schema.yaml

---
title: HPC cluster
description: Oracle Cloud HPC cluster
schemaVersion: 1.1.0
version: "2021040901"
informationalText: Automated HPC cluster deployment
logoUrl: https://cloudmarketplace.oracle.com/marketplace/content?contentId=80357668

source:
  type: marketplace
  reference: 67628143
locale: "en"

outputs:
  bastion:
    title: "Bastion Instance Public IP"
    type: copyableString
    visible: true

variableGroups:
  - title: "Cluster configuration"
    variables:
      - ${targetCompartment}
      - ${ssh_key}
      - ${use_custom_name}
      - ${cluster_name}
      - ${ldap}
  - title: "Headnode options"
    variables:
      - ${bastion_ad}
      - ${bastion_shape}
      - ${bastion_ocpus}
      - ${bastion_custom_memory}
      - ${bastion_memory}
      - ${use_standard_image}
      - ${unsupported_bastion}
      - ${bastion_image_compartment}
      - ${custom_bastion_image}
      - ${unsupported_bastion_image}
      - ${bastion_username}
      - ${bastion_boot_volume_size}
  - title: "Advanced headnode options"
    variables:
      - ${bastion_advanced}
      - ${home_nfs} 
      - ${use_cluster_nfs}
      - ${cluster_nfs_path}
      - ${bastion_block}
      - ${bastion_block_volume_size}
      - ${bastion_block_volume_performance}
  - title: "Compute node options"
    variables:
      - ${ad}
      - ${cluster_network}
      - ${cluster_network_shape}
      - ${instance_pool_shape}
      - ${instance_pool_ocpus}
      - ${instance_pool_custom_memory}
      - ${instance_pool_memory}
      - ${autoscaling}
      - ${node_count}
      - ${hyperthreading}
      - ${boot_volume_size}
      - ${use_marketplace_image}
      - ${compute_username}
      - ${marketplace_listing}
      - ${unsupported}
      - ${compute_image_compartment}
      - ${image}
      - ${image_ocid}
  - title: "API authentication"
    variables:
      - ${inst_prin}
      - ${api_user_ocid}
      - ${api_fingerprint}
      - ${api_user_key}
    visible: ${autoscaling}
  - title: "Additional file system"
    variables:
      - ${add_nfs}
      - ${create_ffs}
      - ${nfs_target_path}
      - ${nfs_source_IP}
      - ${nfs_source_path}
      - ${nfs_options}
      - ${fss_compartment}
      - ${fss_ad}
  - title: "Advanced storage options"
    variables:
      - ${use_advanced}
      - ${use_scratch_nfs}
      - ${scratch_nfs_path}
      - ${scratch_nfs_type_cluster}
      - ${scratch_nfs_type_pool}
      - ${cluster_block_volume_size}
      - ${cluster_block_volume_performance}
  - title: "Network options"
    variables:
      - ${use_existing_vcn}
      - ${vcn_compartment}
      - ${vcn_id}
      - ${public_subnet_id}
      - ${private_subnet_id}
      - ${vcn_subnet}
      - ${public_subnet}
      - ${additional_subnet}
      - ${private_subnet}
  - title: "Software"
    variables:
      - ${slurm}
      - ${spack}
      - ${monitoring}
  - title: "Hidden"
    variables:
      - ${region}
      - ${tenancy_ocid}
      - ${cluster_nfs_mount}
      - ${cluster_nfs_export}
      - ${scratch_nfs_export}
      - ${scratch_nfs_mount}
      - ${marketplace_listing_id}
      - ${ssh_cidr}
      - ${marketplace_source_images}
      - ${marketplace_version_id}
    visible: false
  - title: "Debug"
    variables: 
      - ${configure}
variables:
  targetCompartment:
    title: "target compartment"
    type: oci:identity:compartment:id
    default: ${compartment_ocid}
  ad:
    type: oci:identity:availabilitydomain:name
    visible: complexExpression
    dependsOn:
      compartmentId: ${targetCompartment}
    required: true
    description: "Availability Domain"
    title: "Availability Domain"
  ssh_key:
    type: oci:core:ssh:publickey
    title: "Public SSH key"
    required: true
  use_custom_name:
    type: boolean
    title: "use custom cluster name"
    description: "Use custom name for the cluster"
    required: true
    default: false
  ldap:
    type: boolean
    title: "Configure LDAP authentication from bastion"
    description: "When selected nodes will be configured to use LDAP authentication. User and group management can be performed using cluster commands."
    required: true
    default: true
  cluster_name:
    title: "Name of the cluster"
    description: "Custom cluster name"
    type: string
    visible:
      and:
        - ${use_custom_name}
  bastion_ad:
    type: oci:identity:availabilitydomain:name
    dependsOn:
      compartmentId: ${targetCompartment}
    visible: complexExpression
    required: true
    description: "Availability Domain for bastion host"
    title: "Availability Domain"
    default: ${ad}
  bastion_shape:
    type: oci:core:instanceshape:name
    dependsOn:
      compartmentId: ${targetCompartment}
  bastion_ocpus:
    type: integer
    description: Number of OCPU's for flex shape
    minimum: 1
    maximum: 64
    default: 2
    visible:
      and: 
        - or: 
          - eq:
            - ${bastion_shape}
            - "VM.Standard.E3.Flex"
          - eq: 
            - ${bastion_shape}
            - "VM.Standard.E4.Flex"
          - eq:
            - ${bastion_shape}
            - "VM.Optimized3.Flex"
    required: false
  bastion_custom_memory: 
    title: Use custom memory size
    type: boolean
    default: false
    visible:
      and: 
        - or: 
          - eq:
            - ${bastion_shape}
            - "VM.Standard.E3.Flex"
          - eq:
            - ${bastion_shape}
            - "VM.Optimized3.Flex"
          - eq: 
            - ${bastion_shape}
            - "VM.Standard.E4.Flex"
  bastion_memory:
    title: Memory in GBS
    type: integer
    description: Number of memory for flex shape. Minimum 1GB per core.
    minimum: 1
    maximum: 1024
    default: 16
    visible:
      and: 
        - and: 
          - or: 
            - eq:
              - ${bastion_shape}
              - "VM.Standard.E3.Flex"
            - eq:
              - ${bastion_shape}
              - "VM.Optimized3.Flex"
            - eq: 
              - ${bastion_shape}
              - "VM.Standard.E4.Flex"
        - and: 
            - ${bastion_custom_memory}
    required: false
  use_standard_image:
    type: boolean
    title: "use standard bastion image"
    description: >
      "Use standard bastion image, otherwise provide custom image OCID"
    required: true
    default: true

