WO2014026374A1

WO2014026374A1 - Server system, management method, and device

Info

Publication number: WO2014026374A1
Application number: PCT/CN2012/080293
Authority: WO
Inventors: 郑力; 许利霞; 张羽
Original assignee: 华为技术有限公司
Priority date: 2012-08-17
Filing date: 2012-08-17
Publication date: 2014-02-20
Also published as: CN103210379A; CN103210379B

Abstract

Embodiments of the present invention provide a server system, a management method, and a device. The system comprises a management device, a PCIE exchange plate, and at least one service node device. The management device is connected to each service node device through the PCIE exchange plate. The management device is used for acquiring hardware resource information of each service node device through the PCIE exchange plate, building a physical resource pool according to the acquired information, and establishing a logical partition on the physical resource pool. In the technical solutions of the present invention, the management device and the service node device are connected through the PCIE exchange plate, thereby improving usage flexibility of hardware resources in the server system, and further improving performance of the whole system.

Description

The present invention relates to virtualization technologies, and in particular, to a server system, a management method, and a device. Background technique

In the development of server technology, the concept of logical partitioning effectively improves the utilization of server hardware resources. At present, the server system is generally formed by a plurality of service node device interconnection and combination, and the input/output (10) resources in the service node device are interconnected by a traditional network cable or cable, and the central processing in the service node device The Quick Processing Interconnect (QPI) cable and the dedicated chip are interconnected between the Central Processing Units (CPUs). The server management system divides the logical partitions on the hardware resources of the server system for service use.

However, the use of hardware resources in existing server systems is not flexible enough to further improve the performance of the server system. Summary of the invention

The embodiment of the invention provides a server system, a management method and a device, which are used to improve the flexibility of using hardware resources in a server system, and further improve the performance of the server system.

A first aspect of the present invention provides a server system, including: a management device, a high-speed peripheral component interconnect PCIE switch board, and at least one service node device; the management device through the PCIE switch board and each service Node device connection;

The management device is configured to acquire information about hardware resources of each service node device by using the PCIE switch board, construct a physical resource pool according to the obtained information, and establish a logical partition on the physical resource pool.

In an optional implementation manner of the first aspect, the management device is specifically configured to: present information about a hardware resource included in the physical resource pool to a user, and receive a selected instruction and a division instruction of the user to establish The logical partition, the selected instruction is used to indicate a hardware resource selected by the user from the physical resource pool according to an application requirement, and the dividing instruction is used to indicate the user pair The selected hardware resources are allocated to form the logical partition and the hardware resources included in the logical partition.

In an optional implementation manner of the first aspect, the management device is further configured to install an operating system in the logical user, where the hardware platform includes hardware resources included in the logical partition.

In an optional implementation manner of the first aspect, the management device is further configured to: send an early warning message when a load of the hardware resource included in the logical partition is greater than a preset first load threshold.

In an optional implementation manner of the first aspect, the management device is further configured to: select, from the physical resource pool, when the load of the hardware resource included in the logical partition is greater than a preset first load threshold. Adding hardware resources to the logical partition and triggering a hot plug process of the newly added hardware resource; or when loading a threshold, removing excess hardware resources from the logical partition to the physical resource pool The excess hardware resource is a hardware resource that is not utilized in the logical partition or whose utilization is less than a preset utilization threshold.

In an optional implementation manner of the first aspect, the management device is specifically connected to the hardware resources of each service node device by using the PCIE switch board;

The management device is configured to monitor, by using the PCIE switch board, the status of the hardware resources of each service node device, and obtain information about the hardware resources of each service node device by using the PCIE switch board.

In an optional implementation manner of the first aspect, the management device is specifically connected to the hardware resources of each service node device by using the PCIE switch board, and through the PCIE switch board and each of the The management controller of the service node device is connected;

The management device is configured to monitor, by using the PCIE switch board, a state of a hardware resource of each service node device, and obtain, by using the PCIE switch board, information about hardware resources of each service node device, or Corresponding relationship between the second ports, acquiring information of the hardware resources of each of the service node devices from the management controller of each service node device;

The management controller of each service node device is configured to monitor each of the service nodes. The state of the hardware resource of the device; the first port is a port connected to the hardware resource of each service node device on the PCIE switch board, and the second port is the PCIE switch board and each service node device The port to which the management controller is connected.

In an optional implementation manner of the first aspect, the management device is specifically connected to the hardware resources of each service node device by using the PCIE switch board, and directly managed with each service node device. Controller connection

The management device is configured to monitor, by using the PCIE switch board, a state of a hardware resource of each service node device, and obtain, by using the PCIE switch board, information about hardware resources of each service node device; Corresponding relationship between the second ports, acquiring information of the hardware resources of each of the service node devices from the management controller of each service node device;

The management controller of each of the service node devices is configured to monitor a state of the hardware resources of each of the service node devices; the first port is a hardware resource connection between the PCIE switch board and each of the service node devices. Port, the second port is a port on the management device that is connected to the management controller of each of the service node devices. .

A second aspect of the present invention provides a management device, including:

a monitoring module, connected to the at least one service node device by the high-speed peripheral component interconnection PCIE switch board, configured to acquire, by using the PCIE switch board, information about hardware resources of each service node device of the at least one service node device, Constructing a physical resource pool according to the obtained information;

A partition establishment module is configured to establish a logical partition on the physical resource pool.

In an optional implementation manner of the second aspect, the partition establishment module is specifically configured to: present information of a hardware resource included in the physical resource pool to a user, and receive a selected instruction and a division instruction of the user, The dividing instruction is used to indicate that the user selects a hardware resource from the physical resource pool according to an application requirement, and the dividing instruction is used to instruct the user to perform the selected hardware resource. Allocating the formed logical partition and the hardware resources included in the logical partition.

In an optional implementation manner of the second aspect, the partition establishment module is further configured to install an operating system in the user, where the hardware platform includes hardware resources included in the logical partition. In an optional implementation of the second aspect, the management device further includes: an early warning module, configured to acquire, by the monitoring module, that a load of the hardware resource included in the logical partition is greater than a preset first load gate For a limited time, an early warning message is issued.

In an optional implementation of the second aspect, the partition establishment module is further configured to: when the monitoring module acquires that the load of the hardware resource included in the logical partition is greater than a preset first load threshold, Adding idle hardware resources in the physical resource pool to the logical partition, and triggering a hot plug process of the newly added hardware resources; or

The partition establishment module is further configured to: when the monitoring module acquires that the load of the hardware resource included in the logical partition is less than a preset second load threshold, remove excess hardware resources from the logical partition to the physical In the resource pool, the excess hardware resources are hardware resources that are not utilized in the logical partition or whose utilization is less than a preset utilization threshold.

In an optional implementation manner of the second aspect, the monitoring module includes:

a first monitoring unit, configured to connect, by using the PCIE switch board, hardware resources of each of the service node devices, to monitor, by using the PCIE switch board, a status of hardware resources of each service node device, by using the PCIE The switch board acquires information about the hardware resources of each of the service node devices; and the constructing unit is configured to construct the physical resource pool according to the information about the hardware resources of each service node acquired by the first monitoring unit.

