CN105912431A - Reboot testing method of server, server, control device and system - Google Patents

Abstract

The invention provides a server reboot test method, server, controller and system, the method comprising: through a switch, establishing mutual communication between the server and the controller; the server receiving the boot request sent by the controller, and starting the boot; judging whether there is Ispci-tmp file, if yes, read the device information, write the device information into the Ispci‑$count file, and compare whether the Ispci‑tmp file is consistent with the Ispci‑$count file, otherwise, generate the Ispci‑tmp file for the server’s device information ;When the Ispci-tmp file is consistent with the Ispci-$count file, send the start-up completion information; create the gpu.txt file and the server.txt file; receive the shutdown request sent by the controller, and perform the shutdown operation to realize the server stability test automation.

Description

Server reboot test method, server, controller and system

technical field

The invention relates to the technical field of server applications, in particular to a server reboot test method, server, controller and system.

Background technique

With the continuous development of cloud computing services, the requirements for server stability are getting higher and higher. A kind of important method of server stability test is reboot test at present.

The reboot test method of the existing server is mainly to install the reboot script on each server node, manually connect the power supply to each server node, manually start each server node, reboot the script to run, and check whether the boot process is normal, and then test the server node To perform a shutdown operation, it is necessary to manually disconnect the power of each server node in turn, that is, the existing reboot test method can only be completed by manual participation, and the server stability test cannot be performed automatically.

Contents of the invention

The embodiment of the present invention provides a server reboot test method, a server, a controller and a system, so as to realize the automation of the server stability test.

The server reboot test method establishes communication between the server and the controller through the switch; it also includes:

When the server receives the boot request sent by the controller, it starts the boot;

The server judges whether the Ispci-tmp file exists, if yes, reads the device information in the server, writes the device information into the Ispci-$count file, and compares whether the Ispci-tmp file is consistent with the Ispci-$count file, otherwise , generate the Ispci-tmp file for the device information in the server;

When the Ispci-tmp file is consistent with the Ispci-$count file, send startup completion information to the controller;

Create gpu.txt file and server.txt file;

Receive the shutdown request sent by the controller, and perform the shutdown operation.

Preferably, the establishment of mutual communication between the server and the controller through the switch includes:

The server is connected to the switch through the OS network and the BMC network;

The controller is connected to the switch through the OS network.

Preferably, the above method further includes: setting a first counter count in the server;

After the startup, it further includes: the server judges whether there is a count file, if yes, then stores the first counter count+1 in the count file; otherwise, starts the first counter count, and The first counter count+1, generate a count file, and write the first counter count into the startup item of the server.

Preferably, the server is a Pcie-Switch server, and the Pcie-Switch server includes: a resource server and a server end, wherein a retimer card is inserted in the server end, through which the retimer card and the MiniSASHD cable are connected to the resource server connect;

It further includes: setting the start sequence;

The booting includes: sequentially starting the resource server and the server end according to the set startup sequence.

The server reboot test method is applied to the controller, and the second counter count is set in the controller, and the detection threshold is set; it also includes:

M1. The controller initializes the second counter count;

M2, receiving the startup completion information sent by the server, judging whether the count of the second counter count is less than the detection threshold, if yes, then detecting whether there are gpu.txt files and server.txt files in the server, and if so, calling the server's Shutdown function, control server shutdown;

M3. Send a start-up request to the server, call the start-up function of the server, control the start-up of the server, and set the second counter count+1, and execute M2.

Preferably, the above method further includes: clearing operating system logs in the server.

Preferably, the server is a Pcie-Switch server, and the Pcie-Switch server includes: a resource server and a server end, wherein a retimer card is inserted in the server end, through which the retimer card and the MiniSASHD cable are connected to the resource server connect;

The controlling the shutdown of the server includes: sequentially controlling the shutdown of the server end and the shutdown of the resource server;

The control server startup includes: sequentially controlling the startup of the resource server and the startup of the server.

