Enabling cluster export services

Cluster Export Services (CES) is a key component of the IBM Storage Scale architecture, which is designed to enable access to data stored in the Scale cluster and Object Storage (Storage Scale) system. CES plays a critical role in providing efficient and versatile data access to meet the diverse needs of modern enterprises. This flexibility allows organizations to support a wide range of applications and its use cases.

Enabling colocation designates the subset of Storage server as protocol nodes. If disabled, protocol nodes are created on dedicated virtual servers or bare metal, depending on the specified protocol server profile.

The colocation feature avoids the need to provision extra virtual servers and improves the performance. It is also supported on Bare Metal servers.

Architecture diagram.
CES node on Storage Scale

Before you begin

Before you begin, review the following information:

  1. To begin with the Scale cluster deployment, see Before you begin deploying topic.

  2. For more information on cluster export service see, how CES works topic.

Configuring CES deployment

To enable the CES feature on a Scale cluster, the following variables need to be defined in your workspace.

CES Variables
CES Variable Description Example value
protocol_subnets_cidr Provide the CIDR block required for the creation of the protocol private subnet. Single CIDR block is required. If using a hybrid environment, modify the CIDR block to avoid conflicts with any on-premises CIDR blocks. Ensure the selected CIDR block size can accommodate the maximum number of protocol nodes expected in your cluster. For more information on CIDR block size selection, see Choosing IP ranges for your VPC. "10.241.40.0/24"
protocol_instances Specify the list of virtual server instances to be provisioned as protocol nodes in the cluster. Each object in the list includes the instance profile (machine type), the count (number of instances), the image (OS image to use). This configuration allows for a unified data management solution, enabling different clients to access the same data using NFS protocol. This input can be used to provision virtual server instances (VSI). If persistent, high-throughput storage is required, consider using bare metal instances instead. Ensure you provide valid instance profiles. A maximum of 32 VSI or bare metal nodes are supported. For more information, see Instance Profiles. [{profile = "cx2-32x64" count = 2 image = "hpcc-scale5232-rhel810-v1"}]
filesets_config Specify a list of file sets with client mount paths and optional storage quotas (0 means no quota) to be created within the IBM Storage Scale file system. [{client_mount_path = "/mnt/scale/tools" quota = 0},{client_mount_path = "/mnt/scale/data" quota = 0}]
client_instances Specify the list of virtual server instances to be provisioned as client nodes in the cluster. Each object includes the instance profile (machine type), number of instances (count), OS image to use. This configuration allows customization of the compute tier to suit specific performance and workload requirements. The profile must match a valid IBM Cloud VPC Gen2 instance profile format. For more information, see Instance Profiles. [{profile = "cx2-2x4" count = 0 image = "ibm-redhat-8-10-minimal-amd64-6"}]
colocate_protocol_instances Enable this option to colocate protocol services on the same virtual server instances used for storage. When set to true, the storage nodes will also act as protocol nodes for reducing the need for separate infrastructure. This can optimize resource usage and simplify the cluster setup, especially for smaller environments or cost-sensitive deployments. For larger or performance-intensive workloads, consider deploying dedicated protocol instances instead. true

The successful scale deployment with the CES feature enabled consists of different clusters:

  • Storage cluster with defined storage and CES protocol nodes.
  • Client cluster with defined client nodes that mounts file shares that are exported by protocol nodes with NFS protocol.
  • (Optional) Compute cluster with defined compute nodes.

Verifying CES on the file system

  1. Log in to any of the clusters (storage or compute nodes) by running the following SSH command:

    ssh -J root@BASTION_SERVER vpcuser@STORAGE_NODE
    
  2. To view the cluster shared root configuration on the storage cluster, run the following command:

    mmlsconfig cesSharedRoot
    
  3. To list the protocol nodes in the cluster, run the following command:

    mmces node list
    
  4. To view the protocol cluster information, use the mmlscluster command:

    mmlscluster --ces
    
  5. Use the service list command that provides comprehensive list of the services that are running in the CES cluster, use --verbose and -a flag for detailed information:

    mmces service list --verbose -a
    
  6. Use the mmuserauth command to view the details on the type of authentication used for CES:

    mmuserauth service check
    
  7. Use the mmnfs export command to add, change, list, load, or remove NFS export declarations for IP addresses on nodes that are configured as CES types. Use list to view the current NFS exports:

    mmnfs export list
    
  8. Use the mmlsquota command to display quota information for a user, group, or file set. The -j flag is used for displaying the quota for file set in a file system.

    mmlsquota -j data FILESYSTEM
    

The CES feature is only available with the custom image that is provided by the solution.