Restore EMR Spark plugin startup scripts

scripts/csp-startup-scripts/README.md

@@ -0,0 +1,3 @@
+# Startup Scripts for CSPs with Spark Rapids
+
+With the exception of Dataproc, CSP offerings like EMR have specific set of steps that are required to enable the Spark Rapids Plugin in their environment. The set of scripts here automate parts of that process, for EMR currently. The exact usage can be found in our docs [here](https://docs.nvidia.com/spark-rapids/user-guide/latest/getting-started/aws-emr.html)

scripts/csp-startup-scripts/emr/cgroup-bootstrap-action-emr6.sh

@@ -0,0 +1,21 @@
+#!/bin/bash
+#
+# Copyright (c) 2024-2026, NVIDIA CORPORATION.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+set -ex
+
+sudo chmod a+rwx -R /sys/fs/cgroup/cpu,cpuacct
+sudo chmod a+rwx -R /sys/fs/cgroup/devices

scripts/csp-startup-scripts/emr/cgroup-bootstrap-action-emr7.sh

@@ -0,0 +1,21 @@
+#!/bin/bash
+#
+# Copyright (c) 2024-2026, NVIDIA CORPORATION.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+set -ex
+
+sudo mkdir -p /spark-rapids-cgroup/devices
+sudo mount -t cgroup -o devices cgroupv1-devices /spark-rapids-cgroup/devices
+sudo chmod a+rwx -R /spark-rapids-cgroup

scripts/csp-startup-scripts/emr/config-emr6.json

@@ -0,0 +1,66 @@
+[
+  {
+    "Classification":"spark",
+    "Properties":{
+      "enableSparkRapids":"true"
+    }
+  },
+  {
+    "Classification":"yarn-site",
+    "Properties":{
+      "yarn.nodemanager.resource-plugins":"yarn.io/gpu",
+      "yarn.resource-types":"yarn.io/gpu",
+      "yarn.nodemanager.resource-plugins.gpu.allowed-gpu-devices":"auto",
+      "yarn.nodemanager.resource-plugins.gpu.path-to-discovery-executables":"/usr/bin",
+      "yarn.nodemanager.linux-container-executor.cgroups.mount":"true",
+      "yarn.nodemanager.linux-container-executor.cgroups.mount-path":"/sys/fs/cgroup",
+      "yarn.nodemanager.linux-container-executor.cgroups.hierarchy":"yarn",
+      "yarn.nodemanager.container-executor.class":"org.apache.hadoop.yarn.server.nodemanager.LinuxContainerExecutor"
+    }
+  },
+  {
+    "Classification":"container-executor",
+    "Properties":{
+
+    },
+    "Configurations":[
+      {
+        "Classification":"gpu",
+        "Properties":{
+          "module.enabled":"true"
+        }
+      },
+      {
+        "Classification":"cgroups",
+        "Properties":{
+          "root":"/sys/fs/cgroup",
+          "yarn-hierarchy":"yarn"
+        }
+      }
+    ]
+  },
+  {
+    "Classification":"spark-defaults",
+    "Properties":{
+      "spark.plugins":"com.nvidia.spark.SQLPlugin",
+      "spark.executor.resource.gpu.discoveryScript":"/usr/lib/spark/scripts/gpu/getGpusResources.sh",
+      "spark.submit.pyFiles":"/usr/lib/spark/jars/xgboost4j-spark_3.0-1.4.2-0.3.0.jar",
+      "spark.executor.extraLibraryPath":"/usr/local/cuda/targets/x86_64-linux/lib:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/cuda/compat/lib:/usr/local/cuda/lib:/usr/local/cuda/lib64:/usr/lib/hadoop/lib/native:/usr/lib/hadoop-lzo/lib/native:/docker/usr/lib/hadoop/lib/native:/docker/usr/lib/hadoop-lzo/lib/native",
+      "spark.rapids.sql.concurrentGpuTasks":"2",
+      "spark.executor.resource.gpu.amount":"1",
+      "spark.executor.cores":"${executor_cores}",
+      "spark.task.cpus":"1",
+      "spark.task.resource.gpu.amount":"${task_gpu_amount}",
+      "spark.rapids.memory.pinnedPool.size":"2G",
+      "spark.executor.memoryOverhead":"2G",
+      "spark.sql.files.maxPartitionBytes":"256m",
+      "spark.sql.adaptive.enabled":"false"
+    }
+  },
+  {
+    "Classification":"capacity-scheduler",
+    "Properties":{
+      "yarn.scheduler.capacity.resource-calculator":"org.apache.hadoop.yarn.util.resource.DominantResourceCalculator"
+    }
+  }
+]