The server applied to any of the server reboot test methods described above communicates with the controller of the peripheral device through the switch of the peripheral device, including: a switch unit, a first judgment unit, a read and write unit, and a generation unit ,in,

The switch unit is configured to perform a power-on start when receiving a power-on request sent by the controller of the peripheral device, and trigger the first judging unit to perform a power-off operation when receiving a power-off request sent by the controller of the peripheral device ;

The first judging unit is used to judge whether there is an Ispci-tmp file when receiving the trigger of the boot unit, and if so, trigger the read and write unit; and compare the Ispci-tmp file and Ispci-$ Whether the count file is consistent, otherwise, trigger the generation unit;

The read and write unit is used to read the device information in the server, write the device information into the Ispci-$count file, and when the Ispci-tmp file is consistent with the Ispci-$count file, send Start the completion information to the controller of the peripheral, and create the gpu.txt file and the server.txt file;

The generating unit is configured to generate an Ispci-tmp file for each device information.

Preferably, the above-mentioned server is connected to the switch of the peripheral device through the OS network and the BMC network.

Preferably, the above server further includes: a second judging unit and a first counter, wherein,

The second judging unit is used to judge whether there is a count file, and if so, trigger the first counter; otherwise, start the first counter, generate a count file, and write the first counter into the server to boot startup item;

The first counter is used to count the number of startup times of the switch unit, and count+1 is performed when the switch unit is started up, and the number of startup times is stored in the count file.

Preferably, the above-mentioned server is a Pcie-Switch server, and the Pcie-Switch server includes: a resource server and a server end, wherein a retimer card is inserted in the server end, through which the retimer card and the MiniSASHD cable are connected to the resource server connect.

The controller applied to any of the server reboot testing methods described above, including: a setting unit, a second counter, a detection unit and a call control unit, wherein,

The setting unit is used to set the detection threshold;

The detection unit is used to judge whether the count of the second counter is less than the detection threshold set by the setting unit, if yes, then detect whether there are gpu.txt files and server.txt files in the server of the peripheral device, if yes , then trigger the calling control unit;

The calling control unit is configured to call the shutdown function of the server of the peripheral device when receiving the trigger of the detection unit, control the server of the peripheral device to shut down, send a startup request to the server of the peripheral device, and call the server of the peripheral device The booting function controls the booting of the server of the peripheral device, and adds 1 to the count of the second counter.

The server reboot test system includes: at least one of the above-mentioned any kind of server, a switch and any of the above-mentioned controllers, wherein,

The at least one server and the controller are respectively connected to the switch.

The embodiment of the present invention provides a server reboot test method, server, controller and system. In the method, the mutual communication between the server and the controller is established through a switch; when the server receives the boot request sent by the controller, it starts the boot; The server judges whether the Ispci-tmp file exists, if yes, reads the device information in the server, writes the device information into the Ispci-$count file, and compares whether the Ispci-tmp file is consistent with the Ispci-$count file, otherwise , generate an Ispci-tmp file for the device information in the server; when the Ispci-tmp file is consistent with the Ispci-$count file, send the startup completion information to the controller; create a gpu.txt file and a server.txt file ;Receive the shutdown request sent by the controller, and perform the shutdown operation. Through this method, the server can judge whether the file exists, and compare the consistency between the files to determine whether the server startup is normal. In addition, the server can be started or shut down. Under the control of the controller, it is carried out automatically without manual participation, and the automation of the server stability test is realized.

Description of drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are For some embodiments of the present invention, those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is the flow chart of the server reboot test method provided by one embodiment of the present invention;

Fig. 2 is a flowchart of a server reboot testing method provided by another embodiment of the present invention;

Fig. 3 is a flowchart of a server reboot testing method provided by another embodiment of the present invention;

Fig. 4 is a schematic diagram of the startup/shutdown sequence of the Pcie-Switch server provided by the embodiment of the present invention;

Fig. 5 is a schematic structural diagram of a server provided by an embodiment of the present invention;

Fig. 6 is a schematic structural diagram of a controller provided by an embodiment of the present invention;

FIG. 7 is a schematic structural diagram of a server reboot test system provided by an embodiment of the present invention.

detailed description

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work belong to the protection of the present invention. scope.