  unsupported_bastion: 
    title: "Use unsupported image" 
    description: "Custom image ID for Bastion"
    type: boolean
    default: false
    visible:
      not:
        - ${use_standard_image}

  bastion_username: 
    title: "Default username for bastion" 
    description: "Custom image ID for Bastion"
    type: string
    default: "opc"
    visible:
      not:
        - ${use_standard_image}
 
  unsupported_bastion_image:
    title: "Image OCID"
    description: "Custom image ID for compute nodes"
    type: string
    required: true
    visible:
      and: 
        - ${unsupported_bastion}
        - not: 
            - ${use_standard_image}
  
  bastion_image_compartment:
    title: "bastion image compartment"
    type: oci:identity:compartment:id
    default: ${targetCompartment}
    visible:
      and: 
          - not:
              - ${use_standard_image}
          - not: 
              - ${unsupported_bastion}

  custom_bastion_image:
    title: "Bastion Image ID"
    description: "Custom image ID for bastion nodes. Please note that only Oracle Linux is supported as bastion image at this moment. "
    type: oci:core:image:id
    dependsOn:
      compartmentId: ${bastion_image_compartment}
    visible:
      and: 
          - not:
              - ${use_standard_image}
          - not: 
              - ${unsupported_bastion}

  bastion_boot_volume_size:
    type: integer
    required: true
    minimum: 50
    title: "Size of the boot volume in GB"
    default: 50
  bastion_advanced:
    type: boolean
    title: Show advanced options
    default: false
  bastion_block:
    type: boolean
    title: Additional block volume for shared space
    visible:
      and:
        - ${bastion_advanced}
    default: false
    required: true
  bastion_block_volume_size:
    required: true
    type: integer
    title: "Size of the additional volume in GB"
    default: 1000
    visible:
      and:
        - and:
            - ${bastion_block}
            - ${bastion_advanced}
  bastion_block_volume_performance:
    type: enum
    title: "Block volume performance"
    required: true
    enum:
      - "0.  Lower performance"
      - "10. Balanced performance"
      - "20. High Performance"
    default: "10. Balanced performance"
    visible:
      and:
        - and:
            - ${bastion_block}
            - ${bastion_advanced}
  home_nfs:
    type: boolean
    title: "shared NFS /home from bastion"
    visible:
      and:
        - ${bastion_advanced}
    default: true
  use_cluster_nfs:
    type: boolean
    title: "shared NFS volume from bastion"
    visible:
      and:
        - ${bastion_advanced}
    default: true
  cluster_nfs_path:
    title: "Cluster share mount point"
    description: "Path to NFS share"
    default: "/nfs/cluster"
    type: string
    visible:
      and:
        - ${bastion_advanced}
        - ${use_cluster_nfs}
  cluster_network:
    title: Use cluster network
    type: boolean
    description: Use ROCEv2 cluster network
    default: true

