Your Browser is Out of Date

ShareDemos uses technology that works best in other browsers.
For a full experience use one of the browsers below
TechBook—Dell EMC VxRail System

close

view all chapters

VxRail hardware
VxRail hardware platforms VxRail cluster VxRail models and specifications VxRail node Hardware delivery options VxRail scaling

VxRail scaling

VxRail scaling

  • VxRail scale-out clusters start with as few as three nodes and can grow in one-node increments up to 64 nodes, providing performance and capacity to meet a wide range of use cases. Two node VxRail clusters are supported but cannot be expanded at this time. New systems can be added non-disruptively and different models can be mixed within a VxRail cluster. Flexible storage options also allow a node to start with a few drives and add drives as capacity requirements grow, as shown in the following figure. Single node upgrades and drive scalability protect an optimized initial investment by allowing customers to start with what they need and expand the VxRail cluster by adding nodes and/or drives to increase performance and capacity as needed. Consult the Dell EMC representative for assistance.

    Figure 18.          VxRail scale on demand

    A few basic rules regarding scaling are worth considering for planning:

    Balance

    • At initial deployment, the first three nodes in a cluster must have identical configuration – 2-node vSAN clusters require both nodes to be identical.
    • All nodes must be running the same version of software.
    • Cannot mix hybrid, and all-flash or all-NVMe nodes in the same cluster.
    • Can mix all-flash and all-NVMe nodes in the same cluster. 1 GbE, 10 GbE, and 25 GbE base networking cannot be mixed in the same cluster.
    • 1 GbE must be hybrid and single processor node type.
    • For G Series, all nodes in a chassis must be identical.
    • The E665 must be in its own cluster. This is because AMD processors and Intel processors cannot be placed in the same cluster. This is a VMware limitation.

    Flexibility

    • Systems in a cluster can be different models or series after initial deployment.
    • A cluster can have a varied number of drives, CPU, memory, and model types.
    • A cluster can have between 3-64, but only a max of 8 if 1 GbE networking is used.
    • For G Series, a chassis can be partially populated.

    Upgradeable options

    With VxRail, nodes can upgrade or add memory, NIC cards, cache drives, and capacity drives. GPU can be upgraded or added in the E and V Series. It is not possible to upgrade from a single processor to a dual processor VxRail node. It is not possible to upgrade from all-flash to all-NVMe. See the following table for information about which components are customer installable (replaceable).

    Table 2.              VxRail hardware component replacement list

    Hardware Component

    Customer Replaceable Unit

    (CRU)

    Field Replaceable Unit

    (FRU)

    System Memory

    Y

     

    Hard Drive

    Y

     

    Solid State Drive (cache and capacity)

    Y

     

    NVMe Cache Drive

    Y

     

    Intel Optane Drive

    Y

     

    PCIe Network Interface Cards

    Y

     

    Graphical Processing Unit (GPU)

    Y

     

    Micro SDHC Card

    Y

     

    Power Supply

    Y

     

    Processors

     

    Y

    System Motherboard

     

    Y

    Host Bus Adapter (HBA330)

    Y 

     

    BOSS controller card and M.2 SATA disk

     

    Y

    Network Daughter Card (NDC)

     

    Y