As shown in Figure 1, the embodiment of the present invention provides a server reboot testing method, which may include the following steps:

Step 101: establish mutual communication between the server and the controller through the switch;

Step 102: When the server receives the boot request sent by the controller, start the server;

Step 103: the server judges whether the Ispci-tmp file exists in itself, if yes, then execute step 104; otherwise, execute step 105;

Step 104: read the device information in the server, write the device information into the Ispci-$count file, and compare whether the Ispci-tmp file is consistent with the Ispci-$count file, if yes, then perform step 106, otherwise, perform step 107;

Step 105: generate an Ispci-tmp file for the device information in the server;

Step 106: Send start-up completion information to the controller, and create a gpu.txt file and a server.txt file, and execute step 108;

Step 107: Prompt an error message, and end the current process;

Step 108: Receive a shutdown request sent by the controller, and perform a shutdown operation.

Establish mutual communication between the server and the controller through the switch; when the server receives the boot request sent by the controller, it starts up; the server judges whether there is an Ispci-tmp file, and if so, reads the device information in the server , write the device information into the Ispci-$count file, and compare whether the Ispci-tmp file is consistent with the Ispci-$count file, otherwise, generate the Ispci-tmp file for the device information in the server; when the Ispci-tmp file When it is consistent with the Ispci-$count file, send the startup completion information to the controller; create a gpu.txt file and a server.txt file; receive the shutdown request sent by the controller, and perform a shutdown operation. By this method, the server judges the file Whether it exists, and compare the consistency between files, you can judge whether the server startup is normal. In addition, the startup and shutdown of the server can be carried out automatically under the control of the controller without manual participation, achieving server stability Automation of tests.

In an embodiment of the present invention, in order to ensure communication between the server and the controller, the specific implementation manner of step 101: the server is connected to the switch through the OS network and the BMC network; the controller is connected to the switch through the OS network.

In one embodiment of the present invention, in order to realize counting server startup times, further include: setting the first counter count in the server; The counter count+1 is stored in the count file; otherwise, start the first counter count, add the first counter count+1 to generate a count file, and write the first counter count into the startup item of the server, by writing the counter into the server startup The startup item ensures the accuracy of the statistics of startup times.

In one embodiment of the present invention, the server is a Pcie-Switch server, and the Pcie-Switch server includes: a resource server and a server end, wherein a retimer card is inserted in the server end, through which the retimer card and the MiniSASHD cable Connecting with the resource server; further comprising: setting the startup sequence; the booting includes: starting the resource server and the server in sequence according to the set startup sequence, ensuring the automatic normal startup of the Pcie-Switch server.

As shown in Figure 2, the embodiment of the present invention provides a server reboot test method, which is applied to a controller and may include the following steps:

Step 201: setting the second counter count in the controller, and setting the detection threshold;

Step 202: the controller initializes the second counter count;

Step 203: receiving the startup completion information sent by the server, and judging whether the count of the second counter count is less than the detection threshold, if yes, go to step 204; otherwise, go to step 205;

Step 204: Detect whether there are gpu.txt files and server.txt files in the server, if yes, then execute step 206; otherwise, execute step 207;

Step 205: close the server, exit the control server, and end the current process;

Step 206: call the shutdown function of the server, control the shutdown of the server, and execute step 208;

Step 207: delay for a certain time, and return to step 203;

Step 208: Send a boot request to the server, call a boot function of the server, control the server to boot, set the second counter count+1, and execute step 203.

In an embodiment of the present invention, in order to prevent the operations already executed in the server from affecting the startup of the server, the method further includes: clearing operating system logs in the server.