  cluster_network_shape:
    type: enum
    enum:
      - "BM.HPC2.36"
      - "BM.GPU4.8"
      - "BM.Optimized3.36"
    default: "BM.HPC2.36"
    title: "Shape of the Compute Nodes"
    description: "Shape of compute nodes used in permanent/initial cluster"
    required: true
    visible:
      and:
        - ${cluster_network}

  instance_pool_shape:
    title: "Shape of the Compute Nodes"
    required: true
    type: oci:core:instanceshape:name
    dependsOn:
      compartmentId: ${targetCompartment}
    visible:
      not:
        - ${cluster_network}
    description: "Shape of compute nodes used in permanent/initial cluster"

  instance_pool_ocpus:
    title: Cores
    type: integer
    description: Number of OCPU's for flex shape
    minimum: 1
    maximum: 64
    default: 2
    visible:
      and: 
        - or: 
          - eq:
            - ${instance_pool_shape}
            - "VM.Standard.E3.Flex"
          - eq:
            - ${instance_pool_shape}
            - "VM.Optimized3.Flex"
          - eq: 
            - ${instance_pool_shape}
            - "VM.Standard.E4.Flex"
    required: false

  instance_pool_custom_memory: 
    title: Use custom memory size
    type: boolean
    default: false
    visible:
      and: 
        - or: 
          - eq:
            - ${instance_pool_shape}
            - "VM.Standard.E3.Flex"
          - eq:
            - ${instance_pool_shape}
            - "VM.Optimized3.Flex"
          - eq: 
            - ${instance_pool_shape}
            - "VM.Standard.E4.Flex"
  instance_pool_memory:
    title: Memory in GBS
    type: integer
    description: Number of memory for flex shape. Minimum 1GB per core.
    minimum: 1
    maximum: 1024
    default: 16
    visible:
      and: 
        - and: 
          - or: 
            - eq:
              - ${instance_pool_shape}
              - "VM.Standard.E3.Flex"
            - eq:
              - ${instance_pool_shape}
              - "VM.Optimized3.Flex"
            - eq: 
              - ${instance_pool_shape}
              - "VM.Standard.E4.Flex"
        - and: 
            - ${instance_pool_custom_memory}
    required: false

  node_count:
    required: true
    type: integer
    minimum: 0
    title: "Initial cluster size"
    default: 2
    description: "Number of Compute Instances (Permanent Cluster when autoscaling)"

  hyperthreading:
    required: true
    type: boolean
    title: "Keep Hyperthreading enabled"
    default: true
    description: "When unchecked SMT will be disabled" 

  boot_volume_size:
    type: integer
    required: true
    minimum: 50
    title: "Size of the boot volume in GB"
    default: 50
    description: "Boot volume size in GB of each compute node" 

  use_marketplace_image:
    type: boolean
    title: "use marketplace image"
    description: "Use marketplace image, otherwise provide custom image OCID"
    required: true
    default: true

  marketplace_listing:
    type: enum
    title: "Image version"
    description: "Marketplace listing to use"
    required: true
    enum:
      - "1. Oracle Linux 7.9 OFED 5.3-1.0.0.1 RHCK 20210607"
      - "2. Oracle Linux 7.8 OFED 5.0-1.0.0.0 UEK 20200826"
      - "3. Oracle Linux 7.7 OFED 4.4-2.0.7.0 UEK 20200229"
      - "4. Oracle Linux 7.9 OFED 5.0-2.1.8.0 RHCK 20210709"
    default: "4. Oracle Linux 7.9 OFED 5.0-2.1.8.0 RHCK 20210709"
    visible:
      and:
        - ${use_marketplace_image}

