Skip to main content

Operating-system deployment considerations

Before you attempt to deploy an operating-system image, review the following considerations.

Lenovo XClarity Administrator considerations​

  • Ensure that no jobs are currently running on the target server. To see a list of active jobs, click Monitoring > Jobs.
  • Ensure that the target server does not have a deferred or partially activated server pattern. If a server pattern has been deferred or partially activated on the managed server, you must restart the server to apply all configuration settings. Do not attempt to deploy an operating system to a server with a partially activated server pattern.

    To determine the configuration status of the server, see the Configuration Status field on the Summary page for the managed server (see Viewing the details of a managed server).

  • Ensure that a password for the administrator account that is to be used to deploy the operating system is specified on the Global Settings: Deploy Operating Systems dialog. For more information about setting the password, see Configuring global OS-deployment settings.
  • Ensure that the global default settings are correct for this operating-system deployment (see Configuring global OS-deployment settings).

Managed-device considerations​

  • For information about operating-system deployment limitations for specific devices, see XClarity Administrator Support – Compatibility webpage, click the Compatibility tab, and then click the link for the appropriate device types.
  • Ensure there is no mounted media (such as ISOs) on the target server. Additionally, ensure there are no active Remote Media sessions open to the management controller.
  • Ensure that the timestamp in BIOS is set to the current date and time.

  • For servers with XCC2 that have System Guard enabled and the action set to Prevent OS booting, ensure that System Guard is compliant on the device. If System Guard is not compliant, the devices are prevented from completing the boot process, which causes the OS deployment to fail. To provision these devices, manually respond to the System Guard boot prompt to allow the devices to boot normally.

  • For ThinkSystem and System x servers, ensure that the Legacy BIOS option is disabled. From the BIOS/UEFI (F1) Setup utility, click UEFI Setup > System Settings, and verify that Legacy BIOS is set to Disabled.
  • For Flex System servers, ensure that the chassis is powered on.
  • For Converged, NeXtScale, and System x servers, ensure that a Feature on Demand (FoD) key for remote presence is installed. You can determine whether remote presence is enable, disabled, or not installed on a server from the Servers page (see Viewing the status of a managed server). For more information about FoD keys that are installed on your servers, see Viewing Features on Demand keys.
  • For ThinkSystem servers and ThinkAgile appliances, the XClarity Controller Enterprise feature is required for operating-system deployment. For more information, see Viewing Features on Demand keys.
  • For Converged and ThinkAgile appliances, it is recommended that you do not use XClarity Administrator to perform a bare-metal operating-system deployment.

Operating system considerations​

  • Ensure that you have all applicable operating-system licenses to activate the installed operating systems. You are responsible for obtaining licenses directly from the operating-system manufacturer.
  • Ensure that the operating-system image that you intend to deploy is already loaded in the OS images repository. For information about importing images, see Importing operating-system images.

  • Operating-system images in the XClarity Administrator repository might not be supported only on certain hardware platforms. Only OS-image profiles that are supported by the selected server are listed on the Deploy OS Images page. You can determine whether an operating system is compatible with a specific server from the Lenovo OS Interoperability Guide website.

  • For Windows, you must import a boot file into the OS-images repository before you can deploy a Windows profile. Lenovo bundles the predefined WinPE_64.wim boot file along with a set of device drivers into a single package that can be downloaded from the Lenovo Windows drivers and WinPE Images Repository webpage and then imported into OS-images repository. Because the bundle file contains both device drivers and boot files, you can import the bundle file from the Device Driver or Boot Files tab.

  • For SLES 15 and 15 SP1, you must import both the installer image and the associated package image from the Server OS Support Center webpage. For SLES 15 SP2 or later, you need to import only the Full Installation Media image because the Unified Installer and Packages DVDs from SUSE Linux Enterprise Server 15 and 15 SP1 are deprecated.

