In this chapter, we will walk through the different types of failure scenarios that can occur in Cisco UCS. UCS solution components have excellent redundancy for critical equipment such as chassis and Fabric Interconnects. However, in an unexpected situation such as a physical component's failure, we should be able to identify the failed component and possibly conduct some troubleshooting before contacting Cisco TAC. The most common equipment that fails for the UCS are chassis/Fabric Interconnect power supplies, fan units, IOMs and SFPs for both IOMs, and Fabric Interconnect ports. If proper failover is configured for the network adapters (vNICs) and proper connectivity is configured for the storage adapters (vHBAs), the majority of single component failures do not result in data or management traffic disruption.
UCS failures may also be related to configuration issues, firmware mismatches, temperature issues due to air flow obstruction, and physical cabling issues. In this chapter, we will look into how we can identify these issues from the UCS Manager GUI and also look into LEDs on UCS components.
The following are the topics that will be covered in this chapter:
- Port channel uplink failure and recovery on Fabric Interconnect
- Server link to Fabric Interconnect failure and recovery
- FEX IO modules—failure and recovery
- Fabric Interconnect device failure and recovery
- UCS chassis failure, reporting, and recovery
- Single Fibre Channel failure and recovery on Fabric Interconnects
- Indicating a status with Beacon LEDs
- Creating a tech-support file
In order to dive deep into various troubleshooting scenarios, we will first look at the network and storage connectivity from the mezzanine adapter to the northbound LAN or switch. Different types of ports are involved in this connectivity. They are as follows:
- IOM backplane ports: IOM backplane ports connect to the southbound server mezzanine card through midplane traces. Second and third-generation 2200/2300 series IOM FEXs can provide a maximum of 32 backplane ports.
- IOM Fabric ports: IOM Fabric ports connect to Fabric Interconnect for northbound connectivity. Second-generation 2200 series IOM FEX have four/eight ports 10G each for 2204 and 2208 respectively, whereas third-generation 2300 series IOM FEX have four ports 40G each for 2304.
- Fabric Interconnect server ports: UCS Fabric Interconnect 6332-16UP has a first 16 unified ports. The unified ports and remaining non-unified ports can be configured as server ports to provide southbound connectivity to the UCS chassis IOM module fabric ports.
- Fabric Interconnect uplink ports: UCS ports configured as uplink ports provide northbound connectivity to upstream network switches.
- Fabric Interconnect FC: UCS Fabric Interconnect 6332-16UP includes first 16 unified ports, that can be configured as Fibre Channel ports to provide SAN connectivity. Port reconfiguration between Ethernet and FC requires a reboot of the FI or expansion module.
- Fabric Interconnect FCoE ports: UCS unified Ethernet ports can also be configured as Fibre Channel over Ethernet (FCoE) ports for SAN connectivity using the FCoE protocol.
- Fabric Interconnect Appliance ports: UCS unified Ethernet ports can also be configured as Appliance ports for directly connecting storage appliance to Fabric Interconnect.
The following diagram shows the different ports for both IOM modules and Fabric Interconnects:
