TW201729097A - Rack - Google Patents

Abstract

A rack includes a control system including a plurality of bus, a plurality of baseboard management controllers (BMC), and a control unit. Each BMC generates a heartbeat signal and saves operating data to the control unit after power on and determines to operate at a master mode or a slave mode based on a represented deciding signal. The control unit generates the represented deciding signal based on the heartbeat signals so as to control the first BMC operating normally to operate at the master mode and control the other BMCs to operate at the slave mode. The BMC operating at the master mode reads the operating data saved in the control unit via the represented bus so as to monitor and control a monitored unit.

Description

Cabinet unit

The present invention relates to a cabinet apparatus, and more particularly to a cabinet apparatus having a monitoring system.

The conventional rack device (Rack) includes a rack management controller (Rack/Chassis Management Controller; RMC/CMC), at least one rack backplane (RBP), and at least one tray backplane (Tray Back Plane; TBP). ), at least one node (Node), a fan backplane (Fan Board; FB), and a power distribution board (PDB). Each cabinet backplane is electrically connected to at least one tray backplane, and each tray backplane is electrically connected to at least one node, each node is a server node and includes a Baseboard Management Controller (BMC), the fan The backplane includes at least one fan, and the power distribution board includes at least one power supply unit (PSU).

Referring to FIG. 1, the cabinet includes a cabinet management controller 9, a cabinet backplane 91, a power distribution board 92, three power modules 921, 922, and 923, two tray back boards 93, 94, and a fan. Backplane 95, five nodes 931, 932, 941, 942, 943, five board management controllers 933, 934, 944, 945, 946, three fans 951, 952, 953. The substrate management controllers 933, 934, 944, 945, and 946 are electrically connected to the fans 951, 952, and 953 via the tray back plates 93 and 94 and the fan back plate 95, respectively, to detect the corresponding fans. The operating conditions of 951, 952, and 953, such as the temperature and rotation speed of each fan. The rack management controller 9 is electrically connected to the baseboard management boards 933, 934, 944, 945, and 946 via the cabinet backplane 91 and the tray backplanes 93, 94 to obtain and monitor the fans 951, 952, The working condition of 953. The cabinet management controller 9 is further electrically connected to the fans 951, 952, and 953 via the cabinet backplane 91 and the fan backplane 95, and controls the fans 951 according to the operating conditions of the fans 951, 952, and 953. The rotational speeds of 952 and 953 are used to determine the heat dissipation capability of the fans 951, 952, and 953 to adjust the temperature in the cabinet. However, in this conventional cabinet design, when the rack management controller 9 is implemented with one wafer, the cabinet management controller 9 needs to have sufficient communication interface and sensing interface, and must have sufficient calculation. Capabilities, such as the IPMB (Intelligent Platform Management Bus) using the Intelligent Platform Management Interface (IPMI), etc., require a complex handshake mechanism, resulting in a complex wafer of the rack management controller 9. The level of degrees is often equivalent to the control wafers of the substrate management controllers 933, 934, 944, 945, 946.

Accordingly, it is an object of the present invention to provide a low complexity and low cost cabinet unit.

Thus, the cabinet apparatus of the present invention includes a monitored unit and a monitoring system. The monitoring system includes a plurality of bus bars, a plurality of substrate management controllers, and a control unit.

The baseboard management controllers are electrically connected to the busbars and the monitored units, respectively, and each of the operational data is generated after the startup is started, and the operational data is respectively transmitted through the busbars, and each of the devices is also generated. The heartbeat signals are each determined to operate between a master mode and a slave mode according to a corresponding decision signal.

The control unit is electrically connected to the busbars to receive and store the operational data from the baseboard management controllers, and is also electrically connected to the baseboard management controllers to receive the baseboard management controllers from the baseboard management controllers. Waiting for the heartbeat signal, and generating the corresponding determination signals according to at least the heartbeat signals, to control the first normal operators of the substrate management controllers to operate in the master control mode, and control the baseboard management controllers The rest of the operations operate in this slave mode.

The baseboard management controller operating in the master mode reads the operational data stored by the control unit via the corresponding busbar to monitor the monitored unit.

In some implementations, the control unit of the monitoring system further includes a memory, the memory including a plurality of memory blocks. Each of the baseboard management controllers of the monitoring system further receives a corresponding identification signal, and stores the generated operational data via the corresponding busbar to the memory of the control unit according to the corresponding identification signal. One of the corresponding blocks of the memory block.