In an optional implementation manner of the second aspect, the monitoring module further includes:

a second monitoring unit, connected by the PCIE switch board or directly to the management controller of each service node device, between the first port and the second port connected according to the stored each service node device Corresponding relationship, obtaining information about hardware resources of each service node device from a management controller of each service node device;

The building unit is further configured to: according to the information about the hardware resource of each serving node acquired by the first monitoring unit, and/or the hardware resource of each serving node device acquired by the second monitoring unit Information, constructing the physical resource pool;

The management controller of each of the service node devices is configured to monitor a state of the hardware resources of each of the service node devices; the first port is a hardware resource connection between the PCIE switch board and each of the service node devices. Port, the second port is a port connected to the management controller of each serving node device on the management device or on the PCIE switch board.

A third aspect of the embodiments of the present invention provides a management method, including: The management device interconnects the PCIE switch board through the high-speed peripheral component to acquire information of hardware resources of each of the at least one service node device;

The management device constructs a physical resource pool according to the obtained information;

The management device establishes a logical partition on the physical resource pool;

The management device is connected to each of the service node devices through the PCIE switch board.

In an optional implementation manner of the third aspect, the establishing, by the management device, the logical partition on the physical resource pool includes: receiving a selected instruction and a dividing instruction of the user, to establish the logical partition, and selecting the The instruction is used to instruct the user to select a hardware resource from the physical resource pool according to an application requirement, where the dividing instruction is used to indicate the logical partition formed by the user to allocate the selected hardware resource, and the The hardware resources included in the logical partition.

In an optional implementation manner of the third aspect, the method further includes: the management device, to enable the user to install an operating system, where the hardware platform includes hardware resources included in the logical partition.

In an optional implementation manner of the third aspect, the method further includes: the management device sends an early warning message when a load of the hardware resource included in the logical partition is greater than a preset first load threshold.

In an optional implementation manner of the third aspect, the method further includes: the management device, when the load of the hardware resource included in the logical partition is greater than a preset first load threshold, from the physical resource pool Adding idle hardware resources to the logical partition and triggering a hot plug process of the newly added hardware resource; or threshold, removing excess hardware resources from the logical partition to the physical resource In the pool, the excess hardware resources are hardware resources that are not utilized in the logical partition or whose utilization is less than a preset utilization threshold.

In an optional implementation manner of the third aspect, the management device is connected to the hardware resources of each service node device by using the PCIE switch board; The information that the management device obtains the hardware resources of each of the at least one service node device by interconnecting the PCIE switch board by the high-speed peripheral component includes:

The management device monitors the status of the hardware resources of each of the service node devices through the PCIE switch board, and obtains information about the hardware resources of each service node device through the PCIE switch board.

In an optional implementation manner of the third aspect, the management device is connected to the hardware resources of each service node device by using the PCIE switch board, and through the PCIE switch board and each service The management device of the node device is connected;

The management device accessing the PCIE switch board through the high-speed peripheral component to obtain the hardware resources of each of the service node devices of the at least one service node device includes:

The management device monitors, by using the PCIE switch board, the status of the hardware resources of each service node device, and obtains, by using the PCIE switch board, information about hardware resources of each service node device; or

The management device acquires, according to the stored correspondence between the first port and the second port of each service node device, the management controller of each service node device Information about hardware resources;

The management controller of each of the service node devices is configured to monitor a state of the hardware resources of each of the service node devices; the first port is a hardware resource connection between the PCIE switch board and each of the service node devices. Port, the second port is a port on the PCIE that is connected to the management controller of each of the service node devices.

In an optional implementation manner of the third aspect, the management device is connected to the hardware resources of each service node device by using the PCIE switch board, and directly controls the management of each service node device. Connection

The management device monitors, by using the PCIE switch board, a state of hardware resources of each service node device, and obtains, by using the PCIE switch board, information about hardware resources of each service node device, or

The management device acquires each of the service port devices from each of the service node devices according to the stored correspondence between the first port and the second port of each of the service node devices Information about the hardware resources of the service node device;

The management controller of each of the service node devices is configured to monitor a state of the hardware resources of each of the service node devices; the first port is a hardware resource connection between the PCIE switch board and each of the service node devices. Port, the second port is a port on the management device that is connected to the management controller of each of the service node devices.

In the server system provided by the first aspect of the present invention, the management device and the service node device are connected through a PCIE switch board, and the management device acquires information about hardware resources of each service node device through the PCIE switch board, and builds the information according to the obtained hardware resource information. The physical resource pool is used to establish a logical partition on the physical resource pool. The physical resource pool that shields the physical topology information is implemented at the physical resource level by using the PCIE switch board, which improves the flexibility of the use of hardware resources in the physical resource pool. Improve the performance of the entire system.

The management device provided by the second aspect of the present invention is connected to the service node device through the PCIE switch board, obtains the hardware resource information of each service node device through the PCIE switch board, and constructs a physical resource pool according to the acquired hardware resource information. In addition, a logical partition is established on the physical resource pool, and the physical resource pool for shielding physical topology information is implemented at the physical resource level by using the PCIE switch board, thereby improving the flexibility of using hardware resources in the physical resource pool, thereby effectively improving the entire system. Performance.

According to the management method provided by the third aspect of the present invention, the management device obtains the information of the hardware resources of each service node device by interconnecting the PCIE switch board with the high-speed peripheral component, and constructs a physical resource pool according to the acquired hardware resource information, and then A logical partition is created on the physical resource pool to fully utilize the PCIE switch board to implement physical resource pools that shield physical topology information at the physical resource level, thereby improving the flexibility of using hardware resources in the physical resource pool, thereby effectively improving the performance of the entire system. . BRIEF DESCRIPTION OF THE DRAWINGS In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief description of the drawings used in the embodiments or the prior art description will be briefly described below. The drawings are some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any inventive labor.

1 is a schematic structural diagram of a server system according to an embodiment of the present invention;

2 is a schematic diagram of a specific structure of a server system according to an embodiment of the present invention; 3 is a schematic structural diagram of a management device connected to a CPU and a management controller of a computing node through a PCIE switch board according to an embodiment of the present invention;

4 is a schematic structural diagram of a logical partition according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of an overall structure of a logical partition built on a physical resource pool and a physical resource pool according to an embodiment of the present invention;

FIG. 6 is a schematic structural diagram of a management device according to an embodiment of the present invention;

FIG. 7 is a schematic structural diagram of a management device according to another embodiment of the present invention;

FIG. 8 is a flowchart of a management method according to an embodiment of the present invention;

FIG. 9 is a schematic structural diagram of a management method according to another embodiment of the present invention;

FIG. 10 is a schematic structural diagram of a management method according to another embodiment of the present invention. The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. The embodiments are a part of the embodiments of the invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

FIG. 1 is a schematic structural diagram of a server system according to an embodiment of the present invention. As shown in FIG. 1, the system of this embodiment includes a management device 11, a Peripheral Component Interconnect Express (PCIE) switchboard 12, and at least one service node device 13. The management device 11 is connected to each service node device 13 through the PCIE switch board 12. The PCIE switchboard 12 has features and advantages such as high bandwidth and low latency.

The service node device 13 in this embodiment may be various types of service node devices, for example, a 10-node, a computing node, a storage node, a memory expansion board, a fan board, a power board, and the like.

The management device 11 is configured to acquire information about hardware resources of each service node device 13 through the PCIE switch board 12, construct a physical resource pool according to the obtained information of hardware resources of each service node device, and establish a physical resource pool on the physical resource pool. Logical partition. The hardware resources in this document refer to the actual hardware constituting the server system. More specifically, the hardware resources may be, for example, a CPU, a memory card, a hard disk, etc., which may be implemented by a server system to a computing node, a storage node, and 10 Pool management of hardware resources in a node to form a physical resource pool, which can be subdivided into, for example, a CPU Pool, memory pool, 10 resource pool, network card pool, etc. The information about the hardware resources of each service node device 13 includes status information of hardware resources such as network card, CPU, memory, storage device in-position information, fault information, load, temperature, etc., and may further include, for example, a memory size, a CPU. Device attribute information of hardware resources such as frequency and manufacturer.

