Working with AWS Parallel Cluster - Revision history

Zac: /* AWS Parallel Cluster version 3 */

2022-02-23T21:06:59Z

‎AWS Parallel Cluster version 3

Zac: /* Modeling spot cluster config */

2022-02-23T20:58:45Z

‎Modeling spot cluster config

Zac: /* Modeling spot cluster config */

2022-02-23T20:00:23Z

‎Modeling spot cluster config

Zac: /* Modeling spot cluster config */

2022-02-23T19:57:58Z

‎Modeling spot cluster config

Zac: /* Modeling spot cluster config */

2022-02-23T19:37:57Z

‎Modeling spot cluster config

Zac: /* AWS Parallel Cluster version 3 */

2022-02-23T19:37:22Z

‎AWS Parallel Cluster version 3

Zac at 19:34, 23 February 2022

2022-02-23T19:34:39Z

Zac: /* AWS Parallel Cluster */

2022-02-23T19:18:09Z

‎AWS Parallel Cluster

Zac: /* AWS Parallel Cluster */

2021-03-26T17:28:28Z

‎AWS Parallel Cluster

Zac: /* spot pcluster config */

2021-03-26T17:27:45Z

‎spot pcluster config

@@ Line 125: / Line 125: @@
    - Key: Name
      Value: Feb22Pcluster
 = AWS Parallel Cluster version 2 =

@@ Line 21: / Line 21: @@
 ** m6g.16xlarge - General purpose instance with 64 vCPUs, 256 Gb RAM, AWS Graviton 2 processors
 ** c5.24xlarge - Compute optimized instance with 96 vCPUs, and 192 Gb RAM, Intel Xeon Scalable Processors (Cascade Lake)
-** To access these different instances for a simulation, use a Slurm directive in your job script:
+** To access these different instances for a simulation, use a Slurm directive in your job script (note the connection of the --constraint directive to the ComputeResources -> Name in the config file):
   #SBATCH --partition=slurmq1

@@ Line 21: / Line 21: @@
 ** m6g.16xlarge - General purpose instance with 64 vCPUs, 256 Gb RAM, AWS Graviton 2 processors
 ** c5.24xlarge - Compute optimized instance with 96 vCPUs, and 192 Gb RAM, Intel Xeon Scalable Processors (Cascade Lake)
 * Data storage configuration
 ** A preconfigured "apps" disk is mounted to the cluster as the "/ladco" directory. This disk is based on a snapshot of an EBS volume that has scripts and software (all built on ALinux2) pre-loaded.  We frequently save snapshots of our applications disk to save specific configurations.  New clusters build off recent or newest apps snapshots

@@ Line 11: / Line 11: @@
 * [https://docs.aws.amazon.com/parallelcluster/latest/ug/cluster-configuration-file-v3.html Pcluster 3 YAML configuration file documentation]
 == Modeling spot cluster config ==
   Region: us-east-2
    Image:
@@ Line 17: / Line 34: @@
     InstanceType: c4.large
     Networking:
-     SubnetId: subnet-34d14e4e
+     SubnetId: subnet-????????
      ElasticIp: false  # true|false|EIP-id
     DisableSimultaneousMultithreading: false
     Ssh:
-     KeyName: LADCO WRF
+     KeyName: ????????
     LocalStorage:
      RootVolume:
@@ Line 29: / Line 46: @@
     Iam:
      S3Access:
-     - BucketName: ladco-backup
+     - BucketName: ????????
        EnableWriteAccess: true
   Scheduling:
@@ Line 40: / Line 57: @@
        Networking:
          SubnetIds:
-           - subnet-34d14e4e
+           - subnet-????????
          PlacementGroup:
            Enabled: True
@@ Line 62: / Line 79: @@
      EbsSettings:
        Size: 120
-       SnapshotId: snap-06867f5ffff3ed579
+       SnapshotId: snap-????????
        DeletionPolicy: Snapshot # Delete | Retain | Snapshot
    - MountDir: /data/efs
@@ Line 85: / Line 102: @@
        StorageCapacity: 2400  # 1200 | 2400| 3600
        DeploymentType: SCRATCH_1  # PERSISTENT_1 | SCRATCH_1 | SCRATCH_2
-       ExportPath: s3://ladco-backup/lustre
+       ExportPath: s3://????????/lustre
-       ImportPath: s3://ladco-backup
+       ImportPath: s3://????????
        AutoImportPolicy: NEW_CHANGED  # NEW | NEW_CHANGED
        StorageType: SSD  # HDD | SSD
@@ Line 99: / Line 116: @@
      CloudWatch:
        Enabled: true
-  CustomS3Bucket: ladco-backup
+  CustomS3Bucket: ????????
    Tags:
    - Key: Name