In some embodiments, the cabinet device further includes a plurality of boards, and a plurality of identification pins respectively disposed on the boards. The identification pins determine the logical values of the identification signals. The baseboard management controllers of the monitoring system are respectively disposed on the board, and each of the baseboard management controllers according to the logic value of the identification signal corresponding to the identification pin disposed on the same board The generated operational data is stored via the corresponding busbar to the corresponding one of the memory blocks of the memory of the control unit.

In some implementations, wherein each of the baseboard management controllers of the monitoring system generates a presence signal and generates a jump between a first logic value and a second logic value when the power is turned on and operating normally. The heartbeat signal. The control unit of the monitoring system further determines whether the corresponding determination signal is generated based on the heartbeat signal of the person based on the presence signal from any of the substrate management controllers.

When the logic value of the existence signal of the person is a third logic value, and the heartbeat signal of the first one of the substrate management controllers is between the first logic value and the second logic value When jumping, the logical value of the corresponding decision signal is changed from a fourth logic value to a fifth logic value. When the logical value of the presence signal of the person is a sixth logic value, the logical value of the corresponding determination signal is the fourth logic value.

In some implementations, the baseboard management controller operating in the master mode operates as a rack management controller (Rack/Chassis Management Controller; RMC/CMC) to monitor the monitored unit.

In some implementations, wherein the monitored unit comprises a plurality of fan units. Each of the substrate management controllers of the monitoring system is electrically coupled to one of the fan units. Each of the substrate management controllers detects the electrically connected fan unit during normal operation to obtain the operational data associated with the corresponding fan unit. The substrate management controller operating in the master mode reads the memory stored in the control unit via the corresponding bus bar, and operates the substrate management controllers in the slave mode to detect the Operational information to monitor and control the fan units.

In some implementations, the control unit of the monitoring system further generates a plurality of data flags according to the operational data, and further stores the data flags in the memory, the data flags Indicate whether these operational data are updated regularly.

In some implementations, wherein the busbars of the monitoring system support an intra-integrated circuit (I2C) protocol.

In some implementations, wherein one of the first logic value and the second logic value is logic 1 and logic 0, and the fourth logic value and the fifth logic value are respectively The other and the other are logic 1 and logic 0, and one of the third logic value and the sixth logic value and the other are logic 1 and logic 0, respectively.

The function of the present invention is that the control unit stores the operation data detected by the substrate management controller, that is, the control unit is regarded as a memory (such as an EEPROM), so that the substrate management controller and Communication between the control units requires only simple data access, without the need for complex handshake mechanisms such as IPMI or IPMB implementations as in the prior art. In addition, each of the substrate management controllers respectively uploads the operation data to the control unit, so that the substrate management controller as the cabinet management controller can obtain the operation data detected by the other substrate management controllers at any time, and The complexity of the overall design can be significantly reduced compared to prior art.

Before the present invention is described in detail, it should be noted that in the following description, similar elements are denoted by the same reference numerals.

Referring to FIG. 2, an embodiment of the rack device (Rack) of the present invention is, for example, a server rack (Server Rack), and includes a cabinet shell 7, a plurality of fan units 51-54 disposed in the cabinet shell 7, A plurality of boards 41 to 44, a plurality of identification pins 61 to 64, and a monitoring system 1. The monitoring system includes a plurality of bus bars 31-34, a plurality of baseboard management controllers (BMC) 21-24, and a control unit 11. Each of the boards 41 to 44 is regarded as one node, and each of the fan units 51 to 54 includes a plurality of fans. The identification pins 61 to 64 and the substrate management controllers 21 to 24 are respectively disposed one-to-one on the boards 41 to 44, and are electrically connected to the fan units 51 to 54, respectively. The substrate management controllers 21 to 24 are electrically connected to the control unit 1 to respectively transmit a plurality of presence signals and a plurality of heartbeat signals to the control unit 1, and respectively receive a plurality of determination signals from the control unit 1. The baseboard management controllers 21-24 are also electrically connected to the control unit 1 via the busbars 31-34, respectively, to store a plurality of operational data to the control unit 1, respectively.