In one embodiment of the present invention, the server is a Pcie-Switch server, and the Pcie-Switch server includes: a resource server and a server end, wherein a retimer card is inserted in the server end, through which the retimer card and the MiniSASHD cable connected with the resource server; the control server shutdown includes: sequentially controlling the shutdown of the server end and the shutdown of the resource server; the startup of the control server includes: sequentially controlling the startup of the resource server and the server The terminal is booted and started, which ensures the normal startup of the Pcie-Switch server and realizes the automation of the stability test of the Pcie-Switch server.

In order to make the purpose, technical solution and advantages of the present invention clearer, a further detailed description will be made in conjunction with the interaction process between the server and the controller.

As shown in Figure 3, another embodiment of the present invention provides a server reboot testing method, which may include the following steps:

Step 301: establish mutual communication between the server and the controller through the switch;

In this step, the server is connected to the switch through the OS network and the BMC network; the controller is connected to the switch through the OS network; when the server is a Pcie-Switch server, the Pcie-Switch server includes: a resource server and a server end, wherein the resource The server can contain multiple GPUs, and a retimer card is inserted in the server end, which is connected to the resource server through the retimer card and the MiniSASHD cable; the resource server and the server end are connected to the switch through the OS network and the BMC network.

Step 302: setting the first counter count in the server, setting the second counter count in the controller, and setting the detection threshold;

In this step, when the server is a Pcie-Switch server, the first counter count can be set on the resource server.

Step 303: the controller initializes the second counter count, and clears the operating system log in the server;

In this step, the operating system logs in the server are cleared to avoid the impact of previous operations on the server on the stability test.

Step 304: the controller sends a boot request to the server, calls the server boot function, and starts the server;

In this step, when the server is a non-hot-swappable server such as a Pcie-Switch server, etc., the server can be started according to the startup sequence by setting the startup sequence, as shown in Figure 4, the embodiment of the present invention is set for the Pcie-Switch server Startup/shutdown sequence, since the Pcie-Switch server is not hot-swappable, you need to start the resource server containing the GPU first, and then start the server after the resource server is started; at the same time, in the shutdown process, first shut down the server end, and then shut down the resource server, so as to avoid server downtime caused by startup timing problems.

Step 305: the server judges whether the count file exists, if yes, execute step 306; otherwise execute step 307;

Step 306: store the first counter count+1 in the count file, and execute step 308;

Step 307: start the first counter count, add 1 to the first counter count, generate a count file, and write the first counter count into the server startup item;

The process from step 305 to step 307 is mainly about counting the number of startup times of the server itself, and the counting is done automatically through the counter without human participation.

Step 308: the server judges whether the Ispci-tmp file exists in itself, if yes, then executes step 309; otherwise executes step 310;

Step 309: Read the device information in the server, write the device information into the Ispci-$count file, and compare whether the Ispci-tmp file is consistent with the Ispci-$count file, if yes, then perform step 311; otherwise, perform step 312 ;

Step 310: generate an Ispci-tmp file for the device information in the server;

Steps 308 to 310 are to collect and compare the information of each device in the server. By comparing the device information, it is judged whether the server is fully started. For example, if there are GPU1 and GPU2 in a server, GPU1 is included in the Ispci-tmp file and GPU2 information, but only GPU1 information in the Ispci-$count file, the two files are inconsistent, indicating that the server startup has not been completed.

Step 311: Send the startup completion information to the controller, create a gpu.txt file and a server.txt file, and execute step 313;

Step 312: Prompt an error message, and end the current process;

Step 313: The controller receives the startup completion information sent by the server, and judges whether the count of the second counter count is less than the detection threshold, and if so, executes step 314; otherwise, executes step 315;

For example, if the detection threshold is set to 1000, and the count of the second counter count is the number of server startups, then when the number of server startups is less than 1000, the count of the second counter count is less than 1000.