Specifically, the management device 11 can obtain each service node device through the PCIE switch board 12.

13 hardware resource information, and according to the hardware resource capacity, model and other attribute information and status information contained in the hardware resource information, use the available hardware resources to form a physical resource pool, and then establish logic on the constructed physical resource pool Partition.

In this embodiment, the management device and the service node device are connected through the PCIE switch board, and the management device acquires the hardware resource information of each service node device through the PCIE switch board, and builds a physical resource pool according to the acquired hardware resource information, and then Logical partitions are created on the physical resource pool to fully utilize the PCIE switch board to implement physical resource pools that shield physical topology information at the physical resource level, thereby improving the flexibility of using hardware resources in the physical resource pool and dynamically adjusting the logical partitions. Effectively improve the performance of the entire system.

Further, the hardware resources of the 10 nodes in the embodiment refer to 10 resources, and the hardware resources of the computing nodes refer to CPU resources, and the hardware resources of the storage nodes refer to hard disk drives (HDDs) and solid state disks (Solid State Disk). , SSD) and other storage resources. Optionally, each service node device 13 may include a management controller in addition to the foregoing hardware resources. The management controller can be kept in a power-on state, and is mainly used to obtain information about the hardware resources of the service node device 13 for the management device 11 in an abnormal situation. For example, a certain hardware resource in the service node device 13 is powered off to cause the management device. If the information of the hardware resource cannot be obtained from the hardware resource, the management device 11 can obtain the information of the hardware resource through the management controller of the service node device 13, and in other cases, obtain the service node device through the management controller. 13 information on hardware resources. Optionally, the management controller may be a Baseboard Management Controller (BMC) or a logic device that implements a BMC-like function.

In the server system provided in this embodiment, various types of service node devices may be deployed one or more as needed, and there are usually a plurality of computing nodes, storage nodes, and the like, and the server system may be expanded as needed.

The PCIE switch board 12 may include a Gigabit Ethernet (GE) interface, a Fast Ethernet (FE) interface, and/or a PCIE interface, and the management device 11 can pass. At least one of a GE interface, a FE interface, and a PCIE interface is connected to each of the service node devices 13. The GE interface, the FE interface, and the PCIE interface may be implemented on a PCIE switch board 12 or on multiple PCIE switch boards 12 in a specific implementation.

Based on the above, Fig. 2 shows a specific structure of a server system. In FIG. 2, the management device 11 is connected to the hardware resources of the service node device such as the 10-node 131, the compute node 132, and the storage node 133 through the PCIE interface on the PCIE switch board 12. The management device 11 passes the GE interface on the PCIE switch board 12. It is connected to a management controller of 10 nodes 131, compute nodes 132, storage nodes 133, and the like. In FIG. 2, the shared module such as the fan board 14 in the entire server system is directly connected to the management device 11 through its management controller, but is not limited thereto, and other service node devices other than the fan board 14, for example, 10 nodes 131, calculation The management controller of the node 132, the storage node 133, and the like may also be directly connected to the management device 11. That is to say, the management controller of the service node device can directly connect with the management device 11 without going through the PCIE switch board 12, but the hardware resources of the service node device 13 must be connected to the management device 11 through the PCIE switch board 12.

Based on the above, the connection manner between the management device 11 and the service node device 13 and the information acquisition manner in each connection mode include but are not limited to the following:

The management device 11 can be connected to the hardware resources of each service node device 13 through the PCIE switch board 12, and the management device 11 can monitor the state of the hardware resources of each service node device 13 through the PCIE switch board 12, and then pass through the PCIE switch board. 12 Obtain information of hardware resources of each service node device 13.

In addition, the management device 11 can also be connected to the hardware resources of each service node device 13 through the PCIE switch board 12, and connected to the management controller of each service node device 13 through the PCIE switch board 12, and then manage the device. The status of the hardware resources of each service node device 13 can be monitored by the PCIE switch board 12, and the information of the hardware resources of each service node device 13 can be obtained through the PCIE switch board 12, or the management device 11 can also be stored according to each Corresponding relationship between the first port and the second port connected by the service node device 13 acquires information of hardware resources of each service node device 13 from the management controller of each service node device 13. The management controller of each service node device 13 is used to monitor the state of the hardware resources of each service node device 13. The first port is a port on the PCIE switch board 12 that is connected to the hardware resources of each service node device 13, and the second port is a port on the PCIE switch board 12 that is connected to the management controller of each service node device 13. That is, the management device 11 can obtain each service node setting through the PCIE switch board. The information of the hardware resources of the standby device 13 may also receive the information of the hardware resources of each of the service node devices 13 that are reported from the management controller of each service node device 13 and according to the stored between the first port and the second port. The corresponding relationship determines the hardware resources corresponding to the information reported by the management controller, thereby obtaining the information of the hardware resources of each service node device 13. The management device 11 can obtain all the information about the hardware resources of each service node device 13 through the PCIE switch board, and can also obtain all the information about the hardware resources of each service node device 13 from the management controller, or obtain the hardware through the PCIE switch board. Part of the resource information, get another part of the information from the management controller. For example, the management controller acquires the in-position information of the hardware resources, the manufacturer information, and the like, and acquires the load information of the hardware resources through the PCIE switch board.

In addition, the management device 11 can specifically pass through the PCIE switch board 12 and each service node device.

The hardware resources of the 13 are connected to the management controller of each of the service node devices 13, and the management device 11 can monitor the state of the hardware resources of each of the service node devices 13 through the PCIE switch board 12 through the PCIE switch board 12. Obtaining the information of the hardware resources of each service node device 13, or the management device 11 may also obtain the correspondence between the first port and the second port connected by each of the service node devices 13 from each of the service node devices. The management controller of 13 acquires information of the hardware resources of each of the service node devices 13. The management controller of each service node device 13 is configured to monitor the state of the hardware resources of each service node device 13; the first port is a port on the PCIE switch board 12 that is connected to the hardware resources of each service node device 13, The two ports are ports on the management device 11 that are connected to the management controller of each of the service node devices 13. That is, the management device 11 can obtain the information of the hardware resources of each of the service node devices 13 through the PCIE switch board, and can also receive the information of the hardware resources of each of the service node devices 13 that are reported from the management controller of each service node device 13. And determining, according to the correspondence between the stored first port and the second port, the hardware resources corresponding to the information reported by the management controller, thereby obtaining the information of the hardware resources of each service node device 13. The management device 11 can obtain all the information about the hardware resources of each service node device 13 through the PCIE switch board, and can also obtain all the information about the hardware resources of each service node device 13 from the management controller, or obtain the hardware through the PCIE switch board. Part of the resource information, get another part of the information from the management controller. For example, the management controller obtains the in-position information of the hardware resources, the manufacturer information, and the like, and acquires the load information of the hardware resources through the PCIE switch board.

