Cisco UCS Manager Overview
In the Cisco UCS, the Cisco UCS Manager (UCSM) is the application that takes care of the discovery of the physical components of the system, creates and applies the configuration, and monitors and reports on any changes and issues. It runs on the two Fabric Interconnects that form the Cisco UCS cluster. A single UCS cluster is under the management of one Cisco UCSM, and this combination is called a Cisco UCS domain. To be more precise, a Cisco UCS domain represents a pair of Cisco UCS Fabric Interconnects, connected through the cluster link to create a Cisco UCS cluster, under the management of a single Cisco UCSM. In a single Cisco UCS domain, a Cisco UCSM instance can manage up to 160 B- and C-series servers.
The two Fabric Interconnects in a Cisco UCS cluster operate in a high-availability and redundant mode. When it comes to the data communication of the servers, or the data plane, the two Fabric Interconnects are active (that is, they both process the data communication or provide active data paths). At the level of the management and control planes, they operate in an active-standby manner. This means that after the Fabric Interconnects have booted up, there is a negotiation between the two, based on which is determined to be the primary one. On the primary Fabric Interconnect, the management and control applications and services run in active mode; on the other Fabric Interconnect (the subordinate), they run in standby mode. The active instances will be distributed between the two for load balancing. The cluster link is used to synchronize configuration and state information between the active and standby instances of the applications and services. But what are these applications and services that run on the Fabric Interconnects?
The following software resides and runs on the Fabric Interconnects:
- Cisco Nexus OS (NX-OS): NX-OS consists of two files:
- Kickstart image file: This is the Linux kernel of the operating system.
- System image file: This file contains all the applicable modules of the Nexus OS for the hardware platform on which it is used. Let’s not forget that Cisco uses the Nexus OS on a variety of devices—the Cisco Nexus and MDS switches as well as on the Cisco UCS Fabric Interconnects.
- Cisco UCS Manager application- an XML application that manages the Cisco UCS.
So far, you know that a Cisco UCS domain consists of two Fabric Interconnects running in a redundant mode, also known as a Cisco UCS cluster. On both Fabric Interconnects the NX-OS is loaded and the services needed for the Cisco UCS are up and running in a load-balancing and redundant mode. This means that there will be a single active instance for each service and one standby. Which service will be active on which Fabric Interconnect, regardless of whether it’s the primary or the subordinate, is decided during the negotiations at the start. The primary Fabric Interconnect is the one on which the Cisco UCSM instance runs in active mode. The other is the subordinate Fabric Interconnect because on that one the Cisco UCSM application runs in standby (passive) mode. It will receive configuration and state updates from the primary Fabric Interconnect and then update itself. The following points are important to remember:
- The primary/subordinate mode of operation of the Fabric Interconnects applies only to the management and control planes and does not affect the data plane operation, which is always active through both Fabric Interconnects.
- The primary/standby mode of the Fabric Interconnects is defined by the active instance of the Cisco UCSM application. The rest of the processes, which also run in active/standby mode, are distributed between the two Fabric Interconnects for better utilization of the resources and for load-balancing their consumption. This means that on the subordinate Fabric Interconnect will be processes running in active mode, and their standby instances will be on the active Fabric Interconnect.
Figure 17-1 illustrates the Cisco UCS management components.

Figure 17-1 Cisco UCS Primary/Subordinate Cluster with Nexus OS and the Cisco UCSM Application
The Cisco UCSM is an XML application. It is based on an XML scheme that represents the hierarchical structure of the logical and physical components of the Cisco UCS. This allows for each of these components to be addressed and for the related configuration and information to be applied to them. All the configuration and state information is stored in the Cisco UCSM database in XML format. The physical components under Cisco UCSM management are called managed endpoints, and they include the following:
- The switch elements: The Fabric Interconnects, the Cisco FEXs, the switch modules and ports for the Fabric Interconnects
- The chassis elements: The chassis management controller (CMC), the chassis I/O modules (IOMs), and the chassis power and fan modules.
- Managed server endpoints: The servers and their components, such as the disks, mezzanine adapters, the BIOS, the Cisco Integrated Management Controller (CIMC), and so on
Figure 17-2 illustrates the Cisco UCS Manager interfaces.

