[cluster] section

Defines a cluster template that can be used to create a cluster. A config file can contain multiple [cluster] sections.

The same cluster template can be used to create multiple clusters.

The format is [cluster cluster-template-name]. The [cluster] section named by the cluster_template setting in the [global] section is used by default, but can be overridden on the pcluster command line.

cluster-template-name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

[cluster default]

additional_cfn_template

(Optional) Defines an additional Amazon CloudFormation template to launch along with the cluster. This additional template is used for creating resources that are outside of the cluster but are part of the cluster's lifecycle.

The value must be an HTTP URL to a public template, with all parameters provided.

There is no default value.

additional_cfn_template = https://<bucket-name>.s3.amazonaws.com/my-cfn-template.yaml

Update policy: If this setting is changed, the update is not allowed.

additional_iam_policies

(Optional) Specifies a list of Amazon Resource Names (ARNs) of IAM policies for Amazon EC2. This list is attached to the root role used in the cluster in addition to the permissions required by Amazon ParallelCluster separated by commas. An IAM policy name and its ARN are different. Names can't be used as an argument to additional_iam_policies.

If your intent is to add extra policies to the default settings for cluster nodes, we recommend that you pass the additional custom IAM policies with the additional_iam_policies setting instead of using the ec2_iam_role settings to add your specific EC2 policies. This is because additional_iam_policies are added to the default permissions that Amazon ParallelCluster requires. An existing ec2_iam_role must include all permissions required. However, because the permissions required often change from release to release as features are added, an existing ec2_iam_role can become obsolete.

There is no default value.

additional_iam_policies = arn:aws-cn:iam::123456789012:policy/CustomEC2Policy

Update policy: This setting can be changed during an update.

base_os

(Required) Specifies which OS type is used in the cluster.

Available options are:

Other than the specific Amazon Web Services Regions mentioned in the following table that don't support centos7. All other Amazon commercial Regions support all of the following operating systems.

base_os = alinux2

Update policy: If this setting is changed, the update is not allowed.

cluster_resource_bucket

(Optional) Specifies the name of the Amazon S3 bucket that's used to host resources that are generated when the cluster is created. The bucket must have versioning enabled. For more information, see Using versioning in the Amazon Simple Storage Service User Guide. This bucket can be used for multiple clusters. The bucket must be in the same Region as the cluster.

If this parameter isn't specified, a new bucket is created when the cluster is created. The new bucket has the name of parallelcluster-random_string. In this name, random_string is a random string of alphanumeric characters. All cluster resources are stored in this bucket in a path with the form bucket_name/resource_directory. resource_directory has the form stack_name-random_string, where stack_name is the name of one of the Amazon CloudFormation stacks used by Amazon ParallelCluster. The value of bucket_name can be found in the ResourcesS3Bucket value in the output of the parallelcluster-clustername stack. The value of resource_directory can be found in the value of the ArtifactS3RootDirectory output from the same stack.

The default value is parallelcluster-random_string.

cluster_resource_bucket = my-s3-bucket

Update policy: If this setting is changed, the update is not allowed. Updating this setting cannot be forced.

cluster_type

(Optional) Defines the type of cluster to launch. If the queue_settings setting is defined, then this setting must be replaced by the compute_type settings in the [queue] sections.

Valid options are: ondemand, and spot.

The default value is ondemand.

For more information about Spot Instances, see Working with Spot Instances.

aws iam create-service-linked-role --aws-service-name spot.amazonaws.com

cluster_type = ondemand

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

compute_instance_type

(Optional) Defines the Amazon EC2 instance type that's used for the cluster compute nodes. The architecture of the instance type must be the same as the architecture used for the master_instance_type setting. If the queue_settings setting is defined, then this setting must be replaced by the instance_type settings in the [compute_resource] sections.

If you're using the awsbatch scheduler, see the Compute Environments creation in the Amazon Batch UI for a list of supported instance types.

