TW201523277A - Server system and node detection method - Google Patents

Abstract

A node detection method adapted for a server system, the server system includes an admin node and at least one device under test. The step processes of the method are described below. The device transmits a request to the admin node. The admin node configures an IP address to the device according to the request, and establishes an address table. The admin node determines whether the device is a node according to the address table. If the determination is YES, the admin node transmits a WOL signal to the node. The node executes a PXE mode according to the WOL signal such that obtains the hardware information of the node. The admin node determines the applicable type of the node according to the hardware information of the node. Wherein, the applicable type of the node is at least one of the storage node, the compute node, and the service node.

Description

Server system and node detection method

The invention relates to a server system and a node detecting method, and in particular to a server system and a node detecting method for determining a type applicable to a node.

With the development of technology, computers around the world can be connected through the Internet. A computer can exchange data and access data with another computer through a network connection. On the client and server system architecture, the client and server communicate through the network.

In general, a server system can be configured with multiple nodes, and each node runs multiple virtual machines (VMs) at the same time, thereby providing each user with a separate operating environment. Moreover, each node can be regarded as a separate computer, that is, each node has a memory, a storage space, a computing capability, and a network connection function. Therefore, each node can run a separate operating system, and each node can also communicate and transmit data through a network switch.

After the server system architecture is completed, the nodes in it must be deployed, that is, the operating system and software package information required for installing each node. Firmware package information to enable the server system to operate and provide services to the user. However, when a device to be tested appears in the server system, the user cannot directly know whether the device under test is a node or a network switch, and the device to be tested cannot be effectively detected and utilized.

In view of the above problems, the present disclosure proposes a server system and node detection method, which can automatically detect all nodes in the server system and obtain hardware information of the nodes, and can apply the types to the nodes. Judge.

According to a node detection method in an embodiment of the disclosure, the node detection method is applicable to a server system (for example, a cabinet type data center), wherein the server system includes a management node and at least one device to be tested. The flow of steps of the node detection method is as follows. The request is sent by the device under test to the management node. The management node configures the Internet Protocol address to the device under test according to the request, and establishes an address table according to the Internet Protocol address. The management node determines whether the device under test is a node according to the address table. If the management node determines that the device under test is a node, the network wake-up signal is transmitted to the node. The node performs a pre-boot execution environment mode according to the network wake-up signal, and obtains hardware information of the node accordingly. The management node determines the applicable type of the node according to the hardware information of the node, wherein the applicable type of the node is at least one of the storage node, the computing node and the service node.

In an embodiment, the management node determines the test according to the address table. In the step of whether the device is a node, the step flow described below is further included. Obtain a record indicating the Internet Protocol address and the hardware access control address (MAC address) of the device under test that sends the request to the management node. It is judged whether the record exists in the first address table or the second address table in the address table. If it is determined that the record already exists in the first address table or the second address table, then it is determined whether there is another record of the device under test, and when the determination is yes, the subsequent execution determines whether another record exists in the first The step of the address table or the second address table until it is determined that the record does not exist in the first address table or the second address table. If it is determined that the record does not exist in the first address table or the second address table, then it is determined whether the device to be tested corresponding to the record is a node. If it is determined that the device to be tested corresponding to the record is not a node, the record is updated to the first address table, and the step subsequent to determining whether there is another record of the device to be tested is performed. If it is determined that the device to be tested corresponding to the record is a node, the record is updated to the second address table, and the step after transmitting the network wake-up signal to the node is continued.

In the above-mentioned embodiment, in the step of determining whether the device to be tested corresponding to the record is a node, the step flow described below is further included. The inquiry signal is sent to the device under test according to the Internet Protocol address corresponding to the record. Determine whether the device under test has responded to the inquiry signal within a preset time. If it is determined that the device under test has not responded to the inquiry signal after a preset time, the device to be tested is a node. If it is determined that the to-be-tested device responds to the inquiry signal within the preset time, then it is determined whether the response signal that is responded has a special string. If it is determined that the response signal does not have a special string, the device under test is a node. If judged back The signal has a special string, and the device under test is a network switch.

