Creating Identity and Resource Pools

0
110
6 min read

(For more resources related to this topic, see here.)

Computers and their various peripherals have some unique identities such as Universally Unique Identifiers (UUIDs), Media Access Control (MAC) addresses of Network Interface Cards (NICs), World Wide Node Numbers (WWNNs) for Host Bus Adapters (HBAs), and others. These identities are used to uniquely identify a computer system in a network. For traditional computers and peripherals, these identities were burned into the hardware and, hence, couldn’t be altered easily. Operating systems and some applications rely on these identities and may fail if these identities are changed. In case of a full computer system failure or failure of a computer peripheral with unique identity, administrators have to follow cumbersome firmware upgrade procedures to replicate the identities of the failed components on the replacement components.

The Unified Computing System (UCS) platform introduced the idea of creating identity and resource pools to abstract the compute node identities from the UCS Manager (UCSM) instead of using the hardware burned-in identities. In this article, we’ll discuss the different pools you can create during UCS deployments and server provisioning. We’ll start by looking at what pools are and then discuss the different types of pools and show how to configure each of them.

Understanding identity and resource pools

The salient feature of the Cisco UCS platform is stateless computing . In the Cisco UCS platform, none of the computer peripherals consume the hardware burned-in identities. Rather, all the unique characteristics are extracted from identity and resource pools, which reside on the Fabric Interconnects (FIs) and are managed using UCSM. These resource and identity pools are defined in an XML format, which makes them extremely portable and easily modifiable. UCS computers and peripherals extract these identities from UCSM in the form of a service profile. A service profile has all the server identities including UUIDs, MACs, WWNNs, firmware versions, BIOS settings, and other server settings. A service profile is associated with the physical server using customized Linux OS that assigns all the settings in a service profile to the physical server. In case of server failure, the failed server needs to be removed and the replacement server has to be associated with the existing service profile of the failed server. In this service profile association process, the new server will automatically pick up all the identities of the failed server, and the operating system or applications dependent upon these identities will not observe any change in the hardware. In case of peripheral failure, the replacement peripheral will automatically acquire the identities of the failed component. This greatly improves the time required to recover a system in case of a failure.

Using service profiles with the identity and resource pools also greatly improves the server provisioning effort. A service profile with all the settings can be prepared in advance while an administrator is waiting for the delivery of the physical server. The administrator can create service profile templates that can be used to create hundreds of service profiles; these profiles can be associated with the physical servers with the same hardware specifications. Creating a server template is highly recommended as this greatly reduces the time for server provisioning. This is because a template can be created once and used for any number of physical servers with the same hardware.

Server identity and resource pools are created using the UCSM. In order to better organize, it is possible to define as many pools as are needed in each category. Keep in mind that each defined resource will consume space in the UCSM database. It is, therefore, a best practice to create identity and resource pool ranges based on the current and near-future assessments.

For larger deployments, it is best practice to define a hierarchy of resources in the UCSM based on geographical, departmental, or other criteria; for example, a hierarchy can be defined based on different departments. This hierarchy is defined as an organization, and the resource pools can be created for each organizational unit. In the UCSM, the main organization unit is root, and further suborganizations can be defined under this organization. The only consideration to be kept in mind is that pools defined under one organizational unit can’t be migrated to other organizational units unless they are deleted first and then created again where required.

The following diagram shows how identity and resource pools provide unique features to a stateless blade server and components such as the mezzanine card:

Learning to create a UUID pool

UUID is a 128-bit number assigned to every compute node on a network to identify the compute node globally. UUID is denoted as 32 hexadecimal numbers. In the Cisco UCSM, a server UUID can be generated using the UUID suffix pool. The UCSM software generates a unique prefix to ensure that the generated compute node UUID is unique.

Operating systems including hypervisors and some applications may leverage UUID number binding. The UUIDs generated with a resource pool are portable. In case of a catastrophic failure of the compute node, the pooled UUID assigned through a service profile can be easily transferred to a replacement compute node without going through complex firmware upgrades.

Following are the steps to create UUIDs for the blade servers:

  1. Log in to the UCSM screen.
  2. Click on the Servers tab in the navigation pane.
  3. Click on the Pools tab and expand root.
  4. Right-click on UUID Suffix Pools and click on Create UUID Suffix Pool as shown in the following screenshot:

  5. In the pop-up window, assign the Name and Description values to the UUID pool.
  6. Leave the Prefix value as Derived to make sure that UCSM makes the prefix unique.
  7. The selection of Assignment Order as Default is random. Select Sequential to assign the UUID sequentially.
  8. Click on Next as shown in the following screenshot:

  9. Click on Add in the next screen.
  10. In the pop-up window, change the value for Size to create a desired number of UUIDs.
  11. Click on OK and then on Finish in the previous screen as shown in the following screenshot:

  12. In order to verify the UUID suffix pool, click on the UUID Suffix Pools tab in the navigation pane and then on the UUID Suffixes tab in the work pane as shown in the following screenshot:

Learning to create a MAC pool

MAC is a 48-bit address assigned to the network interface for communication in the physical network. MAC address pools make server provisioning easier by providing scalable NIC configurations before the actual deployment.

Following are the steps to create MAC pools:

  1. Log in to the UCSM screen.
  2. Click on the LAN tab in the navigation pane.
  3. Click on the Pools tab and expand root.
  4. Right-click on MAC Pools and click on Create MAC Pool as shown in the following screenshot:

  5. In the pop-up window, assign the Name and Description values to the MAC pool.
  6. The selection of Default as the Assignment Order value is random. Select Sequential to assign the MAC addresses sequentially.
  7. Click on Next as shown in the following screenshot:

  8. Click on Add in the next screen.
  9. In the pop-up window, change Size to create the desired number of MAC addresses.
  10. Click on OK and then on Finish in the previous screen as shown in the following screenshot:

  11. In order to verify the MAC pool, click on the MAC Pools tab in the navigation pane and then on the MAC Addresses tab in the work pane as shown in the following screenshot:

LEAVE A REPLY

Please enter your comment!
Please enter your name here