Abstract: Management of power consumption in large computing clusters is disclosed herein. According to an aspect, a computing device comprising a power manager may be configured to receive, via a communication interface, information associated with the cluster of computing nodes. The power manager of the computing device may also be configured to determine whether a switch is coupled to an inactive computing node of the cluster of computing nodes based on the received information. Further, the power manager of the computing device may be configured to communicate a command to the switch to remove power supplied to a switch port of the switch coupled to the inactive computing node in response to determining that the switch is coupled to an inactive computing node of the cluster of computing nodes.

Abstract: A method and apparatus are provided for detecting and recovering from a damaged signal line in a communication cable. Signals are transmitted over a first plurality of signal lines within the cable from a first end, signals are transmitted over a second plurality of signal lines within the cable from the second end, and the quality of signals received on each of the first plurality of signal lines at the second end of the cable are monitored. A first one of the first plurality of signal lines that is not providing a signal of a predetermined minimum quality is identified, the identified first signal line is automatically switched out for a first spare signal line within the cable, and one of the second plurality of signal lines that is logically associated with the identified first signal line is automatically switched out for a second spare signal line within the cable.

Abstract: A method includes a first optical transmitter generating a first data signal at a first end of a fiber optic cable, wherein the first optical transmitter and a first photodetector are included in a first optical transceiver. The method further includes a second photodetector receiving a second data signal at a second end of the fiber optic cable. The second photodetector and the a second optical transmitter are included in a second optical transceiver, and the second data signal is the result of the first data signal passing from the first optical transmitter through the fiber optic cable to the second photodetector. A bit error rate in the second data signal is determined and, in response to the bit error rate exceeding a setpoint, the second optical transmitter sends a message to the first photodetector. The power output of the first optical transmitter is increased responsive to the message.

Abstract: A system includes an electronic device including a baseboard management controller (BMC), a power supply unit (PSU), and an external connector, wherein the BMC controls a serial communication bus, and the external connector includes AC voltage contacts and serial communication bus contacts. A power distribution unit (PDU) includes multiple outlets, a management entity, a radio frequency identification (RFID) tag reader, and a visual indicator. A power cable includes a first plug end for coupling to the external connector, a second plug end for coupling to one of the outlets, AC voltage wires extending end-to-end, a read/write RFID tag in the second plug end, and communication wires extending from the first plug end to the read/write RFID tag. Coupling the first plug end to the external connector connects the AC voltage and connects the serial communication bus such that the BMC is in serial communication with the read/write RFID tag.

Abstract: A system includes an electronic device including a baseboard management controller (BMC), a power supply unit (PSU), and an external connector, wherein the BMC controls a serial communication bus, and the external connector includes AC voltage contacts and serial communication bus contacts. A power distribution unit (PDU) includes multiple outlets, a management entity, a radio frequency identification (RFID) tag reader, and a visual indicator. A power cable includes a first plug end for coupling to the external connector, a second plug end for coupling to one of the outlets, AC voltage wires extending end-to-end, a read/write RFID tag in the second plug end, and communication wires extending from the first plug end to the read/write RFID tag. Coupling the first plug end to the external connector connects the AC voltage and connects the serial communication bus such that the BMC is in serial communication with the read/write RFID tag.

Abstract: A method includes supplying power to a physical server from a plurality of power supplies, wherein operation of all hardware components of the server requires more power than any one of the power supplies can provide. A plurality of jobs are run on the server while the plurality of power supplies are supplying power to the physical server. The method further comprises identifying an amount of power required by each of the components, and identifying one or more components that are not required by one or more of the jobs that are running on the server. The method detects a loss of power from one of the power supplies and then selectively removes power from hardware components identified as not required so that at least a central processing unit and a memory device can continue running at least one job using power available from the operational power supplies.