In an embodiment, if the management node determines that the device to be tested is a node, the management node can selectively control the power shutdown, forced power shutdown, light indication, and system reset of the hardware in the node through the control signal. Startup, where the control signal is used to indicate the identification code of the hardware in the node.

In an embodiment, in the step of the node performing the pre-boot execution environment mode according to the network wake-up signal to obtain the hardware information of the node, the hardware information of the node is stored in the database. The database communication is connected between the device under test and the management node to store the hardware information of the node and the result of the applicable type of the node determined by the management node, or provide the hardware information of the node to the management node.

According to a server system according to an embodiment of the present disclosure, the server system includes a device to be tested and a management node, wherein the management node communicates with the device to be tested. The management node includes a server, a processing module, a power control module, and an analysis module. The server configures the Internet Protocol address to the device under test according to the request sent by the device under test, and establishes an address table according to the Internet Protocol address. The processing module is configured to determine, according to the address table, whether the device to be tested is a node. The power control module is configured to transmit a network wake-up signal to the node when the processing module determines that the device to be tested is a node, so that the node performs a pre-boot execution environment mode according to the network wake-up signal and obtains hardware information of the node according to the network wake-up signal. . The analysis module is configured to determine the applicable type of the node according to the hardware information of the node, wherein the applicable type of the node is at least a storage node, a computing node and a service node. One.

In an embodiment, when the processing module determines, according to the address table, whether the device under test is a node, the method further includes obtaining a record, where the record is used to indicate an Internet Protocol address of the device to be tested that sends the request to the management node. Hardware access control address (MAC address). Next, the processing module determines whether the record exists in the first address table or the second address table in the address table. If the processing module determines that the record already exists in the first address table or the second address table, the processing module then determines whether there is another record of the device under test, and when the determination is yes, the processing module continues to execute. It is judged whether another record exists in the first address table or the second address table until it is determined that the record does not exist in the first address table or the second address table. If the processing module determines that the record does not exist in the first address table or the second address table, the processing module then determines whether the device to be tested corresponding to the record is a node. If the processing module determines that the device under test corresponding to the record is not a node, the record is updated to the first address table. If the processing module determines that the device to be tested corresponding to the record is a node, the record is updated to the second address table.

According to the above embodiment, when the processing module determines whether the device to be tested corresponding to the record is a node, the processing module sends an inquiry signal to the device to be tested according to the Internet Protocol address corresponding to the record, and determines that the device is to be tested. Whether the measuring device responds to the inquiry signal within the preset time. If the processing module determines that the device under test has not responded to the inquiry signal after a preset time, the device to be tested is a node. If the processing module determines that the device to be tested responds to the inquiry signal within the preset time, the processing module then determines whether the response signal that is responded has a special word. string. If the processing module determines that the response signal does not have a special string, the device under test is a node. If the processing module determines that the response signal has a special string, the device under test is a network switch.

In an embodiment, if the processing module determines that the device to be tested is a node, the processing module can selectively control the power shutdown, forced power shutdown, light indication, and system of the hardware in the node through the control signal. The reset is initiated, wherein the control signal is used to indicate the identification code of the hardware in the node.

In an embodiment, the server system further includes a database, and the database is communicatively coupled between the device under test and the management node. This database is used to store the result of the node's hardware information and the type of node determined by the analysis module, or to provide hardware information of the node to the analysis module.

In summary, the disclosure provides a server system and a node detection method, which establishes an address table, so that the management node can determine whether the device under test is a node according to the address table, and is determined by the device to be tested as At the time of the node, the device under test is configured to execute a pre-boot execution environment mode, so that the management node can determine the applicable type of the node according to the obtained hardware information of the node. In addition, when the management node determines whether the device under test is a node, it is further possible to determine whether the signal to be responded has a special string through a query response mechanism, and determine whether the device to be tested is a node or a network switch. .

The above description of the disclosure and the following description of the embodiments of the present invention are intended to illustrate and explain the spirit and principles of the invention, and to provide further explanation of the scope of the invention.

1‧‧‧Server system

10‧‧‧Management node

100‧‧‧Server

102‧‧‧Processing module

104‧‧‧Power Control Module

106‧‧‧Analysis module

12‧‧‧Device under test

14‧‧‧Database

S200~S212, S300~S312, S400~S408‧‧‧Step process

1 is a functional block diagram of a server system in accordance with an embodiment of the present disclosure.

