JP5439736B2 - Computer management system, computer system management method, and computer system management program - Google Patents

Description

  The present invention relates to a computer management system, and more particularly to a computer management system for managing a server system, a computer system management method, and a computer system management program.

  A server system using a plurality of servers is known. In such a server system, a management processor is provided in addition to a processor for actually executing processing required for the server (see, for example, Patent Document 1 (Japanese Patent Laid-Open No. 10-301799)). The management processor is a processor for performing management processing of devices provided in each server. Specifically, the management processor performs monitoring of the state (temperature, voltage, etc.) in each server, control of power ON / OFF, control of sequence, and failure of internal devices. If an abnormality is recognized in the server, it is notified outside the server or the operation of the server is stopped.

  However, this management processor itself may also fail. Also, when updating the management program executed by the management processor, it is necessary to stop the management by the management processor. If the management processor cannot be moved, power control and failure processing cannot be performed even if there is no abnormality in the operation of the internal device. Therefore, the operation is stopped regardless of the actual operation system.

  Therefore, even if the management processor breaks down or the management program is being updated, it is desirable that the operation of the server equipped with the management processor can be continued.

  In addition, as a related technique, the computer described in patent document 2 (Unexamined-Japanese-Patent No. 2006-268742) is mentioned. The computer is connected to a computer device and performs at least one of monitoring and power control of the device, and includes a controller connected to a device external to the computer and a failure detection unit that detects a failure of the controller. In this computer, when a fault detection notification is received from the fault detection unit, the path between the controller and the device is disconnected and the switching control is performed so that the path between the device outside the computer and the device is connected. It is possible to perform at least one of device monitoring and power supply control from this apparatus.

  As another related technique, Patent Document 3 (Japanese Patent Laid-Open No. 2005-99918) describes that two management module devices are mounted and the management module device is duplicated.

  As another related technique, Patent Document 4 (Japanese Patent Laid-Open No. 2006-285384) discloses a processor processing method for processing a failure of a management processor in a system including a plurality of processing nodes. This processor fault processing method includes a fault detection unit that detects the occurrence of a fault in an external management processor that is a management processor installed in a processing node different from the processing node in which the management processor is installed, and the fault detection unit. When the occurrence of a failure in the external management processor is detected, an interrupt means for transmitting a predetermined interrupt signal to the external management processor in which the failure has occurred, and when an interrupt signal is received from any of the external management processors, And a failure recovery means for executing a failure recovery process for recovering a failure.

  As another related technique, Patent Document 5 (Japanese Patent Application Laid-Open No. 2004-318498) describes a fail-safe device of a plurality of arithmetic processing devices connected to each other via a network. This fail-safe device includes an alternative control program storage unit that stores an alternative control program for performing alternative control of a plurality of arithmetic processing units. When the arithmetic processing unit detects a faulty arithmetic processing unit, An alternative control program for substituting the failed arithmetic processing unit is read, and alternative control of the faulty arithmetic processing unit is performed via the network.

JP-A-10-301799 JP 2006-268742 A JP-A-2005-99918 JP 2006-285384 A JP 2004-318498 A

  An object of the present invention is to provide a computer management system, a computer system management method, and a computer system management program capable of continuing operation of a server equipped with a management processor even when the management processor cannot be used. There is to do.

  A computer management system according to the present invention includes a plurality of servers and a management module device connected to the plurality of servers. Each of the plurality of servers includes a server management processor that manages each of the servers by executing a management program. The management module device includes a management module processor that manages the plurality of servers as a whole, a replacement processor that replaces and manages a selected server selected from the plurality of servers, and a replacement program that is the same program as the management program As a program storage unit. The management module processor causes the replacement processor to read the replacement program from the program storage unit when the selected server is selected from the plurality of servers. The replacement processor manages the selection server by executing the read replacement program.

  The computer system management method of the present invention is a computer system management method comprising a plurality of servers and a management module device connected to the plurality of servers and managing the plurality of servers as a whole. In this computer system management method, the server management processor provided in each of the plurality of servers normally executes a management program so that each server is normally managed, and in advance in the management module device, Corresponding to the server management processor of the selected server, the step of storing the same program as the management program as a substitute program, the step of selecting a selected server from a plurality of servers, and the replacement processor in the management module device And a step of replacing the selected server by executing the replacement program read by the replacement processor.

  The computer system management program of the present invention is a computer system management program that causes a computer to execute the above-described computer system management method.

  According to the present invention, there are provided a computer management system, a computer system management method, and a computer system management program capable of continuing the operation of a server managed by the management processor even when the management processor cannot be used. .

(Embodiment)
Embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a schematic configuration diagram of a computer system 1 according to the present embodiment. The computer system 1 includes a plurality of servers 2 (2-1, 2-2,..., 2-n) and a management module device 3 that manages the plurality of servers 2 as a whole. This computer system 1 is assumed to be a blade server in which a plurality of servers 2 and a management module device 3 are housed in a single housing (not shown). The management module device 3 is connected to each of the plurality of servers 2 via a system management path 42 and a management processor replacement path 41 so as to be accessible. A backplane 4 is provided between the management module device 3 and the plurality of servers 2. The system management path 42 and the management processor replacement path 41 pass through the backplane 4 and connect the management module device 3 and each server 2 in an accessible manner.

  First, the configuration of each server 2 will be described.