Figure 17-2 Cisco UCSM Application Interfaces
The Cisco UCSM, which is the central point for the management and configuration of the whole Cisco UCS domain, sits between the administrator and the server system. The administrator, or other applications and systems, can communicate with the Cisco UCSM through the available management interfaces:
- Cisco GUI: The graphical user interface initially was based on Java, and an additional client application was used to access it, but with the latest versions of the Cisco UCSM, starting with major version 4, the GUI is based on HTML5 and only a browser is needed.
- Cisco CLI: The command line interface to access the Cisco UCSM
- Third-party access: Using the Cisco UCS XML application programming interface (API), third-party applications can connect to the system.
All the management access—whether you are using the GUI or the CLI, or it’s an application trying to communicate—goes through the Cisco UCSM XML API. This means your actions will be converted into XML and sent to the appropriate managed endpoint as a configuration request, and the operational state will be communicated back.
The operational state communicates information for the managed endpoint. This consists of not only the state of the component but also monitoring and reporting information. The Cisco UCSM supports widely used industry-standard protocols such as the following:
- Serial over LAN
- KVM over IP
- SNMP
- SMASH CLP
- CIM XML
- IPMI
It also supports the following Cisco features for monitoring and reporting:
- Call Home: This feature automates the notifications and even creates Cisco TAC cases based on physical issues occurring within the system.
- Cisco UCSM XML API: With the XML API, the Cisco UCSM can be integrated with third-party monitoring and reporting tools.
- Cisco GUI and CLI: The monitoring and reporting information is available in the GUI and CLI of the UCSM as well.
The Cisco UCSM GUI, shown in Figure 17-3, consists of two major elements: the navigational pane, where you traverse the XML hierarchical tree, and the content pane. Once you select a component, either physical or logical, you will see in the content pane to the right all the related information and options.

Figure 17-3 Cisco UCSM GUI
Additionally, the buttons in the top-right corner allow you to access the quick links, help, administrative session properties, and “about” information as well as to exit.
Another convenient feature, located at the middle top of the GUI, is quick information regarding the critical, major, and minor faults and warning messages (see Figure 17-4). By clicking one of these icons, you will be taken to the appropriate page containing information about what is happening with the system.

Figure 17-4 Cisco UCSM GUI Faults
The navigational pane is divided into eight tabs. The Equipment tab, shown in Figure 17-5, is where you can access all the physical components of the Cisco UCS. This is where you can get information for and access to the Fabric Interconnects, the chassis and its components, and the servers and their components.

Figure 17-5 Cisco UCSM Equipment Tab
The Equipment tab points only to physical components and policies related to the hardware discovery, power redundancy, fan control, and so on. You can access the topology view for the UCS domain in a separate tab for the Fabric Interconnects, the B- and C-series servers’ hardware, the firmware management policies, thermal information, decommissioned servers, and so on. The information you can gather from the Equipment tab is important for understanding the physical connectivity and the resources in the UCS domain. The Equipment tab is where you can monitor the discovery process and configure policies related to it.
The UCS Manager is responsible for managing, pushing configurations to, and monitoring the hardware components of a Cisco UCS domain. To be capable of doing this, the UCS Manager needs to know what is connected and how. The process of learning and acquiring this information is called discovery. In the discovery process, the UCS Manager, which runs on the Fabric Interconnects, communicates with the control and management components from the underlying hardware. These are the Chassis Management Controllers (CMC) on the IOMs, the Cisco Integrated Management Controllers (CIMC) on the B- and C-series servers, and, in the case of C-series servers’ integration through Cisco Nexus 2000 FEXs, their CMCs as well. The whole process of discovery starts when the UCS Manager finds Fabric Interconnect ports that are configured as server ports. This means that servers are connected to these ports. The UCS Manager establishes communication with the CMCs from the IOMs. The CMCs, as they are the management and control components of the IOMs in the server chassis, have already gathered the following information for the hardware in the chassis:
- Thermal sensors
- Fans status
- Power supplies (status and temperature)
- Servers
For the servers’ discovery, the CMCs of the IOMs connect through dedicated management interfaces in the mid-plane of the chassis to the CIMCs of the servers. The CIMCs communicate all the information for the hardware of the server and its status to the CMC. This means that when the UCS Manager starts to communicate with the CMCs, they already have all the information for the hardware in the chassis, including the servers. Thus, the UCS Manager can build a topology and gather the information for the available hardware resources and their connectivity.
Here are some of the available hardware-related policies:
- Global policies include the following (see Figure 17-6):