Abstract: A method includes, for each of a plurality of configured switch ports, identifying switch port settings being used by the configured switch port and device parameters of an installed device connected to the configured switch port. The method further includes correlating one or more of the device parameters to one or more of the switch port settings. In response to an additional device being connected to an additional switch port, the method obtains device parameters of the additional device and automatically configures the additional switch port to use the one or more of the switch port settings that are correlated to one or more of the device parameters matching the device parameters of the additional device.

Abstract: A method comprising flowing a liquid through supply pipe couplings to a fin tube assembly of an air-to-liquid heat exchanger and through the fin tube assembly to return pipe couplings, wherein the fin tube assembly forms a central air flow pathway. The method further comprises passing air through components within a rack secured to the rear door heat exchanger and through the fin tube assembly and detecting an operating condition within the rack. The method then automatically moves the fin tube assembly from a retracted position to an extended position in response to the operating condition, wherein the retracted position directs substantially all of the air to exit the rack through the central air flow pathway, and wherein the extended position allows the air to exit through the central air flow pathway and also through first and second side air flow pathways.

Abstract: Preventing changes to computing devices in a computing system servicing a critical job, including: identifying, by a job protection module, a critical job executing in the computing system; identifying, by the job protection module, one or more computing devices in the computing system utilized during execution of the critical job; and locking, by the job protection module, each of the one or more computing devices in the computing system utilized during execution of the critical job from undergoing a configuration change during execution of the critical job.

Abstract: Automatically orienting hardware ports in a computing device, including: determining a location of a first hardware port configured to receive a first cable connector of a cable; determining a location of a second hardware port configured to receive a second cable connector of the cable; determining, in dependence upon the location of the first hardware port and the location of the second hardware port, orientation information for at least one of the hardware ports; and transmitting an instruction to position at least one of the hardware ports in accordance with the orientation information.

Abstract: Water is circulated through a cooling system within a compute node to remove heat from a heat-generating component within the compute node. Water leakage from the cooling system is collected into a containment reservoir within the compute node. A rate of the water leakage is measured, a temperature of the compute node is measured, a rate of water evaporation from the containment reservoir is determined based upon the measured temperature, and the rate of water leakage is compared to the rate of water evaporation to determine whether water is accumulating in the containment reservoir. A period of time before the containment reservoir reaches a critical level may also be determined.

Abstract: A rack mountable network switch, the rack mountable network switch comprising: a faceplate; a switch chassis coupled to the faceplate, the switch chassis including a plurality of connector openings on a surface of the switch chassis that is perpendicular to the faceplate; a plurality of moveable plug receptacles coupled to a printed circuit board (‘PCB’) mounted inside the switch chassis; and a mechanical positioning component coupled to the moveable plug receptacles, the mechanical positioning component configured to position the moveable plug receptacles to extrude from the switch chassis via the connector openings when the mechanical positioning component is in a first position, the mechanical positioning component further configured to position the moveable plug receptacles inside the switch chassis when the mechanical positioning component is in a second position.

Abstract: Anticipatory protection of critical jobs in a computing system, including: identifying, by a system management module, a problem computing component in the computing system; identifying, by the system management modules, all proximate computing components in the computing system, wherein each proximate computing component is within a predetermined physical proximity of the problem computing component; determining, by the system management module, whether the proximate computing components are executing one or more critical jobs; and responsive to determining that the proximate computing components are executing one or more critical jobs migrating, by the system management module, the one or more critical jobs to distant computing components in the computing system, wherein each distant computing component is not within the predetermined physical proximity of the problem computing component.

Abstract: Anticipatory protection of critical jobs in a computing system, including: identifying, by a system management module, a problem computing component in the computing system; identifying, by the system management modules, all proximate computing components in the computing system, wherein each proximate computing component is within a predetermined physical proximity of the problem computing component; determining, by the system management module, whether the proximate computing components are executing one or more critical jobs; and responsive to determining that the proximate computing components are executing one or more critical jobs migrating, by the system management module, the one or more critical jobs to distant computing components in the computing system, wherein each distant computing component is not within the predetermined physical proximity of the problem computing component.