In the present embodiment, for convenience of explanation, FIG. 2 is composed of four substrate management controllers 21 to 24, four bus bars 31 to 34, four fan units 51 to 54, and four boards 41 to 44. An example is given. These bus bars 31 to 34 support an agreement of an Inter-Integrated Circuit (I2C). In other embodiments, the fan units 51-54 and the number of the fans may also be singular or other plural, and the fan units 51-54 may be respectively disposed on the boards 41-44, or the machines. Except for the plates 41~44, this is not the limit.

The control unit 11 includes a memory 110 that includes a plurality of memory blocks for receiving and storing the operational data from the substrate management controllers 21-24, respectively. The control unit 11 generates the corresponding determination signals according to the presence signals and the heartbeat signals, respectively, to control the first normal operators of the substrate management controllers 21-24 to operate in a master mode, for example Is 21 and controls the rest of the baseboard management controllers 21-24 to operate in a slave mode, for example 22~24. In the present embodiment, the control unit 11 is a microcontroller (Micro Control Unit) including a memory 110.

More specifically, the control unit 11 determines whether or not the corresponding determination signal is generated based on the heartbeat signal of the person based on the presence signal from any of the substrate management controllers 21-24. For example, based on the presence signal of the substrate management controller 21, it is determined whether or not the determination signal output to the substrate management controller 21 is generated based on the heartbeat signal of the substrate management controller 21. When the logic value of the presence signal of the person is a third logic value, and the person is the first one of the baseboard management controllers 21-24, the heartbeat signal is at a first logic value and a second logic When the value jumps, the logical value of the corresponding decision signal is changed from a fourth logic value to a fifth logic value. When the logical value of the presence signal of the person is a sixth logic value, the logical value of the corresponding determination signal is maintained at the fourth logic value. In other words, when the logical value of the corresponding presence signal is the sixth logic value, the control unit 11 may not need to monitor the change of the corresponding heartbeat signal, but can reduce the load of the control unit 11. In this embodiment, the third logic value and the fifth logic value are logic 1, and the fourth logic value and the sixth logic value are logic 0. In other embodiments, the fourth logic value and the fifth logic value may also be one of the other and the other one is a logic 1 and a logic 0, and the third logic value and the sixth logic value may also be one of The other and the other are logic 1 and logic 0, respectively.

Each of the baseboard management controllers 21-24 generates a heartbeat (heartbeat) that jumps between the first logic value and the second logic value after the startup is started and operates normally, and detects the electrically connected The fan units 51 to 54 obtain an operation data relating to the corresponding fan units 51 to 54. One of the first logic value and the second logic value is logic 1 and logic 0, respectively, and the frequency of the heartbeat signal is, for example, 0.5 or 1 Hz, but not limited thereto. Conversely, when each of the baseboard management controllers 21-24 is powered on, or is not operating normally after the booting, the logical values of the heartbeat signals generated by the baseboard management controllers 21-24 remain at The first logical value or the second logical value, as in logic zero. The operational data of each of the fan units 51-54 is, for example, information such as the rotational speed, temperature, and the like of the fans of the fan units 51-54.

Each of the substrate management controllers 21 to 24 further receives a corresponding identification signal, and stores the generated operational data to the memory of the control unit 11 via the corresponding bus bars 31-34 according to the corresponding identification signal. One of the memory blocks of the body 110 corresponds to one of the memory blocks. Referring to FIG. 3, in the embodiment, the memory 110 of the control unit 11 includes five memory blocks 111-115, and the four substrate management controllers 21-24 are respectively disposed in the four of the computer system. When the boards 41 to 44 are used, the four board management controllers 21 to 24 respectively determine the logical values of the corresponding four identification signals according to the four sets of identification pins (ID Pins) 61 to 64, that is, each of the substrates. The management controllers 21 to 24 further cause the four substrate management controllers 21 to 24 to be respectively based on the logical values of the identification signals of the corresponding identification pins 61 to 64 provided on the same boards 41 to 44. The operational data is stored in the address of the corresponding four memory blocks 112-115 of the control unit 11.

In this embodiment, when the logic value of the determination signal received by any one of the substrate management controllers 21-24 is the fifth logic value, that is, logic 1, the substrate management controller 21 operates in The master mode. On the other hand, when the logic value of the determination signal received by any one of the substrate management controllers 21-24 is the sixth logic value, that is, logic 0, the baseboard management controllers 22-24 operate in the Slave mode. The baseboard management controller 21 operating in the master mode operates as a rack management controller (RMC/CMC), and reads the stored in the control unit 11 via the corresponding bus bar 31. The memory 110 operates the operational data detected by the substrate management controllers 22-24 of the slave mode to monitor and control the fan units 51-54. In other words, the baseboard management controller 21 operating in the master mode can monitor the operating conditions of all of the fan units 51-54, and also control the fan units 51-54 according to the operational data. The fans, for example, the rotational speeds, determine the heat dissipation capabilities of the fan units 51-54 to adjust the temperature within the cabinet housing 7.