2 is a flow chart showing the steps of a node detecting method according to an embodiment of the present disclosure.

Fig. 3 is a flow chart showing the detailed steps of step S204 according to Fig. 2.

Fig. 4 is a flow chart showing the detailed steps of step S308 according to Fig. 3.

The detailed features and advantages of the present invention are set forth in the Detailed Description of the Detailed Description of the <RTIgt; </ RTI> <RTIgt; </ RTI> </ RTI> </ RTI> <RTIgt; The objects and advantages associated with the present invention can be readily understood by those skilled in the art. The following examples are intended to describe the present invention in further detail, but are not intended to limit the scope of the invention.

Please refer to FIG. 1. FIG. 1 is a functional block diagram of a server system according to an embodiment of the present disclosure. As shown in FIG. 1, the server system 1 mainly includes a management node 10, a device under test 12, and a database 14, wherein the management node 10, the device under test 12, and the database 14 are in communication with each other. In addition, the admin node 10 further includes a server 100, a processing module 102, and an electric The source control module 104 and the analysis module 106. In practice, the server system 1 can be a rack-type data center, but not limited thereto. The function modules of each part in the server system 1 will be described in detail below.

The server 100 is configured to configure an internet protocol address (IP address) to the device under test 12 according to a request sent by the device under test 12, and according to the internet protocol address. Create an address table. In practice, the server 100 can be a dynamic host configuration protocol server (DHCP server), and the DHCP server automatically allocates IP and related network parameters to the client to provide the client. Automatically set their network with the parameters provided by the DHCP server. Therefore, the request sent by the device under test 12 to the server 100 will be a DHCP request. In addition, the address table can be a MAC/IP mapping table, and the MAC/IP mapping table can be obtained through a command line interface tool (CLI tool, also called a command line tool).

The processing module 102 is configured to determine whether the device under test 12 is a node according to the address table. The power control module 104 is configured to transmit a network wake-up signal to the node (ie, the device under test 12) when the processing module 102 determines that the device under test 12 is a node, so that the node (ie, the device under test 12) The power on self test (POST) mode and the preboot execution environment (PXE) can be performed according to the network wake-up signal. Mode, and obtain the hardware information of this node when executing the pre-boot execution environment mode. In practice, the above-mentioned network wake-up signal is implemented by Wake-on-LAN (WOL) technology, which can make the node that has entered the sleep state or the shutdown state. Through the command of the management node 10, the node wakes up from the sleep state, returns to the operational state, or transitions from the shutdown state to the power-on state.

It should be noted that the node belonging to the device under test 12 in the present invention needs the first startup item of the Boot Option Priorities of the interface of the basic input output system (BIOS). Boot Option #1) is preset to PXE mode. In the actual operation, when the processing module 102 determines that the device under test 12 is a node, the processing module 102 can selectively control the power shutdown, forced power shutdown, and light of the hardware in the node through the control signal. The number indication and system reset start, wherein the control signal is used to indicate the identification code of the hardware in the node. In practice, the control signal may be through a universally unique identifier (UUID) or a globally unique identifier (GUID), which is not limited herein.

In actual operation, the control signal is a form feed character (FF) including 6 bytes, a GUID of 6 repetitions (16 bytes per GUID), and a packet payload of the MAC address. . Thereby, the network interface controller (NIC, also known as the network card) in the device under test 12 belonging to the node supervises this special The GUID field of the payload in the special package, and perform the corresponding operation. For example, if the NIC determines that the GUID field is "E4FB6490-1854-4176-959F-FAFCF3E50152", the light in the node will be turned on; if the NIC determines that the GUID field is "8AB61CE6-01A4-450a- 80E8-287416BC2835", the signal in the node will be turned off; if the NIC determines that the GUID field is "BA3C6C53-F302-4d2e-ACF7-427FF6754902", the power of the node will be turned on or off; if the NIC determines the GUID column If the position is "A994BEEC-5F2D-4e14-AD18-1982C1D44561", the node will be forced to power off; if the NIC determines that the GUID field is "1385BC34-BEAD-4bfa-BC29-84BF9345C880", the node will be reset. The reset system 12 enables the device under test 12 belonging to the node to operate normally without a baseboard management controller (BMC).