  • When deploying Ubuntu with the Virtualization profile, the server host's network must have Internet access to download installation packages from the Ubuntu cloud.
  • For ThinkSystem servers, XClarity Administrator includes out-of-box device drivers to enable the installation of the operating system as well as basic network and storage configuration for the final operating system. For other servers, ensure that the operating-system image that you intend to deploy includes the appropriate Ethernet, Fibre Channel, and storage adapter device drivers for your hardware. If the I/O adapter device driver is not included in the operating system, the adapter is not supported for OS deployment. Always install the latest operating system to ensure that you have the latest inbox I/O adapter device drivers and boot files that you need. You can also add out-of-box device drivers and boot files to operating systems that have been imported into XClarity Administrator (see Customizing OS-image profiles).

    For VMware, use the latest Lenovo Custom Image for ESXi, which includes support for the latest adapters. For information about obtaining that image, see the VMware Support – Downloads webpage.

  • For ThinkSystem servers, if you want to deploy SLES 12 SP2, you must use a kISO profile. To get the kISO profiles, you must import the appropriate SLES kISO image after you import the base SLES operating system. You can download the SLES kISO image from the Linux Support – Downloads webpage.
    Note

    • The SLES kISO image counts towards the maximum number of imported OS images.

      For a list of supported base and custom operating systems, see Supported operating systems.

    • If you delete all kISO profiles, you must delete the base SLES operating system and then import the base operation system and kISO image again to deploy SLES 12 SP2 on a ThinkSystem server.

    • If you create a custom OS-profile based on a kISO profile, the predefined device drivers in the base operating system are not included. The device drivers that are included in the kISO are used instead. You can also add device drivers to the custom OS-profile (see Creating a custom OS-image profile).

For more information about limitations for specific operating systems, see Supported operating systems.

Network considerations​

  • Ensure that all required ports are open (see Port availability for deployed operating systems).

  • Ensure that XClarity Administrator can communicate with the target server (both the baseboard management controller and the servers’ data network) over the interface (Eth0 or Eth1) that was selected when you configured the XClarity Administrator network access.

    To specify an interface to be used for operating-system deployment, see Configuring network access).

    For more information about the operating-system deployment network and interfaces, see Network considerations.

  • Ensure that the IP addresses are unique for the host operating system. XClarity Administrator checks for duplicate IP addresses that you specify for the network address during the deployment process.
  • If the network is slow or unstable, you might see unpredictable results when deploying operating systems.
  • The XClarity Administrator network interface that is used for management must be configured to connect to the baseboard management controller using the same IP address method that you choose on the Global Settings: Deploy Operating Systems dialog. For example, if XClarity Administrator is set up to use eth0 for management, and you choose to use manually assigned static IPv6 addresses when configuring the deployed OS, then eth0 must be configured with an IPv6 address that has connectivity to the baseboard management controller.
  • If you choose to use IPv6 addresses for the OS deployment global settings, the IPv6 address for XClarity Administrator must be routable to the baseboard management controller and the servers’ data network.
  • IPv6 mode is not supported for ThinkServer (see IP configuration limitations).

  • If you are using DHCP to assign IP addresses dynamically, you must configure the MAC address.

  • If you are using static IP addresses, you must configure the following network settings for a specific server before you can deploy an operating system to that server. After these settings are configured, the deployment status of the server changes to Ready. (Note that some fields are not available for static IPv6 addresses.)

    • Hostname

      The hostname must comply with the following rules:

      • The hostname of each managed server must be unique.

      • The hostname can contain strings (labels) that are separated by a period (.).

      • Each label can contain ASCII letters, digits, and dashes (-); however, the string cannot start or end with a dash and cannot contain all digits.

      • The first label can be 2 - 15 characters in length. Subsequent labels can be 2 – 63 characters in length.

      • The total length of the hostname must not exceed 255 characters.

    • MAC address of the port on the host where the operating system is to be installed.

      The MAC address is set to AUTO by default. This setting automatically detects the Ethernet ports that can be configured and used for deployment. The first MAC address (port) that is detected is used by default. If connectivity is detected on a different MAC address, the XClarity Administrator host is automatically restarted to use the newly detected MAC address for deployment.