Wherein, if the management controller of the service node device 13 is connected to the PCIE switch board 12, the second port is a port on the PCIE switch board 12; if the management node and the management node of the service node device 13 The device 11 is directly connected, and the second port is a port on the management device 11. For example, taking the computing node 132 as the serving node device 13 as an example, the CPU of the computing node 132 is interconnected with the management device 11, and the detailed structure of the management controller of the computing node 132 and the management device 11 is as shown in FIG. 3. The management device 11 is connected to the GE interface and the PCIE interface on the PCIE switch board 12, and the management controller of the compute node 132 is connected to the port 10 (PortlO) of the GE interface. The CPU of the node 132 and the port 20 of the PCIE interface (Port 20) )connection. Only one CPU is connected to the management device 11 through the PCIE interface in FIG. 3, but is not limited thereto. When the computing node 132 includes a plurality of CPUs, each CPU is connected to the management device 11 through a PCIE interface. Among them, PortlO belongs to the second port, and Port20 belongs to the first port. The management device 11 stores the correspondence between the Port 10 and the Port 20 connected to the computing node 132. The management device 11 can obtain information about the CPU on the computing node 132, such as the slot number of the CPU. Specifically, the management device 11 further stores a correspondence between the identifier of the CPU on the computing node 132 connecting the port 20 and the connection port 20 and the slot number of the CPU. Based on this, the management device 11 can obtain the computing node 132 through the port 10. The management unit reports part of the information of the CPU, such as the in-position information of the CPU, the manufacturer information, and the like, and further obtains the port 20 connected to the CPU according to the correspondence between the port 10 and the port 20, and then according to the acquired identifier of the CPU. Obtain other information such as the slot number and load of the CPU.

In an optional implementation manner of the embodiment, the process for the management device 11 to establish a logical partition on the physical resource pool is: the management device 11 is configured to present information of the hardware resource included in the physical resource pool to the user, and receive the selected instruction of the user. And dividing instructions to establish the logical partition. The selected instruction is used to indicate a hardware resource selected by the user from the physical resource pool according to the application requirement, and the dividing instruction is used to indicate the hardware resource included in the logical partition and the logical partition formed by the user to allocate the selected hardware resource in the selected instruction. . The function of the management device 11 to present the information of the hardware resources included in the physical resource pool to the user is equivalent to providing the user with a logical partition platform, so that the user uses the PCIE switch board 12 to perform management between the device 11 and the service node device 13. Interconnection, which solves the difference between the hardware resources of the service node devices 13 caused by various interconnection methods such as network cables and cables. The topology relationship of the same type of hardware resources is the same for the user, and therefore, When partitioning logical partitions, you can consider only the application requirements, without considering the differences between hardware resources, and without considering the physical topology of each hardware resource, which is beneficial to improve the division of logical partitions. effectiveness.

The process of selecting hardware resources in a physical resource pool according to application requirements and performing allocation to form different logical partitions is illustrated. For example, when deploying a desktop cloud application, two logical partitions are required to host the desktop cloud terminal service and related database applications. According to the evaluation: The desktop cloud terminal service requires 8 CPUs, several memory, storage, and 100 virtual network cards; the database application requires 2 CPUs, some memory, storage, and a virtual network card. The user can select the CPU, memory, storage, and network card according to the requirements on the logical partition platform presented by the management device 11, and then divide the selected CPU, memory, storage, network card and other hardware resources into corresponding CPUs, memories, and The storage and network card resources may constitute a logical partition that carries the desktop cloud terminal service and a logical partition of the related database application.

In an optional implementation manner of the embodiment, after the logical partition is established, the management device 11 is further configured to: after the logical partition is powered on, present a hardware platform corresponding to the logical partition to the user, so that the user installs the operating system. And middleware or upper-level applications. The hardware platform includes hardware resources included in the logical partition. The operating system can be a normal operating system or a virtual operating system. After the logical partition is established, the management device 11 presents the information of the hardware resources in each logical partition to the user, which is equivalent to presenting one hardware platform to the user, and each hardware platform is similar to the ordinary for the user. Server or computer. The user can then choose to install a normal operating system or virtual operating system on the hardware platform. The function of the management device 11 to present the hardware platform corresponding to the logical partition to the user is equivalent to providing an enhanced adaptation layer on the hardware resource (which can be referred to as the hardware layer) of the logical partition. The enhanced adaptation layer here provides the role of the BIOS for this logical partition, in addition to providing open hardware platform support, various virtualization software support, and more.

Based on the above, the structure of a logical partition is as shown in FIG. 4, which includes a hardware layer, an enhanced adaptation layer, and an operating system in order from bottom to top.

Based on the above, the embodiment provides an overall structure of a logical partition established by the management device 11 on a physical resource pool, as shown in FIG. 5. The physical resource pool mainly includes a CPU pool composed of CPU resources, 10 pools composed of 10 resources, and a memory (MEM) pool composed of MEM resources, wherein the CPU resources, 10 resources, and MEM resources are service node devices. Hardware resources in the network; N logical partitions are established on the physical resource pool, which are logical partitions 1 - logical branches. Each logical partition uses the included hardware resources as a hardware platform, and the hardware platform includes enhanced adaptation. Layers, operating systems, and applications. As shown in Figure 5, the operating system of logical partition 1 is mainly a common operating system, for example, Windows/Linux/unix, etc.; the operating system of logical partition N is a virtual operating system, and a virtualized operating system, such as ESX of vmware, needs to be installed first. , create a virtual machine (VM) on it, and then install the operating system and applications for the VM. Among them, the VMware ESX server is a virtual host software that partitions and consolidates the system in a common environment. Therefore, in the server system of the embodiment, since the PCIE switch board is used for the interconnection between the management device and the service node device, the characteristics and advantages of the high bandwidth and low delay of the PCIE switch board are fully utilized, so that the built The physical resource pool shields the physical topology information of each hardware resource, and the virtualization management software layer can be omitted, which solves the problem that the prior art must shield the hardware resource difference caused by the physical topology through the virtualization technology, and reduce the problem. Manage costs while improving performance.

In an optional implementation manner of the embodiment, the management device 11 is further configured to issue the early warning information when the load of the hardware resource included in the logical partition is greater than a preset first load threshold.

In an optional implementation manner of this embodiment, the management device 11 is further configured to: when the load of the hardware resource included in the logical partition is greater than a preset first load threshold, select an idle hardware resource to be added from the constructed physical resource pool. Go to the logical partition and trigger the hot plug process of the newly added hardware resources. The hot plug process of the newly added hardware resources will be different depending on the newly added hardware resources. For example, if the newly added hardware resource is a NIC resource, the corresponding hot plug process is mainly to add a driver for the newly added hardware resource. If the newly added hardware resource is a CPU, the corresponding hot plug process is mainly the CPU hot swap, online, offline, and the like.

In the above optional implementation manner, the management device 11 may acquire the load status of the hardware resources in the logical partition by monitoring the state of the hardware resources of each service node device 13 during the operation of the logical partition, and include the logical partition in the discovery logical partition. When the load of the hardware resource is greater than the preset first load threshold, the warning information is sent and/or the hardware resource is newly added for the logical partition, and the hot plug process is triggered after the new hardware resource is added.

The first load threshold may be an upper threshold in the load threshold, but is not limited thereto.

Further, the management device 11 may further remove excess hardware resources from the logical partition into the physical resource pool when the load of the hardware resources included in the logical partition is less than the preset second load threshold. The second load threshold may be a lower threshold in the load threshold. The excess hardware resource refers to The excess hardware resources in the partition to the physical resource pool is beneficial to improve the utilization of hardware resources. The hardware resources with lower utilization ratios are usually the hardware resources whose utilization is lower than the preset usage threshold. The thresholds can be set according to different application scenarios.

It is noted that the first load threshold is greater than the second load threshold, and the specific values of the first load threshold and the second load threshold are not limited.

