Deployment values
The following deployment values can be used to configure the Spectrum LSF cluster instance on IBM Cloud®:
NOT_SET refers to an empty value in the IBM Cloud catalog tile. If you are using other modes like CLI, API, or Schematics UI directly to provision the HPC cluster, leave the values empty.
| Value | Description | Is it required? | Default value |
|---|---|---|---|
api_key |
This is the IBM Cloud API key for the IBM Cloud account where the Spectrum LSF cluster needs to be deployed. For more information on how to create an API key, see Managing user API keys. | Yes | None |
app_center_db_pwd |
Password for MariaDB. Note: Password should be at least eight characters, must have one number, one lowercase letter, one uppercase letter, and at least one special character. | No | None |
app_center_gui_pwd |
Password for LSF Application Center GUI. Note: Password should be at least eight characters, must have one number, one lowercase letter, one uppercase letter, and at least one special character. | No | None |
cluster_prefix |
Prefix that is used to name the IBM Spectrum LSF cluster and IBM Cloud resources that are provisioned to build the IBM Spectrum LSF cluster instance. You cannot create more than one instance of the LSF cluster with the same name. Make sure that the name is unique. | No | hpcc-lsf |
dedicated_host_enabled |
Set to 'true' to use dedicated hosts for compute hosts (default is 'false'). Spectrum LSF still dynamically provisions compute hosts at public virtual server instances, and dedicated hosts are used only for static compute hosts that are
provisioned at the time the cluster is created. The number of dedicated hosts and the profile names for dedicated hosts are calculated from worker_node_min_count and worker_node_instance_type. |
No | false |
dedicated_host_placement |
Specify 'pack' or 'spread'. The 'pack' option deploys virtual server instances on one dedicated host until full before moving on to the next dedicated host. The 'spread' option deploys instances in round-robin fashion across all of the dedicated hosts. The 'spread' option should result in mostly even distribution of instances on the hosts, while the 'pack' option might result in one dedicated host being mostly empty. | No | spread |
enable_app_center |
Set to "true" to install and enable the use of IBM Spectrum LSF Application Center GUI. For more information on system requirements, see System requirements for IBM Spectrum LSF Application Center Version 10.2 Fix Pack 13. | No | false |
hyperthreading_enabled |
Setting this to true enables hyper-threading in the worker nodes of the cluster (default). Otherwise, hyper-threading is disabled. | No | true |
ibm_customer_number |
Comma-separated list of the IBM Customer Number(s) (ICN) that is used for the Bring Your Own License (BYOL) entitlement check. For more information on how to find your ICN, see What is my IBM Customer Number (ICN)? | Yes | None |
management_image_name |
Name of the custom image that you want to use to create virtual server instances in your IBM Cloud account to deploy the IBM Spectrum LSF cluster. By default, the automation uses a base image with additional software packages documented here. If you would like to include your application-specific binary files, follow the instructions in Planning for custom images to create your own custom image and use that to build the IBM Spectrum LSF cluster through this offering. | No | hpcc-lsf10-scale5120-rhel88-5-0 |
login_node_instance_type |
Specifies the virtual server instance profile type to be used to create the login node for the Spectrum LSF cluster. For choices on profile types, see Instance profiles. | No | bx2-2x8 |
management_node_count |
Number of management hosts. This is the total number of management hosts and management host candidates. Enter a value in the range 1 - 3. | No | 2 |
management_node_instance_type |
Specify the virtual server instance profile type to be used to create the management hosts for the Spectrum LSF cluster. For choices on profile types, see Instance profiles. | No | bx2-4x16 |
remote_allowed_ips |
Comma-separated list of IP addresses that can access the Spectrum LSF instance through an SSH or RDP interface. For security purposes, provide the public IP addresses that are assigned to the devices that are authorized to establish SSH or RDP connections (for example, ["169.45.117.34"]). To fetch the IP address of the device, use https://ipv4.icanhazip.com/. | Yes | None |