Defaults to t2.micro, optimal when the scheduler is awsbatch.

compute_instance_type = t2.micro

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

compute_root_volume_size

(Optional) Specifies the ComputeFleet root volume size in gibibytes (GiB). The AMI must support growroot.

The default value is 35.

compute_root_volume_size = 35

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

custom_ami

(Optional) Specifies the ID of a custom AMI to use for the head and compute nodes instead of the default published AMIs. For more information, see Modify an AMI or Build a Custom Amazon ParallelCluster AMI.

There is no default value.

custom_ami = ami-00d4efc81188687a0

If the custom AMI requires additional permissions for its launch, these permissions must be added to both the user and head node policies.

For example, if a custom AMI has an encrypted snapshot associated with it, the following additional policies are required in both the user and head node policies:

{
   "Version": "2012-10-17",
   "Statement": [
       {
           "Effect": "Allow",
           "Action": [
               "kms:DescribeKey",
               "kms:ReEncrypt*",
               "kms:CreateGrant",
               "kms:Decrypt"
           ],
           "Resource": [
               "arn:aws:kms:<AWS_REGION>:<AWS_ACCOUNT_ID>:key/<AWS_KMS_KEY_ID>"
           ]                                                    
       }
   ]
}

Update policy: If this setting is changed, the update is not allowed.

cw_log_settings

(Optional) Identifies the [cw_log] section with the CloudWatch Logs configuration. The section name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

For more information, see the [cw_log] section, Amazon CloudWatch dashboard, and Integration with Amazon CloudWatch Logs.

For example, the following setting specifies that the section that starts [cw_log custom-cw] is used for the CloudWatch Logs configuration.

cw_log_settings = custom-cw

Update policy: If this setting is changed, the update is not allowed.

dashboard_settings

(Optional) Identifies the [dashboard] section with the CloudWatch dashboard configuration. The section name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

For more information, see the [dashboard] section.

For example, the following setting specifies that the section that starts [dashboard custom-dashboard is used for the CloudWatch dashboard configuration.

dashboard_settings = custom-dashboard

Update policy: This setting can be changed during an update.

dcv_settings

(Optional) Identifies the [dcv] section with the NICE DCV configuration. The section name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

For more information, see the [dcv] section.

For example, the following setting specifies that the section that starts [dcv custom-dcv] is used for the NICE DCV configuration.

dcv_settings = custom-dcv

Update policy: If this setting is changed, the update is not allowed.

desired_vcpus

(Optional) Specifies the desired number of vCPUs in the compute environment. Used only if the scheduler is awsbatch.

The default value is 4.

desired_vcpus = 4

Update policy: This setting is not analyzed during an update.

disable_cluster_dns

(Optional) Specifies if the DNS entries for the cluster shouldn't be created. By default, Amazon ParallelCluster creates a Route 53 hosted zone. If disable_cluster_dns is set to true, the hosted zone isn't created.

The default value is false.

disable_cluster_dns = true

Update policy: If this setting is changed, the update is not allowed.

disable_hyperthreading

(Optional) Disables hyperthreading on the head and compute nodes. Not all instance types can disable hyperthreading. For a list of instance types that support disabling hyperthreading, see CPU cores and threads for each CPU core for each instance type in the Amazon EC2 User Guide for Linux Instances. If the queue_settings setting is defined, either this setting can be defined, or the disable_hyperthreading settings in the [queue] sections can be defined.

The default value is false.

disable_hyperthreading = true

Update policy: If this setting is changed, the update is not allowed.

ebs_settings

(Optional) Identifies the [ebs] sections with the Amazon EBS volumes that are mounted on the head node. When using multiple Amazon EBS volumes, enter these parameters in a list with each one separated by a comma. The section name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

Up to five (5) additional Amazon EBS volumes are supported.

For more information, see the [ebs] section.

For example, the following setting specifies that the sections that start [ebs custom1] and [ebs custom2] are used for the Amazon EBS volumes.

ebs_settings = custom1, custom2

Update policy: If this setting is changed, the update is not allowed.

ec2_iam_role

(Optional) Defines the name of an existing IAM role for Amazon EC2 that's attached to all instances in the cluster. An IAM role name and its Amazon Resource Name (ARN) are different. ARNs can't be used as an argument to ec2_iam_role.