It can be seen from the above that the management device 11 of the embodiment can flexibly adjust the hardware resources in the logical partition according to the information of the hardware resources, and manage the hardware resources such as the CPU and the CPU by acquiring the information of the hardware resources in the logical partition when the logical partition is running. Hot swapping, enabling dynamic management of logical partitions. In combination with the above, the dynamic management of logical partitions can be specifically implemented by the enhanced adaptation layer provided by the management device 11.

It can be seen from the foregoing that the embodiment of the present invention implements interconnection between the management device and the service node device through the PCIE switch board, and fully utilizes the PCIE switch board to implement free scheduling of hardware resources in the system, so that the logical partition is not subject to physical resource topology. The impact of the creation and use efficiency of the logical partition is improved; the user does not need to perceive the difference of the PCIE switch architecture, and the logical partition operation can be directly performed on the physical resource pool, which can save the virtualization management software layer, reduce the management cost, and improve the time. Performance; In addition, during the operation of the logical partition, through the monitoring of hardware resources, the dynamic logical partition management scheme is realized, and the flexible allocation and hot swapping of 10 physical resources are realized.

FIG. 6 is a schematic structural diagram of a management device according to an embodiment of the present invention. As shown in FIG. 6, the management device of this embodiment includes: a monitoring module 61 and a partition establishing module 62.

The monitoring module 61 is connected to the at least one service node device through the PCIE switch board, and is configured to acquire, by using the PCIE switch board, information about hardware resources of each service node device in the at least one service node device, and construct the information according to the obtained information. Physical resource pool. The partition establishment module 62 is connected to the monitoring module 61 for establishing a logical partition on the physical resource pool constructed by the monitoring module 61.

The service node device in this embodiment may be various types of service node devices, for example, a 10-node, a computing node, a storage node, a memory expansion board, a fan board, a power board, and the like. The hardware resources of the 10 nodes refer to 10 resources, the hardware resources of the computing nodes refer to CPU resources, and the hardware resources of the storage nodes refer to various storage resources such as HDD and SSD.

The management device of the embodiment is connected to the service node device through the PCIE switch board, and the management device obtains the hardware resource information of each service node device through the PCIE switch board, and according to the acquired hardware. Resource information builds a physical resource pool, and then establishes logical partitions on the physical resource pool to make full use of

The PCIE switch board implements a physical resource pool that blocks physical topology information at the physical resource level, which improves the flexibility of using hardware resources in the physical resource pool, thereby effectively improving the performance of the entire system.

FIG. 7 is a schematic structural diagram of a management device according to another embodiment of the present invention. The embodiment is implemented based on the embodiment shown in FIG. 6. As shown in FIG. 7, the management device of this embodiment also includes: a monitoring module 61 and a partition establishing module 62.

Optionally, each service node device may include a management controller in addition to hardware resources. The management controller can be kept in the power-on state, and is mainly used for the management device to obtain the information of the hardware resources of the service node device under abnormal conditions. Of course, under normal circumstances, the management controller can also obtain information about the hardware resources of the service node device.

Based on the above, the monitoring module 61 of the embodiment specifically includes a first monitoring unit 611 and a construction unit 613.

The first monitoring unit 611 is connected to the hardware resources of each service node device through the PCIE switch board, and is configured to monitor the state of the hardware resources of each service node device through the PCIE switch board, and obtain each service node through the PCIE switch board. Information about the hardware resources of the device. The PCIe switch board may include GE, FE, and/or PCIE interfaces, and the first monitoring unit 611 may connect to the hardware resources of each serving node device through any one of the GE, FE, and PCIE interfaces.

The building unit 613 is connected to the first monitoring unit 611, and is configured to construct the physical resource pool according to the information of the hardware resources of each service node acquired by the first monitoring unit 611.

Further, the monitoring module 61 of the embodiment further includes: a second monitoring unit 612. The second monitoring unit 612 is connected to the management controller of each service node device through the PCIE switch board, and is configured to use the correspondence between the first port and the second port connected to each of the stored service node devices. The information of the hardware resources of each service node device is obtained on the management controller of each service node device. Similarly, if the second monitoring unit 612 is connected to the management controller of the service node device through the PCIE switch board, the second monitoring unit 612 can also connect to each service node device through any one of the GE, FE, and PCIE interfaces. Management controller connection. In addition, the second monitoring unit 612 can also be directly connected to the management controller of the serving node device without passing through the PCIE switchboard.

The building unit 613 is also connected to the second monitoring unit 612, and is also used for information of hardware resources of each serving node acquired according to the first monitoring unit 611 and/or hardware of each serving node device acquired by the second monitoring unit 612. Resource information, build a pool of physical resources. The management controller of each service node device is configured to monitor the state of the hardware resources of each service node device; the first port is a port connected to the hardware resource of the service node device on the PCIE switch board, and the second port is The port to which the management node device's management controller is connected. If the second monitoring unit 612 is connected to the management controller of the service node device through the PCIE switch board, the second port may be a port on the PCIE switch board 12 if the second monitoring unit 612 is directly connected to the management controller of the service node device. The second port may be a port on the management device (more specifically, the port may be a port on the second monitoring unit 612). The information of the hardware resources of each service node device may be acquired by the first monitoring unit 611 through the PCIE switch board, or may be received by the second monitoring unit 612 from the management controller of each service node device. The information of the hardware resource is determined according to the correspondence between the stored first port and the second port, and the hardware resources corresponding to the information reported by the management controller are determined, thereby obtaining the information of the hardware resources of each service node device. All the information of the hardware resources of each service node device may be acquired by the first monitoring unit 611 through the PCIE switch board, and all the information of the hardware resources of each service node device may also be obtained by the second monitoring unit 612 from the management controller. A portion of the hardware resources may be acquired by the first monitoring unit 611 through the PCIE switchboard, and another portion of the information may be acquired by the second monitoring unit 612 from the management controller.

For example, the computing node of the computing node is connected to the management controller of the computing node by the first monitoring unit 611 and the second monitoring unit 612 respectively. The structure is shown in Figure 3.

Further, the partition establishment module 62 of the embodiment is specifically configured to present the information of the hardware resources included in the physical resource pool constructed by the monitoring module 61 to the user, and receive the selected instruction and the partitioning instruction issued by the user to establish the logical partition. The selected instruction is used to indicate a hardware resource selected by the user from the physical resource pool according to the application requirement, and the dividing instruction is used to indicate the logical partition formed by the user to allocate the selected hardware resource and the hardware resource included in the logical partition. . The function of the partition establishment module 62 to present the information of the hardware resources included in the physical resource pool to the user is equivalent to providing the user with a logical partition platform, so that the user performs logical partition division on the logical partition platform. Here, since the management device uses the interconnection between the PCIE switch board and the service node device, the difference between the hardware resources of each service node device caused by the various interconnection modes such as the network cable and the cable is solved, and the extension of the same type of hardware resources is solved. The relationship between the PCs is the same for the user. Therefore, when logical partitioning is performed, only the application requirements can be considered without considering the difference between the hardware resources. Different, it is also unnecessary to consider the physical topology of each hardware resource, which is beneficial to improve the efficiency of partitioning logical partitions.

For an example of the process of selecting a hardware resource in a physical resource pool according to an application requirement and performing a process of allocating different logical partitions, refer to the description of the foregoing system embodiment.