resource_group |
Resource group name from your IBM Cloud account where the VPC resources should be deployed. For additional information on resource groups, see Managing resource groups. | No | Default |
scale_compute_cluster_filesystem_mountpoint |
Compute cluster (accessingCluster) file system mount point. The accessingCluster is the cluster that accesses the owningCluster. For more information, see Mounting a remote GPFS file system. | No | /gpfs/fs1 |
scale_compute_cluster_gui_password |
Password for compute cluster GUI. Note: Password should be at least 8 characters, must have one number, one lowercase letter, one uppercase letter, and at least one unique character. Password should not contain username. | No | None |
scale_compute_cluster_gui_username |
GUI user to perform system management and monitoring tasks on the compute cluster. Note: Username should be at least four characters, any combination of lowercase and uppercase letters. | No | None |
scale_filesystem_block_size |
File system block size. Storage Scale supported block sizes (in bytes): 256K, 512K, 1M, 2M, 4M, 8M, 16M. | No | 4M |
scale_storage_cluster_filesystem_mountpoint |
Storage Scale storage cluster (owningCluster) file system mount point. The owningCluster is the cluster that owns and serves the file system to be mounted. For more information, see Mounting a remote GPFS file system. | No | /gpfs/fs1 |
scale_storage_cluster_gui_password |
Password for Storage Scale storage cluster GUI. Note: Password should be at least 8 characters, must have one number, one lowercase letter, one uppercase letter, and at least one unique character. Password should not contain username. | No | None |
scale_storage_cluster_gui_username |
GUI user to perform system management and monitoring tasks on the storage cluster. Note: Username should be at least 4 characters, any combination of lowercase and uppercase letters. | No | None |
scale_storage_image_name |
Name of the custom image that you would like to use to create virtual machines in your IBM Cloud account to deploy the Storage Scale storage cluster. By default, the automation uses a base image plus the Storage Scale software and any other software packages that it requires. If you'd like, you can follow the instructions for Planning for custom images to create your own custom image with additional software and use that to build the Storage Scale storage cluster through this offering. | No | hpcc-scale5120-rhel88 |
scale_storage_node_count |
The number of Storage Scale storage nodes that are provisioned at the time the cluster is created. Enter a value in the range 2 - 18. It must be divisible of 2. | No | 4 |
scale_storage_node_instance_type |
Specify the virtual server instance storage profile type name to be used to create the Storage Scale storage nodes for the Spectrum LSF cluster. For more information, see Instance profiles | No | cx2d-8x16 |
spectrum_scale_enabled |
Setting this to true will enable Storage Scale integration with the cluster. Otherwise, Storage Scale integration will be disabled (default). By entering 'true' for the property you have also agreed to one of the two conditions: (1) You are using the software in production and confirm you have sufficient licenses to cover your use under the International Program License Agreement (IPLA). (2) You are evaluating the software and agree to abide by the International License Agreement for Evaluation of Programs (ILAE). Note: Failure to comply with licenses for production use of software is a violation of IBM International Program License Agreement. | No | false |
ssh_key_name |
Comma-separated list of names of the SSH key that is configured in your IBM Cloud account and is used to establish a connection to the Spectrum LSF management host. Ensure that the SSH key is present in the same resource group and region where the cluster is being provisioned. If you do not have an SSH key in your IBM Cloud account, create one by using the instructions given at SSH keys. | Yes | None |
storage_node_instance_type |
Specify the virtual server instance profile type to be used to create the storage node for the Spectrum Symphony cluster. The storage node is the one that is used to create an NFS instance to manage the data for HPC workloads. For choices on profile types, see Instance profiles. | No | bx2-2x8 |
vpc_name |
Name of an existing VPC in which the cluster resources are deployed. If no value is given, then a new VPC is provisioned for the cluster. Learn more. | No | "" |
