required_providers {
+    helm = {
+    }
+
+    kubernetes = {
+    }
   }
 }
-
-provider "kubernetes" {
-  config_path = pathexpand(var.kube_config)
-  config_context = var.kube_context
-}
\ No newline at end of file
diff --git a/modules/minio/providers.tf b/modules/minio/providers.tf
index f04b17c..d88c184 100644
--- a/modules/minio/providers.tf
+++ b/modules/minio/providers.tf
@@ -1,12 +1,9 @@
-provider "helm" {
-  # Several Kubernetes authentication methods are possible: https://registry.terraform.io/providers/hashicorp/kubernetes/latest/docs#authentication
-  kubernetes {
-    config_path = pathexpand(var.kube_config)
-    config_context = var.kube_context
+terraform {
+  required_providers {
+    helm = {
+    }
+
+    kubernetes = {
+    }
   }
 }
-
-provider "kubernetes" {
-  config_path = pathexpand(var.kube_config)
-  config_context = var.kube_context
-}
\ No newline at end of file
diff --git a/providers.tf b/providers.tf
new file mode 100644
index 0000000..f04b17c
--- /dev/null
+++ b/providers.tf
@@ -0,0 +1,12 @@
+provider "helm" {
+  # Several Kubernetes authentication methods are possible: https://registry.terraform.io/providers/hashicorp/kubernetes/latest/docs#authentication
+  kubernetes {
+    config_path = pathexpand(var.kube_config)
+    config_context = var.kube_context
+  }
+}
+
+provider "kubernetes" {
+  config_path = pathexpand(var.kube_config)
+  config_context = var.kube_context
+}
\ No newline at end of file
diff --git a/spark-image/Dockerfile b/spark-image/Dockerfile
index 0f8dd25..a876682 100644
--- a/spark-image/Dockerfile
+++ b/spark-image/Dockerfile
@@ -1,6 +1,7 @@
-FROM spark:3.5.1-scala2.12-java17-ubuntu
+# docker build -t dnet-spark:1.0.0 . && kind load docker-image -n dnet-data-platform dnet-spark:1.0.0
+FROM spark:3.5.3-scala2.12-java17-ubuntu
 
-user root
+USER root
 RUN curl https://repo1.maven.org/maven2/org/apache/hadoop/hadoop-aws/3.3.4/hadoop-aws-3.3.4.jar -o ${SPARK_HOME}/jars/hadoop-aws-3.3.4.jar
 RUN curl https://repo1.maven.org/maven2/com/amazonaws/aws-java-sdk-bundle/1.12.262/aws-java-sdk-bundle-1.12.262.jar -o ${SPARK_HOME}/jars/aws-java-sdk-bundle-1.12.262.jar