Step 314: Detect whether there are gpu.txt files and server.txt files in the server, if yes, then execute step 316; otherwise, execute step 317;

In this step, it is first necessary to detect whether the server is connected to the switch, that is, the controller can connect to the server through the switch. As mentioned above, when the server is started, the gpu.txt file and the server.txt file will be created. Detect whether the gpu.txt file and server.txt file exist in the server to further confirm that the server has been started.

Step 315: close the server, exit the control server, and end the current process;

Step 316: call the shutdown function of the server to control the shutdown of the server, and execute step 304;

In this step, the second counter count+1, for non-hot-swappable servers such as Pcie-Switch servers, etc., can sequentially control the shutdown of the server end and the shutdown of the resource server according to the sequence set above.

Step 317: Delay for a certain time, and return to step 313.

If the gpu.txt and server.txt files are not detected, it may be because the server has not been started yet, so you can re-receive the start-up completion message sent by the server by extending a certain period of time, such as 5s.

As shown in FIG. 5 , the embodiment of the present invention provides a server applied to any of the server reboot testing methods described above. The server communicates with the controller of the peripheral device through the switch of the peripheral device, including: a switch unit 501 , a first judging unit 502, a reading and writing unit 503 and a generating unit 504, wherein,

The switch unit 501 is configured to perform a power-on start when receiving a power-on request sent by the controller of the peripheral device, and trigger the first judging unit 502 to perform a power-off operation when receiving a power-off request sent by the controller of the peripheral device;

The first judging unit 502 is used to judge whether there is an Ispci-tmp file when receiving the trigger of the boot unit 501, and if so, trigger the reading and writing unit 503; and compare whether the Ispci-tmp file and the Ispci-$count file consistent, otherwise, trigger generating unit 504;

Read and write unit 503, used to read the device information in the server, write the device information into the Ispci-$count file, and when the Ispci-tmp file is consistent with the Ispci-$count file, send the startup completion information to the outside Set the controller, and create gpu.txt file and server.txt file;

The generating unit 504 is configured to generate an Ispci-tmp file for each device information.

In another embodiment of the present invention, the switches connected to the peripheral devices are connected through the OS network and the BMC network.

In yet another embodiment of the present invention, the above server further includes: a second judging unit and a first counter (not shown in the figure), wherein,

The second judging unit is used to judge whether there is a count file, and if so, trigger the first counter; otherwise, start the first counter, generate the count file, and write the first counter into the startup item of the server;

The first counter is used for counting the times of power-on of the switch unit 501. When the switch unit 501 is powered on, count+1 is performed, and the times of power-on and start are stored in the count file.

In another embodiment of the present invention, the above-mentioned server is a Pcie-Switch server, and the Pcie-Switch server includes: a resource server and a server end, wherein a retimer card is inserted in the server end, through which the retimer card and the MiniSASHD cable are connected to the resource server connection.

As shown in FIG. 6 , the embodiment of the present invention provides a controller applied to any of the server reboot testing methods described above, the controller includes: a setting unit 601, a second counter 602, a detection unit 603 and a call control unit 604 ,in,

A setting unit 601, configured to set a detection threshold;

The detection unit 603 is used to judge whether the count of the second counter 602 is less than the detection threshold set by the setting unit 601, if yes, then detect whether there are gpu.txt files and server.txt files in the server of the peripheral device, and if so, trigger call the control unit 604;

Calling the control unit 604, used to call the shutdown function of the server of the peripheral device when receiving the trigger of the detection unit 603, control the shutdown of the server of the peripheral device, send a startup request to the server of the peripheral device, and call the startup function of the server of the peripheral device , control the server of the peripheral to start up, and add 1 to the count of the second counter 602 .

The information exchange and execution process among the units in the above-mentioned device are based on the same concept as the method embodiment of the present invention, and the specific content can refer to the description in the method embodiment of the present invention, and will not be repeated here.

As shown in FIG. 7 , an embodiment of the present invention provides a server reboot test system, including: at least one server 701 of any one of the above, a switch 702 and a controller 703, wherein,

At least one server 701 and a controller 703 are respectively connected to the switch 702 .