The control unit 11 further generates a plurality of data flags according to the operational data, and further stores the data flags in the memory 110, and the data flags respectively indicate whether the operational materials are regularly updated. More specifically, in order to prevent a certain baseboard management controller (BMC) or cabinet management controller (CMC) from being offline or losing its function, the operational data written by it is still regarded as valid data, so it is written. The front end of each operational data is added to the corresponding data flag for judging whether the operational data is valid. For example, the control unit 11 determines whether each of the substrate management controllers 21 to 24 stores the corresponding operation data to the corresponding memory block 112 of the memory 110 within every predetermined time period. In 115, the logical value of the corresponding data flag may be determined, for example, by means of numerical accumulation or random change, thereby indicating whether the corresponding operational data is regularly updated.

It should be particularly noted that in the present embodiment, the monitoring system of the cabinet device is used for monitoring and controlling the fan units 51-54. Similarly, in other embodiments, the monitoring system of the cabinet device is also It can be a plurality of monitored units for monitoring and controlling the cabinet unit. The monitored units may be multiple nodes of the server cabinet, power supply units (such as multiple power supplies), or other computer hardware components. For example, when the monitored unit is a node, the board management controller operating as a cabinet management controller monitors the temperature, power consumption, SN information, ID, power-on state, power-on state, and hardware device of the node. Health status, configuration information (such as CPU, memory, hard drive, BIOS/BMC version), and power on/off control. When the monitored unit is a power supply unit, the board management controller operating as a cabinet management controller monitors the input power consumption, output power consumption, input voltage, output voltage, input current, output current, and switch control of the power unit. , status, and operating temperature.

In addition, in the embodiment, four memory blocks 112-115 of the five memory blocks 111-115 of the memory 110 are used to store the operational data, and the remaining memory block 111 is used. And storing related information generated by the substrate management controller 21 as the cabinet management controller, such as an ambient temperature in the cabinet housing 7, wattage and temperature of the power supply, etc., to provide other such substrates The management controllers 22 to 24 read.

In summary, the control unit stores the operational data detected by the substrate management controllers, that is, the control unit is regarded as a memory (such as an EEPROM), so that the substrate management controllers and Communication between the control units requires only simple data access, without the need for complex handshake mechanisms such as IPMI or IPMB implementations as in the prior art. In addition, each of the substrate management controllers respectively uploads the operation data to the control unit, so that the substrate management controller as the cabinet management controller can obtain the operation data detected by the other substrate management controllers at any time. Furthermore, the substrate management controller as the cabinet management controller writes the detected data into the control unit, so that each substrate management controller does not need to transfer data to each other, and only needs to read data from the control unit. Get the complete status of the entire cabinet unit. Therefore, not only can the complexity of the overall design be greatly reduced compared with the prior art, but also the development of the firmware is relatively simple, and the development cost and the design cost are simplified, so that the object of the present invention can be achieved.

However, the above is only the embodiment of the present invention, and the scope of the invention is not limited thereto, and all the equivalent equivalent changes and modifications according to the scope of the patent application and the patent specification of the present invention are still The scope of the invention is covered.

1‧‧‧Monitoring system
11‧‧‧Control unit
110‧‧‧ memory
111~115‧‧‧ memory block
21~24‧‧‧Baseboard Management Controller
31~34‧‧‧ busbar
41~44‧‧‧ board
51~54‧‧‧Fan unit
61~64‧‧‧Recognition pin
7‧‧‧ cabinet housing
9‧‧‧Cabinet Management Controller
91‧‧‧Cabinet backplane
92‧‧‧Power distribution board
921~923‧‧‧Power Module
93, 94‧‧‧Tray back plate
931~932‧‧‧ nodes
933~934‧‧‧Baseboard management controller
941~943‧‧‧ nodes
944~946‧‧‧Baseboard management controller
95‧‧‧Fan backplane
951~953‧‧‧fan

Other features and advantages of the present invention will be apparent from the embodiments of the present invention, wherein: Figure 1 is a block diagram illustrating a conventional cabinet arrangement; Figure 2 is a block diagram illustrating the cabinet of the present invention An embodiment of the device; and Figure 3 is a schematic diagram illustrating a memory of the embodiment.

