Prerequisites

These prerequisites are applicable for Power9, IBM® Storage Scale System Utility Node, and bring-your-own management server (BYOMS).

Manufacturing and SSR tasks

The following tasks must be complete before the installation:
  1. Verify that all hardware is cleaned and all IP addresses are set and reachable over the appropriate networks by using the Code 20 operation.
  2. Ensure that the /etc/hosts file is blank.
  3. Place the correct IBM Storage Scale System .tgz installation file (appropriate edition) in the /home/deploy directory.
    If an upgrade is needed, download the latest version from IBM Fix Central and replace the existing file.
  4. Clear any existing network bridges.
  5. Remove all images and containers from the system.
  6. Clean and regenerate all SSH keys on the nodes.
  7. Confirm that all system components are updated to the latest supported versions before the system was shipped from manufacturing.
    Components 3000 3200 3500 5000 6000 Utility Node BYOMS
    IBM Storage Scale 6.0.1.0 efix4 6.0.1.0 efix4 6.0.1.0 efix4 6.0.1.0 efix4 6.0.1.0 efix4 6.0.1.0 efix4 6.0.1.0 efix4
    Operating
    system    		 RHEL 8.10 RHEL 8.10 RHEL 9.6 RHEL 8.10 RHEL 9.6 RHEL 9.6 RHEL 8.10
    Kernel 4.18.0-553.124.1.el8_10 4.18.0-553.124.1.el8_10 5.14.0-570.113.1.el9_6 4.18.0-553.124.1.el8_10 5.14.0-570.113.1.el9_6 5.14.0-570.113.1.el9_6 4.18.0-553.124.1.el8_10
    Systemd 239-82.el8_10.16 239-82.el8_10.16 252-51.el9_6.4 239-82.el8_10.16 252-51.el9_6.4 252-51.el9_6.4 239-82.el8_10.16
    Network
    Manager    		 1.40.16-20.el8_10 1.40.16-20.el8_10 1.52.0-10.el9_6 1.40.16-20.el8_10 1.52.0-10.el9_6 1.52.0-10.el9_6 1.40.16-20.el8_10
    Glibc 2.28-251.el8_10.34 2.28-251.el8_10.34 2.34-168.el9_6.24 2.28-251.el8_10.34 2.34-168.el9_6.24 2.34-168.el9_6.24 2.28-251.el8_10.34
    OFED 25.10-2.4.1 25.10-2.4.1 25.10-2.4.1 25.10-2.4.1 25.10-2.4.1 25.10-2.4.1 25.10-2.4.1

Verifying code version

Ensure that you install correct code versions for the IBM Storage Scale System deployment.
  1. Download the unified container package, which supports both Data Management Edition and Data Access Edition from IBM Fix Central.
    Scale_System_UNIFIED-7.0.1.0-EMS.tgz
  2. Ensure deployment nodes are on a supported architecture:
    • x86_64: For IBM Storage Scale Utility Node or bring-your-own management server (BYOMS)
    • ppc64le: For POWER-based management servers. For details about functions supported by respective containers, see Support Matrix.

Client nodes

  • Ensure that client nodes are running MOFED 4.9.x or higher, and these nodes are converted to verbsRDMA core libraries.
  • IBM Storage Scale System cluster node configuration is optimized to run IBM Storage Scale RAID functions. Protocols, other gateways, or any other non-IBM Storage Scale System services must not be run on IBM Storage Scale System management server nodes or I/O server nodes. In a cluster with high I/O load, do not use IBM Storage Scale System nodes as a cluster manager or file system manager. For optimal performance the NSD client nodes that access IBM Storage Scale System nodes must be properly configured. IBM Storage Scale System ships with gssClientConfig.sh script located in /usr/lpp/mmfs/samples/gss/ directory. This script can be used to configure the client as follows:
    /usr/lpp/mmfs/samples/gss/gssClientConfig.sh <Comma Separated list of 
client nodes or nodeclass>
    You can run the following to see configuration parameter settings without setting them:
    /usr/lpp/mmfs/samples/gss/gssClientConfig.sh -D

    After running this script, restart GPFS on the affected nodes for the optimized configuration settings to take effect.

    Important: Start of changeDo not run gssClientConfig.sh unless you fully understand the impact of each setting on the customer environment. Use the -D option to decide if all or some of the settings might be applied. Then, individually update each client node settings as required.End of change

Best practices

General deployment and upgrade guidelines
  • Use only the container image provided with IBM Storage Scale System.

    Red Hat package installation outside the deployment or upgrade flow is not supported.

  • Do not log in or run containers over the high-speed network.

    Use a campus connection, dedicated management network, or virtual console instead.

  • Run the container only from a direct SSH connection or virtual console.

    Avoid starting it from an I/O node or any node that might reboot during the process.

  • Stop all IBM Storage Scale tracing (mmtrace, mmtracectl) before starting the container or deploying any node.

    The container blocks if active tracing is detected.

  • Suspend all IBM Storage Scale and I/O operations before performing an upgrade. Wait for completion of the following commands:
    • mmadddisk, mmdeldisk, mmrestripefile, mmrestripefs, mmrpldisk
    • mmapplypolicy, mmcheckquota, mmfsck, mmlssnapshot, mmrestorefs
    • Snapshot operations: mmcrsnapshot, mmdelsnapshot
  • Place protocol nodes into file system maintenance mode before upgrades to prevent service disruption (recommended, not mandatory).
  • Do not update the management server over the high-speed network.

    Perform updates only via the management network, campus connection, or virtual console.

  • Do not install any additional packages or services that the deployment container repository does not contain. If you want to include any package or service, open an IDEA/RFE for further action from IBM.
  • Start of changeThe declustered arrays must contain minimum 20 drives for a 16+2p RAID configuration (18 code-width drives + 2 spare drives) and 21 drives for a 16+3p RAID configuration (19 code-width drives + 2 spare drives).End of change
Configuration and compatibility
  • Run the essrun config load command on all storage nodes, including management and protocol nodes before:
    • Enabling Admin Mode Central
    • Deploying protocol nodes by using the installation toolkit
  • Ensure all node hostnames and IPs are correctly defined, and that:
    High-speed network names include a suffix of the management network name.
    • Management: IOnode1.test.net
    • High-speed: IOnode1-hs.test.net
  • Do not install additional OS packages manually. If additional packages are required, contact IBM to request inclusion in future container releases.
System-level
  • Enable TRIM support for IBM Storage Scale System 3000 where appropriate. By default, only 80% of space is allocated. Consult IBM Development if using beyond 80% to assess performance trade-offs.
  • Configure a campus or secondary management connection before deploying the container. This enhances remote access and system manageability.
  • Run gnrhealthcheck after adding an I/O node and before you create virtual disk sets to detect issues as follows:

    Duplicate hostid values, which can cause environment instability.

Security and credentials
  • Review cipher usage in your environment, especially for the campus connection.
    • The use of encryption ciphers versus AUTHONLY is under review.
    • Understand the security implications and consult IBM if needed.