scripts/csp-startup-scripts/emr/config-emr7.json

@@ -0,0 +1,64 @@
+[
+  {
+    "Classification": "spark",
+    "Properties": {
+      "enableSparkRapids": "true"
+    }
+  },
+  {
+    "Classification": "yarn-site",
+    "Properties": {
+      "yarn.nodemanager.resource-plugins": "yarn.io/gpu",
+      "yarn.resource-types": "yarn.io/gpu",
+      "yarn.nodemanager.resource-plugins.gpu.allowed-gpu-devices": "auto",
+      "yarn.nodemanager.resource-plugins.gpu.path-to-discovery-executables": "/usr/bin",
+      "yarn.nodemanager.linux-container-executor.cgroups.mount": "true",
+      "yarn.nodemanager.linux-container-executor.cgroups.mount-path": "/spark-rapids-cgroup",
+      "yarn.nodemanager.linux-container-executor.cgroups.hierarchy": "yarn",
+      "yarn.nodemanager.container-executor.class": "org.apache.hadoop.yarn.server.nodemanager.LinuxContainerExecutor"
+    }
+  },
+  {
+    "Classification": "container-executor",
+    "Properties": {},
+    "Configurations": [
+      {
+        "Classification": "gpu",
+        "Properties": {
+          "module.enabled": "true"
+        }
+      },
+      {
+        "Classification": "cgroups",
+        "Properties": {
+          "root": "/spark-rapids-cgroup",
+          "yarn-hierarchy": "yarn"
+        }
+      }
+    ]
+  },
+  {
+    "Classification": "spark-defaults",
+    "Properties": {
+      "spark.plugins": "com.nvidia.spark.SQLPlugin",
+      "spark.executor.resource.gpu.discoveryScript": "/usr/lib/spark/scripts/gpu/getGpusResources.sh",
+      "spark.submit.pyFiles": "/usr/lib/spark/jars/xgboost4j-spark_3.0-1.4.2-0.3.0.jar",
+      "spark.executor.extraLibraryPath": "/usr/local/cuda/targets/x86_64-linux/lib:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/cuda/compat/lib:/usr/local/cuda/lib:/usr/local/cuda/lib64:/usr/lib/hadoop/lib/native:/usr/lib/hadoop-lzo/lib/native:/docker/usr/lib/hadoop/lib/native:/docker/usr/lib/hadoop-lzo/lib/native",
+      "spark.rapids.sql.concurrentGpuTasks": "2",
+      "spark.executor.resource.gpu.amount": "1",
+      "spark.executor.cores": "${executor_cores}",
+      "spark.task.cpus": "1",
+      "spark.task.resource.gpu.amount": "${task_gpu_amount}",
+      "spark.rapids.memory.pinnedPool.size": "2G",
+      "spark.executor.memoryOverhead": "2G",
+      "spark.sql.files.maxPartitionBytes": "256m",
+      "spark.sql.adaptive.enabled": "false"
+    }
+  },
+  {
+    "Classification": "capacity-scheduler",
+    "Properties": {
+      "yarn.scheduler.capacity.resource-calculator": "org.apache.hadoop.yarn.util.resource.DominantResourceCalculator"
+    }
+  }
+]