According to the above solution, the server reboot testing method, server, controller and system provided by each embodiment of the present invention have at least the following beneficial effects:

1. Establish mutual communication between the server and the controller through the switch; when the server receives the boot request sent by the controller, it starts the boot; the server judges whether there is an Ispci-tmp file, and if so, reads the Ispci-tmp file in the server. Device information, write the device information into the Ispci-$count file, and compare whether the Ispci-tmp file is consistent with the Ispci-$count file, otherwise, generate the Ispci-tmp file for the device information in the server; when the Ispci- When the tmp file is consistent with the Ispci-$count file, send the start-up completion information to the controller; create a gpu.txt file and a server.txt file; receive the shutdown request sent by the controller, and perform a shutdown operation. Through this method, through the server Judging whether the file exists, and comparing the consistency between the files, you can judge whether the server startup is normal. In addition, the startup and shutdown of the server can be carried out automatically under the control of the controller without manual participation, realizing the server Automation of Stability Testing.

2. The server is connected to the switch through the OS network and BMC network; the controller is connected to the switch through the OS network; the controller can control the automatic power on and off of the server, ensuring the automation of the server stability test. In addition, by setting the startup sequence; according to The set startup sequence starts the resource server and the server side of the Pcie-Switch server sequentially, so that the non-hot-swappable Pcie-Switch server can also realize the automation of the stability test.

3. After the server startup is completed, the startup completion information is sent to the controller, and the gpu.txt file and the server.txt file are created. After the controller receives the startup completion information, it will still detect whether the gpu.txt file and server exist in the server. txt file to ensure the accuracy of server startup.

4. By setting the detection threshold, only when the count of the second counter count in the controller is less than the detection threshold, the controller will call the startup/shutdown function to control the startup/shutdown of the server and avoid the stability test from entering an endless loop. In addition The controller, by clearing the operating system logs in the server, avoids operations in the server other than switching on and off from affecting the stability of the server, and further improves the accuracy of the server stability test.

It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is a relationship between these entities or operations. There is no such actual relationship or sequence. Furthermore, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus comprising a set of elements includes not only those elements, but also includes elements not expressly listed. other elements of or also include elements inherent in such a process, method, article, or device. Without more limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional same elements in the process, method, article or apparatus comprising said element.

Those of ordinary skill in the art can understand that all or part of the steps for realizing the above-mentioned method embodiments can be completed by hardware related to program instructions, and the aforementioned programs can be stored in a computer-readable storage medium. When the program is executed, the It includes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other various media that can store program codes.

Finally, it should be noted that: the above descriptions are only preferred embodiments of the present invention, and are only used to illustrate the technical solutions of the present invention, and are not used to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims (10)