Figure 17-6 Cisco UCSM Equipment Tab – Global Policies
- Chassis/FEX Discovery Policy: Defines the minimum number of links that must exist between a Fabric Interconnect and an FEX in order for the FEX and the equipment behind it to be discovered.
- Rack Server Discovery Policy: Discovery of the C-series servers.
- Rack Management Communication Policy: How the changes are to be applied to the C-series server—either immediately or after acknowledgment from the administrator.
- Power Policy: Defines the power redundancy. Depending on the setting, there might be requirements for the number of power supplies in the chassis.
- Fan Control Policy: Controls the fan speed to optimize power consumption and noise levels.
- MAC Address Table Aging: Configures how long a MAC address will remain in the MAC address table.
- Global Power Allocation Policy: Specifies the power capping policies (how much power is allowed per server) to apply to the servers in a chassis.
- Info Policy: Enables the information policy to display the uplink switches to which the Fabric Interconnects are connected. By default, this is disabled.
- Global Power Profiling Policy: Defines how the power cap values for the servers are calculated.
- Hardware Change Discovery Policy: Any change in the server hardware component will raise a critical “hardware inventory mismatch” fault.
- Autoconfig Policies: Used to automate the deployment of the newly discovered servers in production. If such a policy exists, the new server will be checked against a set of requirements (a qualification policy). If the server’s hardware matches the qualification policy, it will be associated with a service profile and can be used immediately.
- Server Inheritance Policies: Used to automate the generation of a service profile based on the hardware of the server. In this situation, the hardware identity values of the server are used, and the service profile can be used only with this server.
- Server Discovery Policies: These define the behavior of the UCS Manager when a new server is discovered.
- SEL Policy: A configuration that allows you to export the system event logs from the hardware.
The next tab is the Servers tab, shown in Figure 17-7. The Equipment tab is the only one that shows information for the physical equipment. The Servers tab provides all the logical configuration and state information related to the servers. Here is where you can create a service profile or look at the available one. Also, you can create service profile templates and all the policies related to the servers, such as the boot policy or the BIOS policy. In the Cisco UCS Manager, anything created for a specific aspect of the server configuration is called a policy. That’s why in the Servers tab there is a separate section for all the policies that can be created and used in a service profile. These include not only the forementioned BIOS and boot policies, but also maintenance, memory, scrub, power, adapter, and many more policies. In general, the Cisco UCS Manager is so granular that you can create a policy for any specific piece of configuration you can think of. This allows for extreme flexibility when it comes to the configuration of the servers in your data center. At first glance this might not look very important, but when you start solving the challenges in a real data center, you will appreciate this flexibility.

Figure 17-7 Cisco UCSM Servers Tab
The LAN tab, shown in Figure 17-8, is where you perform all the configurations related to the network connectivity of the Cisco UCS as a whole and the network connectivity of the servers in particular. Here, under the LAN Cloud option, you create the VLANs in which all the server communication will occur, and you configure the communication of directly attached network devices under the Appliances section. In this tab, you also create the specific configuration for the servers’ communication.

Figure 17-8 Cisco UCSM LAN Tab
The fourth tab is the SAN tab (see Figure 17-9). In the SAN tab, you create all the logical configuration related to the storage communication of the servers. This is about the Fibre Channel–based communication of the servers. All of the needed policies, identity pools, VSANs, and HBA adapters’ configuration is created and managed in this tab—both for Cisco UCS communication with the data center SAN and with directly attached FC storage appliances.

Figure 17-9 Cisco UCSM SAN Tab
The VM tab, shown in Figure 17-10, allows for the integration of the Cisco UCS Manager with the VMware and Microsoft hypervisors to reach up to the level of the virtual machines.

Figure 17-10 Cisco UCSM VM Tab
The Storage tab, shown in Figure 17-11, allows for the creation of local storage profiles and policies for the provisioning of the local storage.

Figure 17-11 Cisco UCSM Storage Tab
The Chassis tab, shown in Figure 17-12, is a new one. It allows the DC administrators to focus on tasks specific to the management of multiple chassis in a Cisco UCS domain. In one place, you can create and manage all the needed policies for the chassis of the blade servers.

Figure 17-12 Cisco UCSM Chassis Tab
The Admin tab, shown in Figure 17-13, is where all the configuration related to the administration of, monitoring of, and access to the Cisco UCS is configured. Here you can access the faults, events, and audit logs. You can also configure the authentication and authorization, which can be based on creating a database of local users or using external authentication, authorization, and accounting (AAA) servers such as a Microsoft Windows Active Directory (AD).

Figure 17-13 Cisco UCSM Admin Tab