Further, the partition establishment module 62 of the embodiment is further configured to: after the logical partition is powered on, present a hardware platform corresponding to the logical partition to the user, so that the user installs the operating system. The hardware platform includes hardware resources included in the logical partition. The operating system can be a normal operating system or a virtual operating system. After the logical partition is established, the partition establishing module 62 presents the hardware resource information in each logical partition to the user, which is equivalent to presenting one hardware platform to the user, and each hardware platform is similar to the ordinary user. Server or computer. Users can then choose to install a normal operating system or virtual operating system on the hardware platform. The function of the partitioning module 62 to present the hardware platform corresponding to the logical partition to the user is equivalent to providing an enhanced adaptation layer on the hardware resource (which may be referred to as the hardware layer) of the logical partition. The enhanced adaptation layer here provides the role of the BIOS for the logical partition, in addition to providing open hardware platform support, various virtualization software support, and so on.

A structure of the logical partition constructed by the monitoring device 61 and the partition establishing module 62 of the embodiment management device is as shown in FIG. The physical structure of the physical resource pool constructed by the monitoring module 61 and the partition establishing module 62 and the logical partition established on the physical resource pool in this embodiment are as shown in FIG. 5.

Further, the management device of this embodiment further includes: an early warning module 63. The warning module 63 is connected to the monitoring module 61, and is configured to issue the warning information when the monitoring module 61 acquires that the load of the hardware resource included in the logical partition is greater than the preset first load threshold. Specifically, the early warning module 63 is connected to the first monitoring unit 611 and the second monitoring unit 612, respectively.

The monitoring module 61 can monitor the state of the hardware resources of each service node device through the PCIE switch board, and obtain the load status of the hardware resources in the logical partition. Alternatively, the monitoring module 61 may obtain the state of the hardware resources of each service node device through the management controller of each service node device, thereby acquiring the load status of the hardware resources in the logical partition.

Further, the partition establishment module 62 of the embodiment is further configured to: when the monitoring module 61 acquires that the load of the hardware resource included in the logical partition is greater than a preset first load threshold, select an idle hardware resource from the physical resource pool to add to the In the logical partition, and trigger hot swap of newly added hardware resources Process.

The hot plug process of newly added hardware resources will vary depending on the newly added hardware resources. For example, if the newly added hardware resource is a NIC resource, the corresponding hot plug process is mainly to add a driver for the newly added hardware resource. If the newly added hardware resource is a CPU, the corresponding hot plug process is mainly the CPU hot swap, online, offline, and the like.

Further, the partition establishment module 62 may also remove excess hardware resources from the logical partition into the physical resource pool when the load of the hardware resources included in the logical partition is less than the preset second load threshold. The second load threshold may be a lower threshold in the load threshold. The excess hardware resources, in addition to the excess hardware resources in the logical partition to the physical resource pool, are beneficial to improve the utilization of hardware resources. The lower-utilization hardware resources usually refer to hardware resources whose utilization is less than the preset utilization threshold. The utilization threshold can be set to different values according to actual requirements.

It can be seen from the above that the partition establishment module 62 of the embodiment can flexibly adjust the hardware resources in the logical partition according to the information of the hardware resources, and manage the hardware such as the CPU and the CPU by acquiring the information of the hardware resources in the logical partition when the logical partition is running. Hot swapping of resources to achieve dynamic management of logical partitions. In conjunction with the above, dynamic management of logical partitions may be specifically implemented by an enhanced adaptation layer provided by partition creation module 62.

It can be seen that the management device in this embodiment implements interconnection with the service node device through the PCIE switch board, and implements free scheduling of hardware resources in the system by using the PCIE switch board, so that the logical partition is not affected by the physical resource topology. The efficiency of creating and using logical partitions is improved; the user does not need to perceive the difference of the PCIE switching architecture, and can perform logical partitioning operations directly on the physical resource pool, thereby eliminating the virtualization management software layer, reducing management costs, and improving performance. In addition, during the operation of the logical partition, the management scheme of the dynamic logical partition is realized through the monitoring of the hardware resources, and the flexible allocation and hot plugging of the 10 physical resources are realized.

FIG. 8 is a flowchart of a management method according to an embodiment of the present invention. As shown in FIG. 8, the method in this embodiment includes:

Step 801: The management device acquires each of the at least one service node device through the PCIE switch board. Information about the hardware resources of the service node device.

Step 802: The management device builds a physical resource pool according to the obtained information.

Step 803: The management device establishes a logical partition on the foregoing physical resource pool.

The management device is connected to each service node device through a PCIE switch board.

The service node device in this embodiment may be various types of service node devices, for example, 10 nodes, a computing node, a storage node, a memory expansion board, a fan board, a power board, and the like. The hardware resources of the 10 nodes refer to 10 resources, the hardware resources of the compute nodes refer to CPU resources, and the hardware resources of the storage nodes refer to various storage resources such as HDD and SSD.

In an optional implementation of this embodiment, the management device is connected to the hardware resources of each service node device through a PCIE switch board. The specific implementation manner of step 801 includes: the management device monitors the state of the hardware resources of each service node device through the PCIE switch board, and obtains the hardware resource information of each service node device through the PCIE switch board. Specifically, the interaction between the management device and each service node device is implemented through a PCIE switch board. In this embodiment, the management device can be connected to the management controller of each service node device through the PCIE switch board, or directly to the management controller of each service node device.

In an optional implementation manner of this embodiment, the management device is connected to the hardware resources of each service node device through the PCIE switch board, and is connected to the management controller of each service node device through the PCIE switch board. Based on this, the specific implementation manner of step 801 includes: the management device monitors the state of the hardware resources of each service node device through the PCIE switch board, acquires the hardware resource information of each service node device through the PCIE switch board, or manages the device according to the storage. Corresponding relationship between the first port and the second port of each service node device connection, and acquiring information of hardware resources of each service node device from a management controller of each service node device. The management controller of each service node device is used to monitor the status of hardware resources of each service node device. The first port is a port connected to the hardware resource of each service node device on the PCIE switch board, and the second port is a port connected to the management controller of each service node device on the PCIE switch board. That is, the management device may obtain the information of the hardware resources of each service node device through the PCIE switch board, and may also receive the information of the hardware resources of each service node device that is reported from the management controller of each service node device, and according to the storage. Corresponding relationship between the first port and the second port determines the hardware resources corresponding to the information reported by the management controller, thereby obtaining the information of the hardware resources of each service node device. The management device can obtain all the information about the hardware resources of each service node device through the PCIE switch board. It is also possible to obtain all the information of the hardware resources of each service node device from the management controller, or obtain a part of the hardware resources through the PCIE switch board, and obtain another part of the information from the management controller.

In an optional implementation manner of this embodiment, the management device is connected to the hardware resources of each service node device through the PCIE switch board, and is directly connected to the management controller of the service node device. Based on this, the specific implementation manner of the step 801 includes: the management device monitors the state of the hardware resources of each service node device through the PCIE switch board, acquires the hardware resource information of each service node device through the PCIE switch board, or manages the device according to the storage. Corresponding relationship between the first port and the second port of each service node device connection, and acquiring information of hardware resources of each service node device from a management controller of each service node device. The management controller of each service node device is used to monitor the status of hardware resources of each service node device. The first port is a port on the PCIE switch board that is connected to the hardware resource of each service node device, and the second port is a port on the management device that is connected to the management controller of each service node device. That is, the second port can be a port on the PCIE switch board or a port on the management device. When the management device is connected to the management controller of the service node device through the PCIE switch board, the second port is a port on the PCIE switch board; when the management device is directly connected to the management controller of the service node device, but not through the PCIE switch board The second port is the port on the management device. That is, the management device may obtain the information of the hardware resources of each service node device through the PCIE switch board, and may also receive the information of the hardware resources of each service node device that is reported from the management controller of each service node device, and according to the storage. Corresponding relationship between the first port and the second port determines the hardware resources corresponding to the information reported by the management controller, thereby obtaining the information of the hardware resources of each service node device. The management device can obtain all the information about the hardware resources of each service node device through the PCIE switch board, or obtain all the hardware resources of each service node device from the management controller, or obtain a part of the hardware resources through the PCIE switch board. Information, get another piece of information from the management controller.