If this option is specified, the additional_iam_policies setting is ignored. If your intent is to add extra policies to the default settings for cluster nodes, we recommend that you pass the additional custom IAM policies with the additional_iam_policies setting instead of using the ec2_iam_role settings.

If this option isn't specified, the default Amazon ParallelCluster IAM role for Amazon EC2 is used. For more information, see Amazon Identity and Access Management roles in Amazon ParallelCluster.

There is no default value.

ec2_iam_role = ParallelClusterInstanceRole

Update policy: If this setting is changed, the update is not allowed.

efs_settings

(Optional) Specifies settings related to the Amazon EFS file system. The section name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

For more information, see the [efs] section.

For example, the following setting specifies that the section that starts [efs customfs] is used for the Amazon EFS file system configuration.

efs_settings = customfs

Update policy: If this setting is changed, the update is not allowed.

enable_efa

(Optional) If present, specifies that Elastic Fabric Adapter (EFA) is enabled for the compute nodes. To view the list of EC2 instances that support EFA, see Supported instance types in the Amazon EC2 User Guide for Linux Instances. For more information, see Elastic Fabric Adapter. If the queue_settings setting is defined, either this setting can be defined, or the enable_efa settings in the [queue] section can be defined. A cluster placement group should be used to minimize latencies between instances. For more information, see placement and placement_group.

enable_efa = compute

Update policy: If this setting is changed, the update is not allowed.

enable_efa_gdr

(Optional) Starting with Amazon ParallelCluster version 2.11.3, this setting has no effect. Elastic Fabric Adapter (EFA) support for GPUDirect RDMA (remote direct memory access) is always enabled if it's supported by both the instance type and the operating system.

enable_efa_gdr = compute

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

enable_intel_hpc_platform

(Optional) If present, indicates that the End user license agreement for Intel Parallel Studio is accepted. This causes Intel Parallel Studio to be installed on the head node and shared with the compute nodes. This adds several minutes to the time it takes the head node to bootstrap. The enable_intel_hpc_platform setting is only supported on CentOS 7 (base_os = centos7).

The default value is false.

enable_intel_hpc_platform = true

Update policy: If this setting is changed, the update is not allowed.

encrypted_ephemeral

(Optional) Encrypts the ephemeral instance store volumes with non-recoverable in-memory keys, using LUKS (Linux Unified Key Setup).

For more information, see https://gitlab.com/cryptsetup/cryptsetup/blob/master/README.md.

The default value is false.

encrypted_ephemeral = true

Update policy: If this setting is changed, the update is not allowed.

ephemeral_dir

(Optional) Defines the path where instance store volumes are mounted if they are used.

The default value is /scratch.

ephemeral_dir = /scratch

Update policy: If this setting is changed, the update is not allowed.

extra_json

(Optional) Defines the extra JSON that's merged into the Chef dna.json. For more information, see Building a Custom Amazon ParallelCluster AMI.

The default value is {}.

extra_json = {}

extra_json = { "cluster" : { "skip_install_recipes" : "no" } }

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

fsx_settings

(Optional) Specifies the section that defines the FSx for Lustre configuration. The section name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

For more information, see the [fsx] section.

For example, the following setting specifies that the section that starts [fsx fs] is used for the FSx for Lustre configuration.

fsx_settings = fs

Update policy: If this setting is changed, the update is not allowed.

iam_lambda_role

(Optional) Defines the name of an existing Amazon Lambda execution role. This role is attached to all Lambda functions in the cluster. For more information, see Amazon Lambda execution role in the Amazon Lambda Developer Guide.

An IAM role name and its Amazon Resource Name (ARN) are different. ARNs can't be used as an argument to iam_lambda_role. If both ec2_iam_roleand iam_lambda_role are defined, and the scheduler is sge, slurm, or torque, then there will be no roles created. If the scheduler is awsbatch, then there will be roles created during pcluster start. For example policies, see ParallelClusterLambdaPolicy using SGE, Slurm, or Torque and ParallelClusterLambdaPolicy using awsbatch.