@@ Line 12: / Line 12: @@
 == Modeling spot cluster config ==
   Region: us-east-2
- Image:
+  Image:
-  Os: alinux2  # alinux2 | centos7 | ubuntu1804 | ubuntu2004
+   Os: alinux2  # alinux2 | centos7 | ubuntu1804 | ubuntu2004
- HeadNode:
+  HeadNode:
-  InstanceType: c4.large
+   InstanceType: c4.large
-  Networking:
+   Networking:
      SubnetId: subnet-34d14e4e
      ElasticIp: false  # true|false|EIP-id
-  DisableSimultaneousMultithreading: false
+   DisableSimultaneousMultithreading: false
-  Ssh:
+   Ssh:
      KeyName: LADCO WRF
-  LocalStorage:
+   LocalStorage:
      RootVolume:
        Size: 40
        Encrypted: false
        DeleteOnTermination: true
-  Iam:
+   Iam:
      S3Access:
      - BucketName: ladco-backup
        EnableWriteAccess: true
   Scheduling:
-  Scheduler: slurm
+   Scheduler: slurm
-  SlurmSettings:
+   SlurmSettings:
      ScaledownIdletime: 15
-  SlurmQueues:
+   SlurmQueues:
      - Name: slurmq1 # slurm directive: -p slurmq1
        CapacityType: SPOT
@@ Line 100: / Line 100: @@
        Enabled: true
   CustomS3Bucket: ladco-backup
- Tags:
+  Tags:
    - Key: Name
      Value: Feb22Pcluster

← Older revision		Revision as of 19:34, 23 February 2022
Line 1:		Line 1:
		+	= AWS Parallel Cluster version 3 =
		+
		+	* [https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html Install AWS-CLI]
		+	* [https://docs.aws.amazon.com/parallelcluster/latest/ug/parallelcluster-version-3.html Parallel Cluster version 3]
		+	* [https://docs.aws.amazon.com/parallelcluster/latest/ug/install.html Install Pcluster 3]
		+
		+	== Setting up a multi-instance cluster ==
		+	* You can now configure pcluster to have multiple queues and compute instances in single cluster. This is a nice feature because it allows you to choose from different types of instances based in your job (e.g., memory vs compute optimized) or to choose instances based on spot availability.
		+	* To decide which instances to put into your cluster use the [https://aws.amazon.com/ec2/spot/instance-advisor/ Amazon Spot Instance Advisor] to find instances that have a low frequency of interruption (< 5%). We found that instances that have >5% frequency of interruption don't perform well as spot instances. They either take too long to get into the queue, or they get kicked out of the queue too often. Use the Spot Instance Advisor to find several available instances that meet your needs within a region of interest. You should be pairing your instance search with the [https://aws.amazon.com/ec2/spot/pricing/ EC2 spot pricing table] to find less expensive options.
		+	* After finding the region and instances that will meet your needs build your pcluster config file. See the LADCO modeling instance example below.
		+	* [https://docs.aws.amazon.com/parallelcluster/latest/ug/cluster-configuration-file-v3.html Pcluster 3 YAML configuration file documentation]
		+
		+
		+
	= AWS Parallel Cluster version 2 =		= AWS Parallel Cluster version 2 =
	* [https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-install.html Install AWS-CLI]		* [https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-install.html Install AWS-CLI]

@@ Line 1: / Line 1: @@
-= AWS Parallel Cluster =
+= AWS Parallel Cluster version 2 =
 * [https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-install.html Install AWS-CLI]
 * [https://aws.amazon.com/blogs/opensource/aws-parallelcluster/ Install Pcluster]

@@ Line 6: / Line 6: @@
 Configure the cluster with a config file:
-== spot pcluster config ==
+== [[spot pcluster config]] ==
   [aws]

@@ Line 35: / Line 35: @@
   volume_iops = 1500
   encrypted = false
-−
   # Attach an "apps" volume with pre-loaded software
   [ebs apps]
   shared_dir = ladco
   ebs_volume_id = vol-*****
-−
   # Attach an "input" data volume with pre-loaded input data
   [ebs input]
   shared_dir = input
   ebs_volume_id = vol-*****
-−
   [vpc public]
   master_subnet_id = subnet-******