In an optional implementation manner of this embodiment, the implementation manner of step 803 includes: the management device presents the information of the hardware resource included in the physical resource pool to the user, and receives the selected instruction and the partitioning instruction of the user to establish the logical partition. . The selected instruction is used to indicate a hardware resource selected by the user from the physical resource pool according to the application requirement, and the dividing instruction is used to indicate the logical partition formed by the user to allocate the selected hardware resource and the hardware resource included in the logical partition. . This process is equivalent to providing a logical partition platform to the user, so that the user can divide the logical partition. In the management method provided by the embodiment, the management device and the service node device are connected through the PCIE switch board, and the management device acquires the hardware resource information of each service node device through the PCIE switch board, and constructs the physical resource according to the acquired hardware resource information. The pool, and then the logical partition is established on the physical resource pool, and the PCIE switch board is utilized to implement the physical resource pool for shielding physical topology information at the physical resource level, thereby improving the flexibility of using hardware resources in the physical resource pool, thereby effectively improving the pool. The performance of the entire system.

FIG. 9 is a schematic structural diagram of a management method according to another embodiment of the present invention. As shown in FIG. 9, the method in this embodiment includes:

Step 901: In the case of standby power supply, the management device acquires information about hardware resources of each service node device.

Specifically, the management device can monitor the state of the hardware resources of each service node device through the PCIE switch board, and interact with each service node device through the PCIE switch board to obtain information about the hardware resources of each service node device. Or

The management device may also receive, by using a management controller of each service node device, information about hardware resources of the service node device where the management controller is reported by the management controller.

Step 902: The management device constructs a physical resource pool according to information about hardware resources of each service node device.

Step 903: The management device presents the physical resource pool to the user by using a UI or the like, so that the user performs logical partition division.

Step 904: The user divides the logical partition according to the application requirement.

Step 905: After the logical partition is powered on, the enhanced adaptation layer provided by the management device takes effect, and the hardware platform corresponding to the logical partition is presented to the user.

Optionally, the hardware platform corresponding to each logical partition may include: CPU, 10, MEM, and the like.

Step 906: The user installs an operating system, an intermediate layer, and an upper layer application software on the hardware platform to perform an application.

As shown in the above, the management device and the service node device are connected through the PCIE switch board. The management device obtains the information of the hardware resources of each service node device through the PCIE switch board, and constructs a physical resource pool according to the acquired hardware resource information, and then the physical resource. A logical partition is created on the pool to fully utilize the PCIE switch board to implement a physical resource pool that blocks physical topology information at the physical resource level, that is, no Restricted by the interconnection mode in the prior art, the hardware resources in the physical resource pool can be flexibly allocated, and the logical partition can be dynamically adjusted, thereby effectively improving the performance of the entire system.

Further, after the step 906, the management method shown in FIG. 10 further includes:

Step 907: The management device monitors a load condition of a hardware resource of the logical partition.

Step 908: The management device determines whether the load of the hardware resource included in the logical partition is greater than a preset first load threshold. If the determination result is yes, go to step 909; if the structure is negative, go back to step 907.

Step 909: The management device selects an idle hardware resource from the physical resource pool to be added to the logical partition, and triggers a hot plugging process of the newly added hardware resource.

The newly added hardware resource is of the same type as the hardware resource whose load exceeds the first load threshold. For example, if the hardware resource whose load exceeds the first load threshold is the CPU, the newly added hardware resource is also the CPU resource; The hardware resource of the load threshold is 10, and the newly added hardware resources are also 10, and so on.

Further, an optional implementation of step 908 may be: the management device sends the early warning information. The first load threshold may be an upper threshold in the load threshold, but is not limited thereto. When the second load threshold is preset, excess hardware resources are removed from the logical partition into the physical resource pool. The second load threshold may be a lower threshold in the load threshold. The excess hardware resources, in addition to the excess hardware resources in the logical partition to the physical resource pool, are beneficial to improve the utilization of hardware resources. The lower-utilization hardware resources usually refer to hardware resources whose utilization is less than the preset utilization threshold. The utilization threshold can be set to different values according to actual requirements.

It can be seen from the above that the management method provided in this embodiment, when the logical partition is running, obtains the hardware resources in the logical partition according to the information of the hardware resources, and manages the hardware resources in the logical partition according to the information of the hardware resources. Hot swapping, enabling dynamic management of logical partitions. In combination with the above, the dynamic management of logical partitions can be implemented by an enhanced adaptation layer provided by the management device.

One of ordinary skill in the art can understand that all or part of the steps of the above method embodiments are implemented. The foregoing program may be stored in a computer readable storage medium, and when executed, the program includes the steps of the foregoing method embodiment; and the foregoing storage medium includes: ROM, RAM , a variety of media that can store program code, such as a disk or an optical disk.

Finally, it should be noted that the above embodiments are only for explaining the technical solutions of the present invention, and are not intended to be limiting thereof; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art will understand that The technical solutions described in the foregoing embodiments may be modified, or some or all of the technical features may be equivalently replaced; and the modifications or substitutions do not deviate from the technical solutions of the embodiments of the present invention. range.

Claims

claims

1. A server system, characterized in that it includes: a management device, a high-speed peripheral component interconnecting a PCIE switching board and at least one service node device; the management device communicates with the at least one service node device through the PCIE switching board Each service node device in the connection is connected;

The management device is configured to obtain information about the hardware resources of each service node device through the PCIE switching board, construct a physical resource pool based on the obtained information, and establish logical partitions on the physical resource pool.

2. The server system according to claim 1, wherein the management device is specifically configured to present information about hardware resources included in the physical resource pool to a user, and receive selection instructions and division instructions from the user, To establish the logical partition, the selection instruction is used to instruct the user to select hardware resources from the physical resource pool according to application requirements, and the partition instruction is used to instruct the user to perform operations on the selected hardware resources. The logical partition formed by the allocation and the hardware resources included in the logical partition.

3. The server system according to claim 2, wherein the management device is also used to enable the user to install an operating system, and the hardware platform includes hardware resources included in the logical partition.

4. The server system according to claim 2 or 3, wherein the management device is further configured to issue an early warning message when the load of the hardware resources included in the logical partition is greater than the preset first load threshold.

5. The server system according to claim 2 or 3, characterized in that, when a load threshold is reached, idle hardware resources are selected from the physical resource pool and added to the logical partition, and trigger the newly added Hot swap process of hardware resources; or when the load threshold is reached, remove excess hardware resources from the logical partition to the physical resource pool, and the excess hardware resources are not utilized in the logical partition. Or hardware resources whose utilization is less than the preset utilization threshold.

6. The server system according to any one of claims 1 to 5, characterized in that, the management device is specifically connected to the hardware resources of each service node device through the PCIE switching board; The management device is specifically configured to monitor the status of the hardware resources of each service node device through the PCIE switching board, and obtain information about the hardware resources of each service node device through the PCIE switching board.