vpn_enabled |
Set the value as true to deploy a VPN gateway for VPC in the cluster. | No | false |
vpn_peer_address |
The peer public IP address to which the VPN will be connected. | No | "" |
vpn_peer_cidrs |
Comma-separated list of peer CIDRs (for example, 192.168.0.0/24) to which the VPN will be connected. | No | "" |
vpn_preshared_key |
The pre-shared key for the VPN. | No | "" |
worker_node_instance_type |
Specify the virtual server instance profile type to be used to create the worker nodes for the Spectrum LSF cluster. The worker nodes are the ones where the workload execution takes place and the choice should be made according to the characteristics
of the workloads. For choices on profile types, see Instance profiles. Note: If dedicated_host_enabled == true, the available instance prefix (for example, bx2 and cx2) can be
limited depending on your target region. Check ibmcloud target -r {region_name}; ibmcloud is dedicated-host-profiles. |
No | bx2-4x16 |
worker_node_max_count |
The maximum number of worker nodes that can be deployed in the Spectrum LSF cluster. In order to use the Resource Connector feature to dynamically create and delete worker nodes based on workload demand, the value that is selected for this parameter must be larger than worker_node_min_count. If you plan to deploy only static worker nodes in the
LSF cluster, for example, when you use Storage Scale storage, the value for this parameter should be equal to worker_node_min_count. Enter a value in the range 1 - 500. |
No | 10 |
worker_node_min_count |
The minimum number of worker nodes. This is the number of static worker nodes that will be provisioned at the time the cluster is created. If using NFS storage, enter a value in the range 0 - 500. If using Storage Scale storage, enter a
value in the range 1 - 64. Note: Storage Scale requires a minimum of 3 compute nodes (combination of primary, secondary, management, and worker nodes) to establish a quorum and maintain data consistency in the event of a node failure. Therefore, the minimum value of 1 might need to be larger if the value specified for management_node_count is less than
2. |
No | 0 |
zone |
IBM Cloud zone name within the selected region where the Spectrum LSF cluster should be deployed. To get a full list of zones within a region, see Get zones by using the CLI. | Yes | None |
vpc_cidr_block |
Creates the address prefix for the new VPC, when the vpc_name variable is empty. Only a single address prefix is allowed. For more information, see Setting IP ranges. |
No | ["10.241.0.0/18"] |
vpc_cluster_private_subnets_cidr_blocks |
This is the CIDR block required for the creation of the compute and storage cluster private subnet. Modify the CIDR block if it has already been reserved or used for other applications within the VPC or conflicts with any on-premises CIDR blocks when using a hybrid environment. Provide only one CIDR block for the creation of the compute and storage subnet. Make sure to select a CIDR block size that will accommodate the maximum number of management, storage, both static and dynamic worker nodes that you expect to have in your cluster. For more information on CIDR block size selection, see Choosing IP ranges for your VPC. | No | ["10.241.0.0/20"] |
vpc_cluster_login_private_subnets_cidr_blocks |
This is the CIDR block required for the creation of the login cluster private subnet. Modify the CIDR block if it has already been reserved or used for other applications within the VPC or conflicts with any on-premises CIDR blocks when using a hybrid environment. Provide only one CIDR block for the creation of the login subnet. Since login subnet is used only for creation of login virtual server instance provide a CIDR range of /28. | No | ["10.241.16.0/28"] |
subnet_id |
Existing subnet ID under the VPC, where the cluster will be provisioned. | No | None |
dns_domain |
IBM Cloud DNS Services domain name used for the IBM Cloud LSF cluster. | No | lsf.com |
cluster_id |
Unique ID of the cluster used by LSF for configuration of the resources. This can be up to 39 alphanumeric characters including the underscore (_), the hyphen (-), and the period (.) characters. Other special characters and spaces are not allowed. Do not use the name of any host or user as the name of your cluster. You cannot change the cluster ID after deployment. | No | HPCCluster |
enable_customer_managed_encryption |