The analysis module 106 is configured to determine the applicable type of the node according to the hardware information of the node, where the applicable type of the node is at least a storage node, a compute node, and a service node. One, that is, the applicable type of this node may also be a generic node having the function of the computing node and the function of the service node. Generally, the storage node is used to provide storage space for the user of the client and to store the image file and configuration file of the virtual machine (VM), and the computing node is used to provide multiple virtual machines to the client. Make The service node is used to provide services to the user of the client.

The database 14 is configured to store the result of the applicable type of the node determined by the hardware information and analysis module 106 of the device 12 to be tested belonging to the node, or provide the hardware information of the node to the analysis module 106. . In practice, the database 14 can be a Berkeley database that can hold any type of key/value pair. In other words, the data storage method of the Berkeley database is (key). , value) for it.

It is to be noted that, because the node in the embodiment of the present invention can perform the power on/off, forced shutdown, restart, or signal control of the node through the management node 10, and determine the type applicable to the node, and the management node 10 can automatically The nodes in the embodiment of the present invention can detect the baseboard management controller (BMC), which can save costs. In one embodiment, the node has an Advanced Configuration and Power Interface (ACPI) and a network interface controller (NIC, also known as a network card), so that the node can be controlled In the management node 10.

In order to more clearly illustrate the actual operation mode of the server system 1 and the node detecting method of the present invention, please refer to FIG. 1 and FIG. 2 together. FIG. 2 is a node detecting method according to an embodiment of the present disclosure. Step flow chart. As shown in FIG. 2, in step S200, the device under test 12 transmits a request to the management node 10. In step S202, the management node 10 will be based on this request. The Internet Protocol address is configured to be sent to the device under test 12, and an address table is established according to the Internet Protocol address.

In step S204, the management node 10 determines whether the device under test 12 is a node according to the address table described above. If the management node 10 determines that the device under test 12 is a node, step S208 is performed; if the management node 10 determines that the device under test 12 is not a node, the device under test 12 is a network switch ( That is, in step S206), the network switch is used to enable each node in the Internet to communicate and transmit data.

In step S208, the management node 10 transmits a network wake-up signal to the device under test 12 determined to be a node. Next, in step S210, it is determined that the device under test 12 of the node performs the pre-boot execution environment mode according to the above-mentioned network wake-up signal, and obtains the hardware information of the node accordingly. Finally, in step S212, the management node 10 determines the applicable type of the node according to the hardware information of the node.

Please refer to FIG. 3, which is a flow chart of detailed steps according to step S204 of FIG. 2. As shown in FIG. 3, in the step of determining whether the device under test 12 is a node according to the address table according to the address table (ie, step S204), the detailed step flow described below may be further included. In step S300, the management node 10 first obtains a record. The record is used to indicate the Internet Protocol address and the MAC address of the device under test 12 that sends the request to the management node 10.

In step S302, the management node 10 determines whether the record is It exists in the first address table or the second address table in the address table. If the management node 10 determines that the record already exists in the first address table or the second address table in the address table, step S304 is performed; if the management node 10 determines that the record does not exist in the address table. The first address table or the second address table is executed in step S308.

In step S304, the management node 10 determines whether there is another record of another device under test. If the management node 10 determines that there is another record, step S302 is re-executed until the management node 10 determines that the record does not exist in the first address table or the second address table. If the management node 10 determines that there is no more new record, the sleep of step S306 is performed for a preset time until the management node 10 re-establishes another address table, and another address table is established at the management node 10. When step S300 is re-executed.

In step S308, the management node 10 determines whether the device under test 12 corresponding to the record is a node. If the management node 10 determines that the device under test 12 corresponding to the record is not a node (ie, the device under test 12 is determined to be a network switch), step S310 is performed; if the management node 10 determines that the record corresponds to the If the measuring device 12 is a node, step S312 is performed. In step S310, since the management node 10 has determined that the device under test 12 is a network switch, the management node 10 updates the record corresponding to the network switch to the first address table, and the subsequent steps S304. In step S312, since the management node 10 has judged that the device under test 12 is a node, the management node 10 updates the record corresponding to the node to the second address table, and simultaneously proceeds to steps S208 and S304.