  compute_image_compartment:
    title: "compute image compartment"
    type: oci:identity:compartment:id
    default: ${targetCompartment}
    visible:
      and: 
        - not:
            - ${use_marketplace_image}
        - not:
            - ${unsupported}

  image:
    title: "Image"
    description: "Custom image ID for compute nodes"
    type: oci:core:image:id
    required: true
    dependsOn:
      compartmentId: ${compute_image_compartment}
    visible:
      and: 
        - not:
            - ${use_marketplace_image}
        - not:
            - ${unsupported}

  compute_username: 
    title: "Default username for compute hosts" 
    description: "Custom image ID for compute hosts"
    type: string
    default: "opc"
    visible:
      not:
        - ${use_marketplace_image}

  unsupported: 
    title: "Use unsupported image" 
    description: "Custom image ID for compute nodes"
    type: boolean
    default: false
    visible:
      not:
        - ${use_marketplace_image}
 
  image_ocid:
    title: "Image OCID"
    description: "Custom image ID for compute nodes"
    type: string
    required: true
    visible:
      and: 
        - ${unsupported}

  use_advanced:
    type: boolean
    title: "Show advanced options"
    default: false
  use_scratch_nfs:
    type: boolean
    title: "Shared NFS scratch space from NVME or Block volume"
    visible:
      and:
        - ${use_advanced}
    default: true
  scratch_nfs_type_cluster:
    type: enum
    title: "Scratch storage configuration"
    enum:
      - "none"
      - "nvme"
      - "block"
    default: "nvme"
    visible:
      and:
        - ${cluster_network}
        - ${use_advanced}
        - ${use_scratch_nfs}
  scratch_nfs_type_pool:
    type: enum
    title: "Scratch storage configuration"
    enum:
      - "none"
      - "block"
    default: "none"
    visible:
      and: 
        - not:
            - ${cluster_network}
            - ${use_scratch_nfs}
        - and: 
            - ${use_advanced}
  cluster_block_volume_size:
    required: true
    type: integer
    title: "Size of the additional volume in GB"
    default: 1000
    visible:
      and:
        - or:
            - and: 
                - eq:
                    - ${scratch_nfs_type_cluster}
                    - "block"
                - and:
                    - ${use_advanced}
                    - ${cluster_network}
                    - ${use_scratch_nfs}
            - and: 
                - eq:
                    - ${scratch_nfs_type_pool}
                    - "block"
                - and:
                    - not:
                        - ${cluster_network}
                    - and: 
                        - ${use_scratch_nfs}
                        - ${use_advanced}
  cluster_block_volume_performance:
    type: enum
    title: "Block volume performance"
    required: true
    enum:
      - "0.  Lower performance"
      - "10. Balanced performance"
      - "20. High Performance"
    default: "10. Balanced performance"
    visible:
      and:
        - or:
            - and: 
                - eq:
                    - ${scratch_nfs_type_cluster}
                    - "block"
                - and:
                    - ${cluster_network}
                    - ${use_scratch_nfs}
                    - ${use_advanced}
            - and: 
                - eq:
                    - ${scratch_nfs_type_pool}
                    - "block"
                - and:
                    - not:
                        - ${cluster_network}
                    - and: 
                        - ${use_scratch_nfs}
                        - ${use_advanced}
  scratch_nfs_path:
    title: "NFS scratch space mount point"
    description: "Path to NFS share"
    default: "/nfs/scratch"
    type: string
    visible:
      and:
        - ${use_advanced}
        - ${use_scratch_nfs}
  use_existing_vcn:
    type: boolean
    title: "Use Existing VCN"
    description: "Use existing VCN or create new one"
    default: false
    required: true
  vcn_compartment:
    title: "VCN compartment"
    type: oci:identity:compartment:id
    visible: ${use_existing_vcn}
    default: ${targetCompartment}
  vcn_id:
    type: oci:core:vcn:id
    visible:
      and:
        - ${use_existing_vcn}
    title: "Existing network"
    default: ''
    required: true
    dependsOn:
      compartmentId: ${vcn_compartment}
  public_subnet_id:
    type: oci:core:subnet:id
    dependsOn:
      compartmentId: ${vcn_compartment}
      vcnId: ${vcn_id}
      hidePrivateSubnet: true
    visible: ${use_existing_vcn}
  private_subnet_id:
    type: oci:core:subnet:id
    dependsOn:
      compartmentId: ${vcn_compartment}
      vcnId: ${vcn_id}
      hidePublicSubnet: true
    visible: ${use_existing_vcn}
  vcn_subnet:
    type: string
    title: "VCN IP range"
    description: "VCN subnet"
    default: "172.16.0.0/21"
    required: true
    visible:
      not:
        - ${use_existing_vcn}
  public_subnet:
    type: string
    title: "Public subnet IP range"
    default: "172.16.0.0/24"
    description: "Must be within VCN subnet"
    required: true
    visible:
      not:
        - ${use_existing_vcn}
  additional_subnet:
    type: string
    title: "Additional subnet IP range"
    default: "172.16.1.0/24"
    description: "Must be within VCN subnet"
    required: true
    visible:
      not:
        - ${use_existing_vcn}
  private_subnet:
    type: string
    title: "Private subnet IP range"
    default: "172.16.4.0/22"
    description: "Must be within VCN subnet"
    required: true
    visible:
      not:
        - ${use_existing_vcn}
  ssh_cidr:
    type: string
    title: "Initial CIDR range allowed to SSH"
    default: "0.0.0.0/0"
    description: "Allowed SSH network in CIDR notation"
    required: true
    visible:
      not:
        - ${use_existing_vcn}
  slurm:
    type: boolean
    title: "Install SLURM"
    default: true
    description: "Needed for autoscaling" 