Abstract: A computer program product includes computer usable program code for: identifying a plurality of power distribution units (PDUs) disposed in a rack, wherein each PDU receives power from a main power source and includes a circuit breaker; identifying a plurality of devices disposed in the rack, wherein each device receives power from one of the PDUs, and wherein the plurality of devices are selected from server nodes, network switches and external data storage devices; obtaining vital product data from a service processor in each device, wherein the vital product data identifies the device by a model identification code; and powering on, for each of the PDUs, the plurality of devices that are connected to the PDU in a sequence to prevent an inrush current from tripping the circuit breaker within the PDU, wherein the sequence powers on devices in order of ascending commonality of the model identification code.

Abstract: A method of powering on a plurality of devices includes identifying a plurality of power distribution units disposed in a rack, wherein each power distribution units is connected to receive power from a main power source, and wherein each power distribution unit includes a circuit breaker. The method further includes identifying a plurality of devices disposed in the rack, wherein each device is connected to receive power from one of the power distribution units, and wherein the plurality of devices are selected from server nodes, network switches and external data storage devices. For each of the power distribution units, the plurality of devices that are connected to the power distribution unit are powered on in a sequence to prevent an inrush current from tripping the circuit breaker within the power distribution unit. The sequence powers on the devices identified as network switches and external data storage devices prior to powering on the devices identified as server nodes.

Abstract: A method and computer program product for implementing the method, where the method comprises obtaining boot dependencies among a plurality of systems, wherein a boot dependency identifies a dependent system, a service system that provides a service to the dependent system, a provide state of the service system, and a need state of the dependent system that requires the service system to have reached the provide state. The method further comprises obtaining historical measurements of the time periods between states for each of the systems. Then, during a process of booting the plurality of systems, the method initiates boot of each dependent system at a time that is determined, based on the historical measurements, to allow the dependent system to reach the need state no earlier than the time at which the service system is determined, based on the historical measurements, to reach the provide state.

Abstract: A unit of information technology equipment (ITE), such as a compute node or a network switch, comprises a system board in communication with network connectors that selectively connect to a network interconnect, a gravity sensor for detecting a first orientation and a second orientation of the system board, and a basic input/output system (BIOS) in communication with the gravity sensor to receive a signal identifying whether the system board is in the first or second orientation. The ITE further comprises a network ASIC (application specific integrated circuit) that inverts the pinout of transmit and receive signals in the network connectors in response to a command from the BIOS indicating that the system board is in the second orientation. A system may comprise a first ITE in a first orientation that interlocks with a second ITE in a second orientation that is inverted 180 degrees from the first orientation.

Abstract: A computer program product includes computer usable program code embodied on a tangible computer usable storage medium for: identifying a plurality of power distribution units (PDUs) disposed in a rack, wherein each PDU receives power from a main power source, and wherein each PDU includes a circuit breaker; identifying a plurality of devices disposed in the rack, wherein each device receives power from one of the PDUs, and wherein the plurality of devices are selected from server nodes, network switches and external data storage devices; and powering on, for each of the PDU, the plurality of devices that are connected to the PDU in a sequence to prevent an inrush current from tripping the circuit breaker within the PDU, wherein the sequence powers on the devices identified as network switches and external data storage devices prior to powering on the devices identified as server nodes.

Abstract: Management of power consumption in large computing clusters is disclosed herein. According to an aspect, a method includes using a power manager to receive, via a communication interface, information associated with the cluster of computing nodes. The method includes determining whether a switch is coupled to an inactive computing node of the cluster of computing nodes based on the received information. Further, the method includes communicating a command to the switch to remove power supplied to a switch port of the switch coupled to the inactive computing node in response to determining that the switch is coupled to an inactive computing node of the cluster of computing nodes.