1‧‧‧Monitoring system

11‧‧‧Control unit

110‧‧‧ memory

21~24‧‧‧Baseboard Management Controller

31~34‧‧‧ busbar

41~44‧‧‧ board

51~54‧‧‧Fan unit

61~64‧‧‧Recognition pin

7‧‧‧ cabinet housing

Claims (9)

A cabinet device comprising: a monitored unit; and a monitoring system comprising a plurality of busbars, a plurality of baseboard management controllers respectively electrically connecting the busbars and the monitored unit, and respectively generating each after starting up An operational data, and the operational data are respectively transmitted through the busbars, and each of which generates a heartbeat signal, and each of which is determined to operate between a master mode and a slave mode according to a corresponding determination signal. And a control unit electrically connecting the bus bars to receive and store the operational data from the baseboard management controllers, and electrically connecting the baseboard management controllers to receive the baseboard management controllers The heartbeat signals, and based on the heartbeat signals, generate the corresponding determination signals to control the first normal operators of the substrate management controllers to operate in the master mode and control the substrate management The rest of the controller operates in the slave mode, wherein the substrate operating in the master mode The controller reads the processing unit through the control bus corresponding to those stored in the operation information monitoring unit to monitor the subject. The cabinet device of claim 1, wherein: the control unit of the monitoring system further comprises a memory, the memory includes a plurality of memory blocks; and each of the substrate management controllers of the monitoring system further receives a corresponding And identifying, according to the corresponding identification signal, the generated operational data to the corresponding one of the memory blocks of the memory of the control unit via the corresponding bus. The cabinet device of claim 2, further comprising a plurality of boards, and a plurality of identification pins respectively disposed on the boards, wherein the identification pins determine logic values of the identification signals, wherein The baseboard management controllers of the monitoring system are respectively disposed on the board, and each of the baseboard management controllers generates the generated values according to the logic values of the identification signals corresponding to the identification pins disposed on the same board. The operation data is stored to the corresponding one of the memory blocks of the memory of the control unit via the corresponding bus bar. The cabinet device of claim 3, wherein: each of the baseboard management controllers of the monitoring system further generates a presence signal, and generates a first logic value and a second logic value when the power is turned on and normally operated. And the control unit of the monitoring system further determines, according to the presence signal from any one of the substrate management controllers, whether to generate the corresponding determination according to the heartbeat signal of the person a signal, when the logic value of the presence signal of the one is a third logic value, and the person is the first one of the baseboard management controllers, the heartbeat signal is at the first logic value and the second logic value When the jitter occurs, the logical value of the corresponding determination signal is changed from a fourth logic value to a fifth logic value, and when the logical value of the presence signal of the person is a sixth logic value, the corresponding determination signal The logical value is the fourth logical value. The cabinet apparatus of claim 4, wherein the baseboard management controller operating in the master mode operates as a rack management controller (Rack/Chassis Management Controller; RMC/CMC) to monitor the monitored unit. The cabinet device of claim 5, wherein the monitored unit comprises a plurality of fan units, each of the baseboard management controllers of the monitoring system is electrically connected to one of the fan units, each of the baseboard management controllers During normal operation, the fan unit that is electrically connected is detected to obtain the operation data related to the corresponding fan unit, and the baseboard management controller operating in the main control mode reads and stores via the corresponding bus bar. In the memory of the control unit, the operational data detected by the substrate management controllers in the slave mode are operated to monitor and control the fan units. The cabinet device of claim 6, wherein the control unit of the monitoring system further generates a plurality of data flags according to the operational data, and further stores the data flags in the memory, The data flag indicates whether the operational data is updated regularly. The cabinet apparatus of claim 7, wherein the busbars of the monitoring system support an agreement of an Inter-Integrated Circuit (I2C). The cabinet apparatus of claim 8, wherein one of the first logic value and the second logic value and the other one are logic 1 and logic 0, respectively, the fourth logic value and the fifth logic value One of the other and the other are logic 1 and logic 0, and one of the third logic value and the sixth logic value and the other are logic 1 and logic 0, respectively.