  spack:
    type: boolean
    title: "Install Spack package manager"
    default: false
    description: "Install Spack package manager. Requires shared folder" 

  monitoring:
    type: boolean
    title: "Install HPC Cluster Monitoring Tools"
    default: false
    description: "Install Grafana, Telegrapf, and InfluxDB tools for system monitoring."

  autoscaling:
    type: boolean
    title: "Scheduler based autoscaling"
    default: false
    description: "Requires SLURM installation. Scheduler will launch new clusters based on job requirements"
    
  inst_prin:
    type: boolean
    title: "Use Instance Principal" 
    description: "You will need to set a dynamic group and policy to allow the bastion to authenticate " 
    default: true
    visible:
      and:
        - ${autoscaling}

  api_user_key:
    type: string
    title: "API key"
    default: ""
    visible:
      not:
        - ${inst_prin}

  api_fingerprint:
    type: string
    title: "API fingerprint"
    default: ""
    visible:
      not:
        - ${inst_prin}

  api_user_ocid:
    type: string
    title: "API User OCID"
    default: ""
    visible:
      not:
        - ${inst_prin}
        
  configure: 
    type: boolean
    title: "Configure system"
    default: true
    description: "If unchecked, cluster will be launched but left unconfigured"

  add_nfs:
    type: boolean
    title: "Add another NFS filesystem"
    default: false
    description: "For FSS, leave options filed empty."
    visible: true

  create_ffs:
    type: boolean
    title: "Create FSS"
    default: false
    description: "For FSS, leave options filed empty."
    visible: ${add_nfs}

  fss_compartment:
    title: "FSS compartment"
    description: "Compartment to add the FSS Mount Target and File System"
    type: oci:identity:compartment:id
    default: ${targetCompartment}
    visible: 
      and:
        - ${add_nfs}
        - ${create_ffs}

  fss_ad:
    type: oci:identity:availabilitydomain:name
    dependsOn:
      compartmentId: ${fss_compartment}
    visible: 
      and:
        - ${add_nfs}
        - ${create_ffs}
    default: ${ad}
    description: "FSS Availability Domain"
    title: "FSS Availability Domain"
        
  nfs_target_path:
    type: string
    title: "NFS Path"
    default: "/app"
    description: "Value of the path on which you can mount the drive"
    visible: 
      and:
        - ${add_nfs}
  nfs_source_IP:
    type: string
    title: "NFS server IP"
    default: "0.0.0.0"
    description: "IP address of the NFS server"
    visible: 
      and:
        - ${add_nfs} 
        - not:
          - ${create_ffs}
  nfs_source_path:
    type: string
    title: "NFS server Path"
    default: "/app"
    description: "Value of the path on the NFS server"
    visible: ${add_nfs}
  nfs_options:
    type: string
    title: "Options"
    default: ""
    description: "Mount options, FSS requires empty string"
    visible: 
      and:
        - ${add_nfs} 
        - not:
          - ${create_ffs}