There is no default value.

iam_lambda_role = ParallelClusterLambdaRole

Update policy: This setting can be changed during an update.

initial_queue_size

(Optional) Sets the initial number of Amazon EC2 instances to launch as compute nodes in the cluster. If the queue_settings setting is defined, then this setting must be removed and replaced by the initial_count settings in the [compute_resource] sections.

This setting is applicable only for traditional schedulers (SGE, Slurm, and Torque). If the maintain_initial_size setting is true, then the initial_queue_size setting must be at least one (1).

If the scheduler is awsbatch, use min_vcpus instead.

Defaults to 2.

initial_queue_size = 2

Update policy: This setting can be changed during an update.

key_name

(Optional) Names an existing Amazon EC2 key pair with which to enable SSH access to the instances.

key_name = mykey

Update policy: If this setting is changed, the update is not allowed.

maintain_initial_size

(Optional) Maintains the initial size of the Auto Scaling group for traditional schedulers (SGE, Slurm, and Torque).

If the scheduler is awsbatch, use desired_vcpus instead.

This setting is a Boolean flag. If set to true, the Auto Scaling group doesn't ever have fewer members than the value of initial_queue_size, and the value of initial_queue_size must be one (1) or greater. The cluster can still scale up to the value of max_queue_size. If cluster_type = spot then the Auto Scaling group can have instances interrupted and the size can drop under initial_queue_size.

If set to false, the Auto Scaling group can scale down to zero (0) members to prevent resources from sitting idle when they aren't needed.

If the queue_settings setting is defined then this setting must be removed and replaced by the initial_count and min_count settings in the [compute_resource] sections.

Defaults to false.

maintain_initial_size = false

Update policy: This setting can be changed during an update.

master_instance_type

(Optional) Defines the Amazon EC2 instance type that's used for the head node. The architecture of the instance type must be the same as the architecture used for the compute_instance_type setting.

In Amazon Web Services Regions that have a Free Tier, defaults to the Free Tier instance type (t2.micro or t3.micro). In Amazon Web Services Regions that do not have a Free Tier, defaults to t3.micro. For more information about the Amazon Free Tier, see Amazon Free Tier FAQs.

master_instance_type = t2.micro

Update policy: If this setting is changed, the update is not allowed.

master_root_volume_size

(Optional) Specifies the head node root volume size in gibibytes (GiB). The AMI must support growroot.

The default value is 35.

master_root_volume_size = 35

Update policy: If this setting is changed, the update is not allowed.

max_queue_size

(Optional) Sets the maximum number of Amazon EC2 instances that can be launched in the cluster. If the queue_settings setting is defined, then this setting must be removed and replaced by the max_count settings in the [compute_resource] sections.

This setting is applicable only for traditional schedulers (SGE, Slurm, and Torque).

If the scheduler is awsbatch, use max_vcpus instead.

Defaults to 10.

max_queue_size = 10

Update policy: This setting can be changed during an update, but the compute fleet should be stopped if the value is reduced. Otherwise, existing nodes may be terminated.

max_vcpus

(Optional) Specifies the maximum number of vCPUs in the compute environment. Used only if the scheduler is awsbatch.

The default value is 20.

max_vcpus = 20

Update policy: This setting can't be decreased during an update.

min_vcpus

(Optional) Maintains the initial size of the Auto Scaling group for the awsbatch scheduler.

If the scheduler is SGE, Slurm, or Torque, use maintain_initial_size instead.

The compute environment never has fewer members than the value of min_vcpus.

Defaults to 0.

min_vcpus = 0

Update policy: This setting can be changed during an update.

placement

(Optional) Defines the cluster placement group logic, enabling either the whole cluster or only the compute instances to use the cluster placement group.

