Compute

This section answers questions related to compute/storage node management and upsizing.

  1. If I add four compute nodes, is each node added one at a time or would all four pull in at the same time?
    Nodes can be purchased in any number of units, and you can add them in groups from the user interface.
    Node addition is a 3-step procedure:
    1. Compute configuration that readies the hardware and configures XCC.
    2. Add node to OpenShift® Container Platform.
    3. Add node to IBM Storage Scale.
    All three steps can be done with multiple nodes at a time. However, in Step a, it is recommended to fill the nodes from lower RU to higher RU to maintain the center of gravity of the rack.
  2. How long does each node take to add?
    It takes approximately 30 minutes to add one node into IBM Storage Fusion HCI System ecosystem. Contact IBM support to help with node addition. For more information about node upsize, see Adding nodes to the cluster.

    Compute configuration steps take 20 minutes to configure. If in case the node needs firmware updates, it will take another 90 minutes to update the firmware and complete the configuration.

  3. Where does the Service Support Representative (SSR) configure the XCC from?
    The SSR does not configure the XCC. Based on the Link-Local Address (LLA), the IBM Storage Fusion HCI System management software connects to the XCC using the default credentials. It also performs the remaining configurations.

  4. After the SSR does the rack-and-stack all at once, can I add the nodes later? Does the SSR verify the correctness of the installation or connectivity?
    SSR assists you with Step 1: Configure node wherein new nodes are added to cluster for configuration. The configuration completion includes verification. After the completion, you can do the remaining two phases of adding to OpenShift Container Platform and IBM Storage Scale.
  5. Would it be possible to add on drives to the nodes after they join the IBM Storage Fusion HCI System cluster? For example, there is a need for more storage over time.
    Yes, if you started with less than the full disk count, you can add more drives to the storage nodes. Drives need to be added across all the storage nodes. You can also add more storage nodes but they also have to match the number of existing disks.

  6. Do the 64-core compute nodes provide double the number of cores and quadruple the amount of RAM compared to the original compute nodes that were released a year ago?
    IBM Storage Fusion HCI System supports 64-core servers (9155-C04, 9155-C05, and 9155-C14) that come with a base memory of 1024 GB RAM, upgradeable to 2048 GB RAM. It is an alternative to 32-core servers (9155-C00, 9155-C01, and 9155-C10), so you can get the most out of demanding applications that are ideal for businesses with high data processing needs.
  7. Are there any interop concerns if compute nodes are not on the same firmware release?
    It is always recommended to have all the compute nodes at the same firmware level.
  8. Is it possible to remove the expansion rack from the cluster or IBM Storage Fusion HCI System after adding an expansion rack to a base rack?
    There is no provision to remove the expansion after the nodes are added to the OpenShift Cluster. If such a need arises, then contact IBM support.