Servicing IBM Storage Scale protocol nodes

Use these steps to upgrade the OS and to install kernel updates, firmware updates, and driver updates on IBM Storage Scale protocol nodes.

It is advisable to match OS, kernel, firmware, and driver across all nodes within a cluster to help with performance and to ease debugging. As a part of this procedure, ensure the following:
  • Before upgrading OS or installing kernel updates, refer to the IBM Storage Scale FAQ to ensure that the version of IBM Storage Scale currently running is supported on the version of OS that you are planning to upgrade to or the kernel update that you are planning to install: IBM Storage Scale FAQ in IBM® Documentation.

    It is possible that an IBM Storage Scale upgrade might be required to get to a version of IBM Storage Scale that supports the newly wanted OS. If so, upgrade IBM Storage Scale first and then the OS.

  • If kernel update is for a new OS (RHEL 7.3 vs RHEL 7.2, for example), always update the OS before applying the kernel update.
  • When taking nodes offline to upgrade the OS or installing kernel updates, firmware updates, or driver updates, ensure the following:
    • The Quorum does not break.
    • Enough NSD nodes remain up to access NSDs.
    • The remaining nodes can handle the desired workload.
  1. Uninstall the applicable drivers as follows.
    1. Obtain and extract the applicable drivers.
    2. Suspend CES on the node being upgraded. 
      mmces node suspend -N NodeBeingUpgraded --stop
    3. Shut down GPFS on the node being upgraded. 
      mmshutdown -N NodeBeingUpgraded
    4. Find the uninstallation script within the drivers package and execute it on the node being upgraded.
  2. Create a local repository for the OS upgrade.

    A repository must be created so that the OS can be upgraded. This repository can be DVD or ISO based. Make sure that you remove any repositories pointing to old OS versions.

  3. Upgrade the OS and reboot the node after the OS upgrade.

    For more information, see Guidance for upgrading the operating system on IBM Storage Scale nodes.

  4. Install the kernel update and reboot the node after the kernel update.
  5. Rebuild the GPFS portability layer using the mmbuildgpl command.

    For more information, see Building the GPFS portability layer on Linux nodes.

  6. Install the required firmware updates and reboot the node after the firmware update.
  7. Install the latest drivers, as applicable and reboot the node after the driver update.
  8. Verify that GPFS is active on the node and then resume CES. 
    mmgetstate -a
mmces node resume -N NodeBeingUpgraded --start
mmces node list
mmces service list -a
mmces address list
  9. Repeat the preceding steps on each protocol node that needs to be serviced, one node at a time.