If the queue_settings setting is defined, then this setting should be removed and replaced with placement_group settings for each of the [queue] sections. If the same placement group is used for different instance types, it’s more likely that the request might fail due to an insufficient capacity error. For more information, see Insufficient instance capacity in the Amazon EC2 User Guide for Linux Instances. Multiple queues can only share a placement group if it’s created in advance and configured in the placement_group setting for each queue. If each [queue] sections defines a placement_group setting, then the head node can't be in the placement group for a queue.

Valid options are cluster or compute.

This parameter isn't used when the scheduler is awsbatch.

The default value is compute.

placement = compute

Update policy: If this setting is changed, the update is not allowed.

placement_group

(Optional) Defines the cluster placement group. If the queue_settings setting is defined, then this setting should be removed and replaced by the placement_group settings in the [queue] sections.

Valid options are the following values:

When set to DYNAMIC, a unique placement group is created and deleted as part of the cluster stack.

This parameter isn't used when the scheduler is awsbatch.

For more information about placement groups, see Placement groups in the Amazon EC2 User Guide for Linux Instances. If the same placement group is used for different instance types, it’s more likely that the request might fail due to an insufficient capacity error. For more information, see Insufficient instance capacity in the Amazon EC2 User Guide for Linux Instances.

There is no default value.

Not all instance types support cluster placement groups. For example, the default instance type of t3.micro doesn't support cluster placement groups. For information about the list of instance types that support cluster placement groups, see Cluster placement group rules and limitations in the Amazon EC2 User Guide for Linux Instances. See Placement groups and instance launch issues for tips when working with placement groups.

placement_group = DYNAMIC

Update policy: If this setting is changed, the update is not allowed.

post_install

(Optional) Specifies the URL of a post-install script that's run after all of the node bootstrap actions are complete. For more information, see Custom Bootstrap Actions.

When using awsbatch as the scheduler, the post-install script is run only on the head node.

The parameter format can be either http://hostname/path/to/script.sh or s3://bucketname/path/to/script.sh.

There is no default value.

post_install = s3://<bucket-name>/my-post-install-script.sh

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

post_install_args

(Optional) Specifies a quoted list of arguments to pass to the post-install script.

There is no default value.

post_install_args = "argument-1 argument-2"

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

pre_install

(Optional) Specifies the URL of a pre-install script that's run before any node deployment bootstrap action is started. For more information, see Custom Bootstrap Actions.

When using awsbatch as the scheduler, the pre-install script is run only on the head node.

The parameter format can be either http://hostname/path/to/script.sh or s3://bucketname/path/to/script.sh.

There is no default value.

pre_install = s3://<bucket-name>/my-pre-install-script.sh

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

pre_install_args

(Optional) Specifies a quoted list of arguments to pass to the pre-install script.

There is no default value.

pre_install_args = "argument-3 argument-4"

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

proxy_server

(Optional) Defines an HTTP or HTTPS proxy server, typically http://x.x.x.x:8080.

There is no default value.

proxy_server = http://10.11.12.13:8080

Update policy: If this setting is changed, the update is not allowed.

queue_settings

(Optional) Specifies that the cluster uses queues instead of a homogenous compute fleet, and which [queue] sections are used. The first [queue] section listed is the default scheduler queue. The queue section names must start with a lowercase letter, contain no more than 30 characters, and only contain lowercase letters, numbers, and hyphens (-).

Up to five (5) [queue] sections are supported.

For more information, see the [queue] section.

For example, the following setting specifies that the sections that start [queue q1] and [queue q2] are used.

queue_settings = q1, q2

Update policy: The compute fleet must be stopped for this setting to be changed for an update.

raid_settings

(Optional) Identifies the [raid] section with the Amazon EBS volume RAID configuration. The section name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

For more information, see the [raid] section.

For example, the following setting specifies that the section that starts [raid rs] be used for the Auto Scaling configuration.

raid_settings = rs

Update policy: If this setting is changed, the update is not allowed.

s3_read_resource

(Optional) Specifies an Amazon S3 resource to which Amazon ParallelCluster nodes are granted read-only access.