Setting this value to true will enable customer managed encryption. Otherwise, encryption will be provider managed. |
No | false |
kms_instance_id |
Unique identifier of the Key Protect instance associated with the Key Management Service (KMS). This ID can be found under the details of the KMS, see View key-protect ID. | No | None |
kms_key_name |
Provide the existing KMS encryption key name that you want to use for the IBM Cloud LSF cluster, (for example; kms_key_name: my-encryption-key). | No | None |
custom_file_shares |
This value is used to mount points and sizes in GB and IOPS range of file shares that can be used to customize shared file storage layout. Provide the details for up to 5 shares. Each file share size in GB supports different range of IOPS. For more information, see file share IOPS value. | No | [{ mount_path = "/mnt/binaries", size = 100, iops = 2000 }, { mount_path = "/mnt/data", size = 100, iops = 6000 }] |
enable_ldap |
Set this option to true to enable LDAP for IBM Cloud HPC, with the default value set to false. |
No | false |
ldap_basedns |
This DNS domain name is used for configuring the LDAP server. If an LDAP server is already existing, then ensure to provide the associated DNS domain name. | No | hpcaas.com |
ldap_server |
Provide the IP address for an existing LDAP server. If no address is given, a new LDAP server will be created. | No | None |
ldap_admin_password |
The LDAP administrative password should be 8 to 20 characters long, with a mix of at least three alphabetic characters, including one uppercase and one lowercase letter. It must also include two numerical digits and at least one special character from (~@_+:). It is important to avoid including the username in the password for enhanced security. [This value is ignored for an existing LDAP server]. | No | None |
ldap_user_name |
Custom LDAP user for performing cluster operations. Note: Username should be between 4 to 32 characters, (any combination of lowercase and uppercase letters).[This value is ignored for an existing LDAP server]. | No | None |
ldap_user_password |
The LDAP user password should be 8 to 20 characters long, with a mix of at least three alphabetic characters, including one uppercase and one lowercase letter. It must also include two numerical digits and at least one special character from (~@_+:). It is important to avoid including the username in the password for enhanced security. [This value is ignored for an existing LDAP server]. | No | None |
ldap_vsi_profile |
Profile to be used for LDAP virtual server instance. | No | cx2-2x4 |
ldap_vsi_osimage_name |
Image name to be used for provisioning the LDAP instances. | No | ibm-ubuntu-22-04-3-minimal-amd64-3 |
compute_image_name |
Name of the custom image that you use to create virtual server instances in your IBM Cloud account to deploy the IBM Spectrum LSF cluster worker nodes. By default, the solution uses a RHEL 8-8 OS image with additional software packages mentioned here. If you would like to include your application-specific binary files, follow the instructions in Planning for custom images to create your own custom image and use that to build the IBM Spectrum LSF cluster through this offering. | No | hpcaas-lsf10-rhel88-compute-v3 |
enable_vpc_flow_logs |
Flag to enable VPC flow logs. If true, a flow log collector will be created. | No | false |
create_authorization_policy_vpc_to_cos |
Set it to true if authorization policy is required for VPC to access COS. This can be set to false if authorization policy already exists. For more information on how to create authorization policy manually, see creating authorization policies for VPC flow log. |
No | false |
existing_cos_instance_guid |
This is the GUID of the COS instance to create a flow log collector. | No | None |
existing_storage_bucket_name |
Name of the COS bucket to collect VPC flow logs. | No | None |
is_flow_log_collector_active |
This value indicates whether the collector is active. If false, then this collector is created in inactive mode. | No | None |
cluster_subnet_id |
This is the existing cluster subnet ID under the VPC, where the cluster will be provisioned. | No | None |
login_subnet_id |
This is the existing login subnet ID under the VPC, where the bastion/login will be provisioned. | No | None |
dns_instance_id |
This is the IBM Cloud HPC DNS service resource id. | No | None |
dns_custom_resolver_id |
This is the IBM Cloud DNS custom resolver id. | No | None |