In practice, the first address table is a location table for storing records of the network switch, and the second address table is a location table for storing the records of the nodes, that is, the first address table. It is a network switch address table, and the second address table is a node address table, and the first address table and the second address table are updated as the number of devices 12 under test increases.

Please refer to FIG. 4, which is a flow chart of detailed steps according to step S308 of FIG. As shown in FIG. 4, in the step of determining whether the device to be tested corresponding to the record is a node (ie, step S308), the detailed step flow described below may be further included. In step S400, the management node 10 sends an inquiry signal to the device 12 to be tested according to the Internet Protocol address corresponding to the record. In practice, the query signal can be a simple network management protocol command (SNMP command), and the simple network management protocol command can be recognized by the network switch, that is, the network switch. This simple network management protocol directive can be supported.

In step S402, the management node 10 determines whether the device under test 12 has responded to the inquiry signal for a predetermined period of time. If the management node 10 determines that the device under test 12 does not respond to the inquiry signal in the preset time period, it indicates that the device under test 12 is a kind of node (ie, step S404); if the management node 10 determines that the device under test 12 is If the inquiry signal has been answered in the preset time period, step S406 is performed.

In step S406, the management node 10 determines the response back. Whether the signal should have a special string. If the management node 10 determines that the response signal does not have a special string, it indicates that the device under test 12 is a kind of node (ie, step S404); if the management node 10 determines that the response signal has a special string, It means that the device under test 12 is a network switch (ie, step S408). In practice, the response signal can be a system description with a "switch" string, and the "switch" string is a special string, but not limited to this.

In summary, the embodiment of the present invention provides a server system and a node detection method, which establishes an address table, so that the management node can determine whether the device under test is a node according to the address table, and is When it is determined as a node, the device under test is configured to execute a pre-boot execution environment mode, so that the management node can determine the applicable type of the node according to the obtained hardware information of the node. In addition, when the management node determines whether the device under test is a node, it is further possible to determine whether the signal to be responded has a special string through a query response mechanism, and determine whether the device to be tested is a node or a network switch. . Thereby, the server system and the node detection method of the present invention can automatically detect all the nodes in the server system, and obtain the hardware information of the nodes by controlling the power on of the nodes, and then according to the hardware. The information is used to judge the applicable types of these nodes, and the node does not need to be provided with a backplane management controller, which can save cost and is very practical.

Although the present invention has been disclosed above in the above embodiments, it is not intended to limit the invention. Without departing from the spirit and scope of the invention Both the movement and the retouching are within the scope of patent protection of the present invention. Please refer to the attached patent application for the scope of protection defined by the present invention.

Claims (10)