For example, arn:aws-cn:s3:::my_corporate_bucket* provides read-only access to the my_corporate_bucket bucket and to the objects in the bucket.

See working with Amazon S3 for details on format.

There is no default value.

s3_read_resource = arn:aws-cn:s3:::my_corporate_bucket*

Update policy: This setting can be changed during an update.

s3_read_write_resource

(Optional) Specifies an Amazon S3 resource which Amazon ParallelCluster nodes are granted read/write access to.

For example, arn:aws-cn:s3:::my_corporate_bucket/Development/* provides read/write access to all objects in the Development folder of the my_corporate_bucket bucket.

See working with Amazon S3 for details on format.

There is no default value.

s3_read_write_resource = arn:aws-cn:s3:::my_corporate_bucket/*

Update policy: This setting can be changed during an update.

scaling_settings

Identifies the [scaling] section with the Auto Scaling configuration. The section name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

For more information, see the [scaling] section.

For example, the following setting specifies that the section that starts [scaling custom] is used for the Auto Scaling configuration.

scaling_settings = custom

Update policy: If this setting is changed, the update is not allowed.

scheduler

(Required) Defines the cluster scheduler.

Valid options are the following values:

scheduler = slurm

Update policy: If this setting is changed, the update is not allowed.

shared_dir

(Optional) Defines the path where the shared Amazon EBS volume is mounted.

Don't use this option with multiple Amazon EBS volumes. Instead, provide shared_dir values under each [ebs] section.

See the [ebs] section for details on working with multiple Amazon EBS volumes.

The default value is /shared.

The following example shows a shared Amazon EBS volume mounted at /myshared.

shared_dir = myshared

Update policy: If this setting is changed, the update is not allowed.

spot_bid_percentage

(Optional) Sets the on-demand percentage used to calculate the maximum Spot price for the ComputeFleet, when awsbatch is the scheduler.

If unspecified, the current spot market price is selected, capped at the On-Demand price.

spot_bid_percentage = 85

Update policy: This setting can be changed during an update.

spot_price

(Optional) Sets the maximum Spot price for the ComputeFleet on traditional schedulers (SGE, Slurm, and Torque). Used only when the cluster_type setting is set to spot. If you don't specify a value, you are charged the Spot price, capped at the On-Demand price. If the queue_settings setting is defined, then this setting must be removed and replaced by the spot_price settings in the [compute_resource] sections.

If the scheduler is awsbatch, use spot_bid_percentage instead.

For assistance finding a Spot Instance that meets your needs, see the Spot Instance advisor.

spot_price = 1.50

Update policy: This setting can be changed during an update.

tags

(Optional) Defines tags to be used by Amazon CloudFormation.

If command line tags are specified via --tags, they are merged with config tags.

Command line tags overwrite config tags that have the same key.

Tags are JSON formatted. Don't use quotes outside of the curly braces.

For more information, see Amazon CloudFormation resource tags type in the Amazon CloudFormation User Guide.

tags = {"key" : "value", "key2" : "value2"}

Update policy: If this setting is changed, the update is not allowed.

template_url

(Optional) Defines the path to the Amazon CloudFormation template that's used to create the cluster.

Updates use the template that was originally used to create the stack.

Defaults to https://aws_region_name-aws-parallelcluster.s3.amazonaws.com/templates/aws-parallelcluster-version.cfn.json.

template_url = https://us-east-1-aws-parallelcluster.s3.amazonaws.com/templates/aws-parallelcluster-2.11.9.cfn.json

Update policy: This setting is not analyzed during an update.

vpc_settings

(Required) Identifies the [vpc] section with the Amazon VPC configuration where the cluster is deployed. The section name must start with a letter, contain no more than 30 characters, and only contain letters, numbers, hyphens (-), and underscores (_).

For more information, see the [vpc] section.

For example, the following setting specifies that the section that starts [vpc public] is used for the Amazon VPC configuration.

vpc_settings = public

Update policy: If this setting is changed, the update is not allowed.