scripts/csp-startup-scripts/emr/emr-spark-plugin-startup.py

@@ -0,0 +1,124 @@
+#
+# Copyright (c) 2024-2026, NVIDIA CORPORATION.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+import argparse
+import subprocess
+import json
+import boto3
+from botocore.exceptions import NoCredentialsError, PartialCredentialsError
+
+def upload_file_to_s3(file_name, bucket_name, object_name=None):
+    s3 = boto3.client('s3')
+
+    # If no object name is specified, use the file name
+    if object_name is None:
+        object_name = file_name
+
+    try:
+        s3.upload_file(file_name, bucket_name, object_name)
+        print(f"File '{file_name}' uploaded successfully to bucket '{bucket_name}' as '{object_name}'")
+    except FileNotFoundError:
+        print(f"Error: The file {file_name} was not found.")
+    except NoCredentialsError:
+        print("Error: AWS credentials not found.")
+    except PartialCredentialsError:
+        print("Error: Incomplete AWS credentials.")
+    except Exception as e:
+        print(f"An error occurred: {e}")
+
+g4dn_instance_map = {
+    "g4dn.xlarge": 4,
+    "g4dn.2xlarge": 8,
+    "g4dn.4xlarge": 16,
+    "g4dn.12xlarge": 48,
+    "g4dn.16xlarge": 64
+}
+
+def create_emr_cluster(release_label, key_name, service_role, subnet_id, az, instance_profile, worker_instance, s3_bucket_name):
+    try:
+        conf_json_fn = None
+        bootstrap_fn = None
+        if "emr-7" in release_label:
+            conf_json_fn="config-emr7.json"
+            bootstrap_fn="cgroup-bootstrap-action-emr7.sh"
+        else:
+            conf_json_fn="config-emr6.json"
+            bootstrap_fn="cgroup-bootstrap-action-emr6.sh"
+        # Replace the fields in the json
+        exec_cores = g4dn_instance_map.get(worker_instance)
+
+        print("Config Json" + conf_json_fn)
+        with open(conf_json_fn, 'r') as file:
+            data = json.load(file)
+        json_string = json.dumps(data)
+
+        # Replace the placeholder with the actual variable
+        json_string = json_string.replace("${task_gpu_amount}", str(1/exec_cores))
+        json_string = json_string.replace("${executor_cores}", str(exec_cores))
+        updated_data = json.loads(json_string)
+
+        print(json.dumps(updated_data, indent=4))
+        with open(conf_json_fn, "w") as file:
+            file.write(json_string)
+        upload_file_to_s3(bootstrap_fn, s3_bucket_name, bootstrap_fn)
+
+
+        command = [
+            "aws", "emr", "create-cluster",
+            "--release-label", release_label,
+            "--applications", "Name=Hadoop", "Name=Spark", "Name=Livy", "Name=JupyterEnterpriseGateway",
+            "--service-role", service_role,
+            "--ec2-attributes", f"AvailabilityZone={az},InstanceProfile={instance_profile}",
+            "--instance-groups",
+            "InstanceGroupType=MASTER,InstanceCount=1,InstanceType=m4.4xlarge",
+            f"InstanceGroupType=CORE,InstanceCount=1,InstanceType={worker_instance}",
+            "--configurations", f"file://{conf_json_fn}",
+            "--bootstrap-actions", f"Name='Setup cgroups bootstrap',Path=s3://{s3_bucket_name}/{bootstrap_fn}"
+        ]
+
+        result = subprocess.run(command, check=True, text=True, capture_output=True)
+
+        print("Cluster created successfully!")
+        print(result.stdout)
+
+    except subprocess.CalledProcessError as e:
+        print("Error creating EMR cluster:", e.stderr)
+
+
+parser = argparse.ArgumentParser(description="A script that takes command-line arguments.")
+
+# Define arguments
+parser.add_argument("-r", "--release_label", type=str, default="emr-7.1.0",  help="EMR Release Label, emr-7.1.0 for example")
+parser.add_argument("-k", "--key_name", type=str, required=True, help="Access Key Name")
+parser.add_argument("-s", "--service_role", type=str, required=True, help="AWS EMR service Role")
+parser.add_argument("-n", "--subnet", type=str, required=True, help="Subnet ID")
+parser.add_argument("-z", "--availability_zone", type=str, default="us-west-2b", help="Availability Zone")
+parser.add_argument("-i", "--instance_profile", type=str, required=True, help="Instance Profile")
+parser.add_argument("-w", "--worker_instance", type=str, default="g4dn.2xlarge",  help="Worker Instance g4dn.xxxx")
+parser.add_argument("-b", "--s3_bucket_name", type=str, required=True, help="S3 Bucket Name to store the bootstrap and config info")
+
+args = parser.parse_args()
+
+release_label = args.release_label
+key_name = args.key_name
+service_role = args.service_role
+subnet_id = args.subnet
+az = args.availability_zone
+instance_profile = args.instance_profile
+worker_instance = args.worker_instance
+s3_bucket_name = args.s3_bucket_name
+
+create_emr_cluster(release_label, key_name, service_role, subnet_id, az, instance_profile, worker_instance, s3_bucket_name)