7. The server system according to any one of claims 1 to 5, characterized in that, the management device is connected to the hardware resources of each service node device through the PCIE switching board, and is connected to the hardware resources of each service node device through the PCIE switching board. The switching board is connected to the management controller of each service node device;

The management device is specifically configured to monitor the status of the hardware resources of each service node device through the PCIE switching board, and obtain the information of the hardware resources of each service node device through the PCIE switching board; or The corresponding relationship between the second ports is to obtain the information of the hardware resources of each service node device from the management controller of each service node device;

Wherein, the management controller of each service node device is used to monitor the status of the hardware resources of each service node device, and the first port is connected to the hardware resources of each service node device on the PCIE switching board. The second port is a port on the PCIE switching board connected to the management controller of each service node device.

8. The server system according to any one of claims 1 to 5, characterized in that, the management device is specifically connected to the hardware resources of each service node device through the PCIE switching board, and is directly connected to the Management controller connections for each service node device;

Wherein, the management controller of each service node device is used to monitor the status of the hardware resources of each service node device, and the first port is connected to the hardware resources of each service node device on the PCIE switching board. The second port is a port on the management device connected to the management controller of each service node device.

9. A management device, characterized by including: a monitoring module, connected to at least one service node device through a high-speed peripheral component interconnection PCIE switching board, and used to obtain information about the hardware resources of each service node device in the at least one service node device through the PCIE switching board, Construct a physical resource pool based on the obtained information;

A partition establishment module is used to establish logical partitions on the physical resource pool.

10. The management device according to claim 9, wherein the partition establishment module is specifically configured to present the information of the hardware resources included in the physical resource pool to the user, and receive the user's selection instructions and partitions. The instruction is to establish the logical partition. The selection instruction is used to instruct the user to select hardware resources from the physical resource pool according to application requirements. The partition instruction describes the hardware resources included in the logical partition.

11. The management device according to claim 10, wherein the partition establishment module is further configured to present the hardware platform corresponding to the logical partition to the user after the logical partition is powered on, so that the The user installs an operating system, and the hardware platform includes hardware resources included in the logical partition.

12. The management device according to claim 10 or 11, further comprising: an early warning module, configured to obtain in the monitoring module that the load of the hardware resources included in the logical partition is greater than the preset first load gate For a limited time, an early warning message will be issued.

13. The management device according to claim 10 or 11, characterized in that,

The partition establishment module is also configured to select idle hardware resources from the physical resource pool and add them to the physical resource pool when the monitoring module obtains that the load of the hardware resources included in the logical partition is greater than the preset first load threshold. logical partition, and trigger the hot-swapping process of the newly added hardware resources; or

The partition establishment module is also configured to remove excess hardware resources from the logical partition to the physical partition when the monitoring module obtains that the load of the hardware resources included in the logical partition is less than a preset second load threshold. In the resource pool, the excess hardware resources are hardware resources that are not utilized in the logical partition or have a utilization rate less than a preset utilization threshold.

14. The management device according to any one of claims 9-13, characterized in that the monitoring module includes:

A first monitoring unit is connected to the hardware resources of each service node device through the PCIE switching board, and is used to monitor the hardware resources of each service node device through the PCIE switching board. Status, obtain the information of the hardware resources of each service node device through the PCIE switching board; The construction unit is used to construct the information of the hardware resources of each service node according to the information obtained by the first monitoring unit. Describe the physical resource pool.

15. The management device according to claim 14, characterized in that the monitoring module further includes:

The second monitoring unit is connected to the management controller of each service node device through the PCIE switching board or directly, and is used to connect the first port and the second port of each service node device according to the storage corresponding relationship, obtaining information about the hardware resources of each service node device from the management controller of each service node device;

The building unit is further configured to obtain information about the hardware resources of each service node according to the information obtained by the first monitoring unit and/or obtain information about the hardware resources of each service node device obtained by the second monitoring unit. information to construct the physical resource pool;

Wherein, the management controller of each service node device is used to monitor the status of the hardware resources of each service node device; the first port is connected to the hardware resources of each service node device on the PCIE switching board The second port is a port on the PCIE switching board or the management device connected to the management controller of each service node device.

16. A management method characterized by including:

The management device obtains information about the hardware resources of each service node device in at least one service node device through the high-speed peripheral component interconnection PCIE switching board;

The management device constructs a physical resource pool based on the obtained information;

The management device establishes logical partitions on the physical resource pool;

Wherein, the management device is connected to each service node device through the PCIE switching board.

17. The management method according to claim 16, wherein the management device establishing a logical partition on the physical resource pool includes: receiving a selection instruction and a division instruction from the user to establish the logical partition, so The selection instruction is used to instruct the user to select hardware resources from the physical resource pool according to application requirements, and the partition instruction is used to instruct the user to allocate the logical partitions formed by the selected hardware resources. The hardware resources included in the logical partition.

18. The management method according to claim 17, further comprising: after the logical partition is powered on, the management device presents the hardware platform corresponding to the logical partition to the user, so that all The user installs an operating system, and the hardware platform includes hardware resources included in the logical partition.

19. The management method according to claim 17 or 18, further comprising: issuing an early warning message when the threshold is reached.

20. The management method according to claim 17 or 18, further comprising: at a threshold, selecting idle hardware resources from the physical resource pool to add to the logical partition, and triggering the new addition. hot-swapping process of hardware resources; or at a threshold, remove excess hardware resources from the logical partition to the physical resource pool, and the excess hardware resources are not utilized in the logical partition or Hardware resources whose utilization is less than the preset utilization threshold.

21. The server method according to any one of claims 16 to 20, characterized in that: the management device is connected to the hardware resources of each service node device through the PCIE switching board; the management device is connected through a high-speed The peripheral component interconnection PCIE switching board obtains the information of the hardware resources of each service node device in at least one service node device, including:

The management device monitors the status of the hardware resources of each service node device through the PCIE switching board, and obtains the information of the hardware resources of each service node device through the PCIE switching board.

22. The management method according to any one of claims 16 to 20, characterized in that: the management device is connected to the hardware resources of each service node device through the PCIE switching board, and is connected to the hardware resources of each service node device through the PCIE switching board. The board is connected to the management controller of each service node device;

The management device obtains the information of the hardware resources of each service node device in at least one service node device through the high-speed peripheral component interconnection PCIE switching board, including:

The management device monitors the status of the hardware resources of each service node device through the PCIE switching board, and obtains the information of the hardware resources of each service node device through the PCIE switching board; or The management device obtains the information of each service node device from the management controller of each service node device according to the stored correspondence between the first port and the second port connected to each service node device. Hardware resource information;

Wherein, the management controller of each service node device is used to monitor the status of the hardware resources of each service node device; the first port is connected to the hardware resources of each service node device on the PCIE switching board The second port is a port on the PCIE switching board connected to the management controller of each service node device.

23. The management method according to any one of claims 16 to 20, characterized in that the management device is connected to the hardware resources of each service node device through the PCIE switching board, and is directly connected to the hardware resources of each service node device. Management controller connection for each service node device;

The management device monitors the status of the hardware resources of each service node device through the PCIE switching board, and obtains the information of the hardware resources of each service node device through the PCIE switching board, or

The management device obtains the information of each service node device from the management controller of each service node device according to the stored correspondence between the first port and the second port connected to each service node device. Hardware resource information;

Wherein, the management controller of each service node device is used to monitor the status of the hardware resources of each service node device; the first port is connected to the hardware resources of each service node device on the PCIE switching board The second port is a port on the management device connected to the management controller of each service node device.