When a host is set up to use stateless caching, the host uses a cached image if the vSphere Auto Deploy Server is not available. To use stateless caching, you must configure a host profile. You can apply that host profile to other hosts that you want to set up for stateless caching.

Prerequisites

  • Decide which disk to use for caching and determine whether the caching process will overwrite an existing VMFS partition.
  • In production environments, protect the vCenter Server system and the vSphere Auto Deploy server by including them in a highly available environment. Having the vCenter Server in a management cluster guarantees that VDS and virtual machine migration are available. If possible, also protect other elements of your infrastructure. See Set Up Highly Available vSphere Auto Deploy Infrastructure.
  • Set up your environment for vSphere Auto Deploy. See Preparing for vSphere Auto Deploy.
  • Verify that a disk with at least 4GB of free space is available. If the disk is not yet partitioned, partitioning happens when you apply the host profile.
  • Set up the host to attempt a network boot first and to boot from disk if network boot fails. See your hardware vendor's documentation.
  • Create a host profile. See the Host Profiles documentation.

Procedure

  1. Navigate to Home > Policies and Profiles > Host Profiles.
  2. Click the host profile you want to configure and select the Configure tab.
  3. Click Edit Host Profile.
  4. On the Edit host profile page of the wizard, select Advanced Configuration Settings > System Image Cache Configuration > System Image Cache Configuration.
  5. In the System Image Cache Profile Settings drop-down menu, choose a policy option.
    Option Description
    Enable stateless caching on the host Caches the image to disk.
    Enable stateless caching to a USB disk on the host Caches the image to a USB disk attached to the host.
  6. (Optional) If you select Enable stateless caching on the host, specify the information about the disk to use.
    Option Description
    Arguments for first disk
    When configuring a System Image Install disk, you have multiple options to define the device you want ESXi to be installed to and booted from. You can use the following arguments to define the disk for the installation:
    • localesx – The first disk detected containing a valid installation of ESXi
    • local – The first local disk detected by ESXi after boot
    • remoteesx - The first remote disk detected containing a valid installation of ESXi
    • sortedremoteesx- The first remote disk sorted by the lowest LUN ID detected containing a valid installation of ESXi
    • remote - The first remote disk detected by ESXi after boot
    • sortedremote - The first remote disk sorted by the lowest LUN ID detected by ESXi after boot
    • device model
    • device vendor
    • vmkernel device driver name

    You can get the values for the device model and vendor arguments by running the command esxcli storage core device list in a console to the ESXi host, logging in as the root. You get the vmkernel device driver name argument by running the command esxcli storage core adapter list. You must then identify the storage adapter to which your boot device is connected.

    By default, the system attempts to replace an existing ESXi installation, and then attempts to write to the local disk.

    You can use the Arguments for first disk field to specify a comma-separated list of disks to use, in order of preference. You can specify more than one disk. Use localesx for the first disk with ESX installed on it, model and vendor information, or specify the name of the vmkernel device driver. For example, to have the system first look for a disk with the model name ST3120814A, second for any disk that uses the mptsas driver, and third for the local disk, specify ST3120814A,mptsas,local as the value of this field.

    The first disk setting in the host profile specifies the search order for determining which disk to use for the cache. The search order is specified as a comma delimited list of values. The default setting localesx,local specifies that vSphere Auto Deploy should first look for an existing local cache disk. The cache disk is identified as a disk with an existing ESXi software image. If vSphere Auto Deploy cannot find an existing cache disk, it searches for an available local disk device. When searching for an available disk vSphere Auto Deploy uses the first empty disk that does not have an existing VMFS partition.

    You can use the first disk argument only to specify the search order. You cannot explicitly specify a disk. For example, you cannot specify a specific LUN on a SAN.

    Check to overwrite any VMFS volumes on the selected disk If you select this check box, the system overwrites existing VMFS volumes if not enough space is available to store the image, image profile, and host profile.
    Check to ignore any SSD devices connected to the host If you select this check box, the system ignores any existing SSD devices and does not store image profiles and host profiles on them.
  7. Click Save to complete the host profile configuration.

What to do next

Apply the host profile to individual hosts by using the Host Profiles feature in the vSphere Client. See the Host Profiles documentation. Alternatively, you can create a rule to assign the host profile to hosts with the vSphere Client or by using PowerCLI. See Write a Rule and Assign a Host Profile to Hosts.

  • Create a rule that applies the host profile to all hosts that you want to provision with the settings specified in the reference host. For writing a rule in a PowerCLI session, see Write a Rule and Assign a Host Profile to Hosts.
  • For hosts that are already provisioned with vSphere Auto Deploy, perform the test and repair compliance operations in a PowerCLI session, see Test and Repair Rule Compliance.
  • Power on unprovisioned hosts to provision them with the new host profile.