A node detection method is applicable to a server system, the server system includes a management node and at least one device to be tested, and the node detection method includes: sending, by the device under test, a request to the management node; the management The node configures an internet protocol address to the device under test according to the request, and establishes an address table according to the internet protocol address; the management node determines, according to the address table, whether the device to be tested is one a node; if the management node determines that the device to be tested is the node, transmitting a network wake-up signal to the node; the node performs a pre-launch execution environment mode according to the network wake-up signal, and obtains the node according to the The hardware information is determined by the management node according to the hardware information of the node, wherein the applicable type of the node is at least one of a storage node, a computing node, and a service node. The method for detecting a node according to claim 1, wherein the step of determining, by the management node, whether the device to be tested is the node according to the address table further comprises: obtaining a record, the record is used to indicate the management The internet protocol address of the device under test to which the node sends the request; Determining whether the record exists in a first address table or a second address table in the address table; if it is determined that the record already exists in the first address table or the second address table, Then, it is determined whether there is another record of the other device to be tested, and when the determination is yes, the step of determining whether another record exists in the first address table or the second address table is continued; Determining that the record does not exist in the first address table or the second address table; if it is determined that the record does not exist in the first address table or the second address table, then determining that the record corresponds to Whether the device to be tested is the node; if it is determined that the device to be tested corresponding to the record is not the node, the record is updated to the first address table, and the subsequent execution is performed to determine whether there is another Another step after the recording of the device; and if it is determined that the device to be tested corresponding to the record is the node, updating the record to the second address table, and subsequently performing the transmission of the network wake-up signal Give the step after the node. The method for detecting a node according to claim 2, wherein the step of determining whether the device to be tested corresponding to the record is the node further comprises: according to the internet protocol address corresponding to the record, The device under test sends an inquiry signal; and determines whether the device to be tested has a response signal within a preset time; If it is determined that the device under test has not returned the inquiry signal after the preset time, the device to be tested is the node; if it is determined that the device to be tested should be queried within the preset time, then the signal is determined, and then it is determined Whether the response signal of the response has a special string; if it is determined that the response signal does not have the special string, the device to be tested is the node; and if it is determined that the response signal has the special string, then The device under test is a network switch. The node detection method of claim 1, wherein if the management node determines that the device to be tested is the node, the management node can selectively control the power shutdown of the hardware in the node through a control signal. The forced power off, the light indication, and the system reset start, wherein the control signal is used to indicate the identification code of the hardware in the node. The node detecting method according to claim 1, wherein in the step of the node executing the pre-boot execution environment mode according to the network wake-up signal to obtain the hardware information of the node, the hardware information of the node is stored in the In a database, the database is connected between the device under test and the management node to store the hardware information of the node and the result of the applicable type of the node determined by the management node, or provide the node Hardware information to the management node. A server system comprising: a device to be tested; and a management node, the communication device is connected to the device under test, the management node includes: a server, configured to configure an internet protocol address to be tested according to a request sent by the device under test a device, and establishing an address table according to the Internet Protocol address; a processing module for determining whether the device to be tested is a node according to the address table; and a power control module for processing The module determines that the device to be tested is a node and transmits a network wake-up signal to the node, so that the node performs a pre-launch execution environment mode according to the network wake-up signal and obtains hardware information of the node according to the network wake-up signal. And an analysis module, configured to determine a suitable type of the node according to the hardware information of the node, wherein the applicable type of the node is at least one of a storage node, a computing node, and a service node. The server system of claim 6, wherein the processing module determines whether the device to be tested is the node according to the address table, and further includes obtaining a record for indicating to send to the management node. The requested internet protocol address of the device under test, and then the processing module determines whether the record exists in a first address table or a second address table in the address table, if The processing module determines that the record already exists in the first address table or the second address table, and the processing module then determines whether there is another record of the other device to be tested, and determines that it is When the processing module is connected Performing to determine whether another record exists in the first address table or the second address table until it is determined that the record does not exist in the first address table or the second address table, if the processing module Determining whether the record does not exist in the first address table or the second address table, the processing module then determines whether the device to be tested corresponding to the record is the node, and if the processing module determines the If the device to be tested corresponding to the record is not the node, the record is updated to the first address table, and if the processing module determines that the device to be tested corresponding to the record is the node, the record is recorded. Update to the second address table. The server system of claim 7, wherein the processing module determines whether the device to be tested corresponding to the record is the node, and the processing module further determines the internet protocol bit corresponding to the record. Sending an inquiry signal to the device to be tested, and determining whether the device under test has a response signal for a predetermined time, if the processing module determines that the device to be tested has not returned after the preset time If the signal is to be queried, the device to be tested is the node. If the processing module determines that the device to be tested is back to the inquiry signal within the preset time, the processing module then determines whether the response signal is responsive. Having a special string, if the processing module determines that the response signal does not have the special string, the device to be tested is the node, and if the processing module determines that the response signal has the special string, The device under test is a network switch. The server system of claim 6, wherein the processing module is further capable of transmitting a control signal if the processing module determines that the device to be tested is the node The power supply shutdown, forced power shutdown, light indication, and system reset startup of the hardware in the node are selectively controlled, wherein the control signal is used to indicate the identification code of the hardware in the node. The server system of claim 6, wherein the server system further comprises a database, the database is communicatively connected between the device under test and the management node, and the database is used to store the hardware of the node. The information and the result of the applicable type of the node determined by the analysis module, or the hardware information of the node is provided to the analysis module.