      You can determine the status of the MAC address port that is used for OS deployment from the MAC address drop-down menu on the Network Settings dialog. If multiple ports are up or if all ports are down, AUTO is used by default.

      Note

      • Virtual network ports are not supported. Do not use one physical network port to simulate multiple virtual network ports.

      • When the server's network setting is set to AUTO, XClarity Administrator can automatically detect network ports in slots 1 – 16. At least one port in slots 1 – 16 must have a connection to XClarity Administrator.

      • If you want to use a network port in slot 17 or greater for the MAC address, you cannot use AUTO. Instead, you must set the server's network setting to the MAC address of the specific port that you want to use.

      • For ThinkServer servers, not all host MAC addresses are displayed. In most cases, MAC addresses for AnyFabric Ethernet adapters are listed on the Edit Network Settings dialog. MAC addresses for other Ethernet adapters (such as the Lan-On-Motherboard) are not listed. In cases where the MAC address for an adapter is not available, use the AUTO method for non-VLAN deployments.

    • IP address and subnet mask
    • IP gateway
    • Up to two domain name system (DNS) servers
    • Maximum transmission unit (MTU) speed
    • VLAN ID, if VLAN IP mode is enabled

  • If you choose to use VLANs, you can assign a VLAN ID to the host network adapter that is being configured.

For more information about the operating-system deployment network and interfaces, see Configuring network settings for managed servers and Network considerations.

Storage and boot-option considerations​

  • Ensure that the UEFI boot option on the target server is set to UEFI boot only before you deploy an operating system. The Legacy-only and UEFI first, then legacy boot options are not supported for operating-system deployment.

  • Each server must have a hardware RAID adapter that is installed and configured.

    Attention

    • Only storage that is set up with hardware RAID is supported.

    • The software RAID that is typically present on the onboard Intel SATA storage adapter or storage that is set up as JBOD are not supported; however, if a hardware RAID adapter is not present, setting the SATA adapter to AHCI SATA mode enabled for operating-system deployment or setting unconfigured good disks to JBOD might work in some cases. For more information, see OS installer cannot find the disk drive on which you want to install.

      This exception does not apply to M.2 drives.

    • If a managed device has both local drives (SATA, SAS, or SSD) that are not configured for hardware RAID and M.2 drives, you must disable the local drives if you want to use M.2 drives, or you must disable the M.2 drives if you want to use local drives. You can disable on-board storage controller devices and legacy and UEFI storage option ROMs using the using Configuration Patterns by selecting Disable local disk on the Local Storage tab of the wizard or by creating a Configuration Pattern from an existing server and then disabling the M.2 devices in the extended UEFI pattern.

    • If a SATA adapter is enabled, the SATA mode must not be set to IDE.

  • The NVMe storage that is connected to a server motherboard or HBA controller is not supported and must not be installed in the device; otherwise, OS deployment to non-NVMe storage will fail.

  • When deploying RHEL, multi ports that are connected to same LUN on the target storage is not supported.

  • Ensure that secure-boot mode is disabled for the server. If you are deploying a secure-boot mode enabled operating system (such as Windows), disable secure-boot mode, deploy the operating system, and then re-enable secure-boot mode.

  • When deploying Microsoft Windows to a server, attached drives must not have existing system partitions (see OS deployment fails due to existing system partitions on an attached disk drive).

  • For ThinkServer servers, ensure that the following requirements are met:

    • The boot settings on the server must include a Storage OpROM Policy that is set to UEFI Only. For more information, see OS installer cannot boot on a ThinkServer server.

    • If you are deploying ESXi and there are network adapters that are PXE bootable, disable PXE support on the network adapters before deploying the operating system. The deployment is completed, you can re-enable PXE support, if desired.

    • If you are deploying ESXi and there are bootable devices in the boot-order list other than the drive on which the operating system is to be installed, remove the bootable devices from the boot-order list before deploying the operating system. After deployment is complete, you can add the bootable device back to the list. Ensure that the installed drive is at the top of the list.

For more information about storage-location settings, see Choosing the storage location for managed servers.