1. The server reboot test method is characterized in that, through a switch, the mutual communication between the server and the controller is established; it also includes: When the server receives the boot request sent by the controller, it starts the boot; The server judges whether the Ispci-tmp file exists, if yes, reads the device information in the server, writes the device information into the Ispci-$count file, and compares whether the Ispci-tmp file is consistent with the Ispci-$count file, otherwise , generate the Ispci-tmp file for the device information in the server; When the Ispci-tmp file is consistent with the Ispci-$count file, send startup completion information to the controller; Create gpu.txt file and server.txt file; Receive the shutdown request sent by the controller, and perform the shutdown operation. 2. The method of claim 1, wherein, The mutual communication between the server and the controller is established through the switch, including: The server is connected to the switch through the OS network and the BMC network; The controller is connected to the switch through the OS network; and / or, It further includes: setting a first counter count in the server; After the startup, it further includes: the server judges whether there is a count file, if yes, then stores the first counter count+1 in the count file; otherwise, starts the first counter count, and The first counter count+1, generate a count file, and write the first counter count into the startup item of the server. 3. The method according to claim 1 or 2, characterized in that, Described server is Pcie-Switch server, and this Pcie-Switch server comprises: resource server and server end, and wherein, described server end is inserted with a retimer card, is connected with described resource server by this retimer card and MiniSASHD cable; It further includes: setting the start sequence; The booting includes: sequentially starting the resource server and the server end according to the set startup sequence. 4. The server reboot test method is characterized in that it is applied to the controller, and the second counter count is set in the controller, and the detection threshold is set; it also includes: M1. The controller initializes the second counter count; M2, receiving the startup completion information sent by the server, judging whether the count of the second counter count is less than the detection threshold, if yes, then detecting whether there are gpu.txt files and server.txt files in the server, and if so, calling the server's Shutdown function, control server shutdown; M3. Send a start-up request to the server, call the start-up function of the server, control the start-up of the server, and set the second counter count+1, and execute M2. 5. The method according to claim 4, further comprising: clearing operating system logs in the server. 6. according to the described method of claim 4 or 5, it is characterized in that, described server is Pcie-Switch server, and this Pcie-Switch server comprises: resource server and server end, wherein, described server end is inserted with a retimer card , connected to the resource server through the retimer card and the MiniSASHD cable; The controlling the shutdown of the server includes: sequentially controlling the shutdown of the server end and the shutdown of the resource server; The control server startup includes: sequentially controlling the startup of the resource server and the startup of the server. 7. The server that is applied to the server reboot test method described in any one of claims 1 to 3 communicates with the controller of the peripheral through the switch of the peripheral, and is characterized in that it includes: a switch unit, a first judgment unit, read-write unit, and generation unit, where, The switch unit is configured to perform a power-on start when receiving a power-on request sent by the controller of the peripheral device, and trigger the first judging unit to perform a power-off operation when receiving a power-off request sent by the controller of the peripheral device ; The first judging unit is used to judge whether there is an Ispci-tmp file when receiving the trigger of the boot unit, and if so, trigger the read and write unit; and compare the Ispci-tmp file and Ispci-$ Whether the count file is consistent, otherwise, trigger the generation unit; The read and write unit is used to read the device information in the server, write the device information into the Ispci-$count file, and when the Ispci-tmp file is consistent with the Ispci-$count file, send Start the completion information to the controller of the peripheral, and create the gpu.txt file and the server.txt file; The generating unit is configured to generate an Ispci-tmp file for each device information. 8. The server according to claim 7, wherein: Switches connected to peripherals through OS network and BMC network; and / or, Further comprising: a second judging unit and a first counter, wherein, The second judging unit is used to judge whether there is a count file, and if so, trigger the first counter; otherwise, start the first counter, generate a count file, and write the first counter into the server to boot startup item; The first counter is used to count the number of startup times of the switch unit. When the switch unit is started up, count+1 is performed, and the number of startup times is stored in the count file; and / or, The server is a Pcie-Switch server, and the Pcie-Switch server includes: a resource server and a server end, wherein a retimer card is inserted in the server end, and is connected to the resource server through the retimer card and the MiniSASHD cable. 9. The controller applied to the server reboot test method according to any one of claims 4 to 6, characterized in that, comprising: a setting unit, a second counter, a detection unit and a call control unit, wherein, The setting unit is used to set the detection threshold; The detection unit is used to judge whether the count of the second counter is less than the detection threshold set by the setting unit, if yes, then detect whether there are gpu.txt files and server.txt files in the server of the peripheral device, if yes , then trigger the calling control unit; The calling control unit is configured to call the shutdown function of the server of the peripheral device when receiving the trigger of the detection unit, control the server of the peripheral device to shut down, send a startup request to the server of the peripheral device, and call the server of the peripheral device The booting function controls the booting of the server of the peripheral device, and adds 1 to the count of the second counter. 10. server reboot test system, is characterized in that, comprises: at least one server described in claim 7 or 8, switch and the controller described in claim 9, wherein, The at least one server and the controller are respectively connected to the switch.