Configuring servers using configuration patterns

Server patterns are used to quickly provision or preprovision multiple servers (rack and tower servers and compute nodes) from a single set of defined configuration settings.

Procedure

The following figure illustrates the workflow for configuring managed servers. The solid arrows indicate actions taken by you. The dashed arrows indicate actions that are performed automatically by XClarity Administrator.

1. Create address pools.
   An address pool is a defined set of address ranges. Lenovo XClarity Administrator uses address pools to assign IP and I/O addresses to individual servers when the server patterns are deployed to those servers.

   For more information about creating address pools, see Defining address pools.

2. Create category patterns.
   A category pattern groups together related firmware settings and can be reused in multiple server patterns. You can create patterns for the following firmware categories:

   • System information
   • Management interfaces
   • Devices and I/O ports
   • FC boot targets
   • I/O adapter ports

   For more information about category patterns, see Working with server patterns.

3. Create a server pattern.

   A server pattern represents pre-OS server configurations, including local storage configuration, I/O adapter configuration, boot settings, and other baseboard management controller and UEFI firmware settings. A server pattern is used as an overall pattern to quickly configure multiple servers at one time.

   You can define multiple server patterns to represent different configurations that are used in your data center.

   When defining a server pattern, select category patterns and address pools as needed to construct the desired configuration for a specific group of servers. A category pattern groups together related configuration settings that can be reused by multiple server patterns.

   You can create a server pattern from scratch for Converged, Flex System, NeXtScale, and System x servers to define the desired configuration before the hardware arrives. Or, you can create a server pattern from an existing managed server. When you create a server pattern from an existing server, XClarity Administrator learns category patterns from the selected server.

   For more information about creating server patterns, see Creating a server pattern.

4. Deploy the server pattern.

   You can deploy a server pattern to one or more individual servers or to groups of servers at the same firmware level. For example, you can deploy a server pattern to a chassis so all the compute nodes in that chassis are configured the same. During deployment, XClarity Administrator creates a server profile for each server to which the server pattern was deployed. Each server profile represents the specific configuration for a single server. It inherits settings from the server pattern and also contains server-specific information (such as assigned IP addresses and MAC addresses). Because the server profile inherits settings from the server pattern, if you change the server pattern, changes are automatically updated in the server profile. This way, you can maintain common configurations in one place.

   
   
   

   You can deploy a server pattern to:

   • Existing servers. A server profile is created for each server. The server profile is activated after the associated server is rebooted.
   • Empty bays in an existing chassis. A server profile is created for each empty bay. The server profile that is associated with the empty bay can then be activated after the compute node is physically installed.
   • Placeholder for a chassis that you do not yet have. You can pre-provision compute nodes in a chassis that you do not yet have by defining a placeholder chassis to act as a target for the server pattern before the hardware arrives. The placeholder chassis bundles all of the server profiles that are created for each empty compute-node bay. So, when the hardware arrives, you can assign the server profiles to all compute nodes in the new chassis by deploying the placeholder chassis to the new chassis. Each server profile is activated after the associated compute node is rebooted.

   Note: You can deploy a server pattern to multiple servers; however, multiple patterns cannot be deployed to a single server.

   Attention:

   • Ensure that all target servers are at the same firmware level. When you learn a pattern from a specific server, the pattern contains configuration settings for the versions of firmware that are installed on that server.
   • The settings on a server can become out of compliance with its server profile if settings are changed without using Configuration Patterns or if an issue occurred during deployment, such a firmware issue or an invalid setting. You can determine the compliance status of each server from the Configuration Patterns: Server Profiles page.

   For more information about deploying a server pattern, see Deploying a server pattern to a server and Deploying a placeholder chassis.

5. Modify server configuration by editing the server pattern.

   You use server patterns to control a common configuration from a single place. You no longer update settings directly on servers. Instead, you update category patterns and server patterns, and the changes are automatically deployed to all associated profiles and their servers.

   For more information about editing a server pattern, see Modifying a server pattern.