  Each server 2 includes an operational device 25, a server management processor 22, and a first switching device 21. The active device 25 and the first switching device 21 are connected via a server management path 23. The server management processor 22 and the first switching device 21 are connected via a normal management path 24. The server management processor 22 is connected to the management module device 3 via a system management path 42. The first switching device 21 is connected to the management module device 3 via a management processor replacement path 41.

  The operational device 25 is a device for actually executing processing requested by each server, and includes a processor such as a CPU, for example.

  The server management processor 22 is a processor provided separately from the operational device 25 and manages the operational device 25. A management program 22-1 is installed in the server management processor 22. The server management processor 22 realizes its function by executing the management program 22-1. Specifically, the server management processor 22 monitors the failure of the active device 25, power ON / OFF control when a failure occurs, chip set register access, host internal temperature monitoring, voltage monitoring, and hardware (HW) Ware) Interrupt processing, etc.

  The first switching device 21 is a device for switching between management of the active device 25 by the server management processor 22 and management of the active device 25 by the management module device 3. For example, the first switching device 21 realizes its function by executing a program in which a CPU is installed.

  Next, the configuration of the management module device 3 will be described. The management module device 3 includes a management module processor 31, a replacement management processor 32, a program storage unit 33, and a second switching device 50.

  The management module processor 31 is a processor for managing the entire plurality of servers 2. The management module processor 31 is connected to the server management processor 22 of each server 2 via a system management path 42. Further, the management module processor 31 is connected to the second switching device 50 via the path 38, the replacement management processor 32 via the path 35, and the program storage unit 33 via the path 39. The management module processor 31 monitors the state of each server 2 and performs power supply control via the system management path 42. The management module processor 31 has a CPU, and realizes its function by a computer system management program stored in advance in a ROM or the like being executed by the CPU.

  The replacement management processor 32 is provided to replace the server management processor 22. The replacement management processor 32 is connected to the second switching device 50 via the path 36 and the program storage unit 33 via the path 37.

  The program storage unit 33 stores alternative programs. The alternative program is the same program as the management program installed in the server management processor 22 of each server 2. The program storage unit 33 stores alternative programs for all management programs that may be installed in the plurality of servers 2. The program storage unit 33 is realized by, for example, a hard disk or a ROM.

  The second switching device 50 is a device for switching the connection between the replacement management processor 32 and each server 2. The second switching device 21 is connected to the first switching device 50 of each server 2 via the management processor replacement path 41. For example, the second switching device 50 realizes its function by executing a program in which a CPU is installed.

  Subsequently, the operation of the computer system 1 according to the present embodiment will be described.

  Normally, in each server 2, the server management processor 22 manages the active device 25. That is, the first switching device 21 connects the normal management path 24 to the server management path 23, and the management processor replacement path 41 is in an invalid state.

  Here, it is assumed that a failure has occurred in the server management processor 22 of a certain server 2. FIG. 2 is a flowchart showing an operation method when a failure occurs in the server management processor 22.

Step S1; Failure Detection When a failure occurs in the server management processor 22, the management module processor 31 detects a failure in the server management processor 22 via the system management path 42 on the management module device 3 side. The management module processor 31 selects the server 2 equipped with the server management processor 22 in which the failure is detected as the failure server 2 (selected server). At the same time, the management module processor 31 identifies the management program that has been executed by the server management processor 22 of the failed server 2. In order to specify the management program, for example, information for specifying the management program may be stored in advance in association with information for specifying the server 2 or the server management processor 22.

Step S2: Designation of Alternative Program Subsequently, the management module processor 31 accesses the program storage unit 33 via the path 39. Then, it searches for an alternative program corresponding to the management program specified in S1. After the search, the management module processor 31 accesses the alternative management processor 32 via the path 35 and instructs to read the corresponding alternative program by the search.

Step S3: Path Switching The management module processor 31 accesses the second switching device 50 via the path 38 and validates the management processor replacement path 41 connected to the failed server 2.

Step S4: Execution of Alternative Program The alternative management processor 32 that has received an instruction to read an alternative program reads the designated alternative program from the program storage unit 33 and executes it. Further, the replacement management processor 32 accesses the first switching device 21 of the failed server 2 through the management processor replacement path 41 and connects the management processor replacement path 41 to the server management path 23. Thereafter, the active device 25 of the failed server 2 is managed not by the server management processor 22 but by the alternative management processor 32.

Step S5: Recovery of Server Management Processor The management module processor 31 overwrites the management program of the server management processor 22 of the failed server 2 while the failure server 2 is being managed by the replacement management processor 32. Then, the server management processor 22 is restarted to attempt recovery. When recovered, the first switching device 21 and the second switching device 50 are switched to return to normal management. In other words, the operational device 25 is managed by the server management processor 22.

  By the operations in steps S1 to S5 described above, even if the server management processor 22 fails, the replacement management processor 32 performs the replacement management, and the operation of the failed server 2 can be continued.

  Next, an operation when the management program 22-1 of a certain server is updated will be described. FIG. 3 is a flowchart showing an operation when the management program 22-1 is updated.

Step S11: Acceptance of update instruction An instruction to update the management program 22-1 of a certain server 2 to a program stored in a designated location in the management module apparatus 3 from a user via an input device (not shown). Is input. In the management module device 3, the management module processor 31 receives an update instruction. The management module processor 31 selects the update target server 2 as the update server 2 (selected server).

Step S12: Update of Alternative Program The management module processor 31 that has received the update instruction first accesses the program storage unit 39 to update the alternative program corresponding to the management program of the update server.

Step S13: Alternative Program Reading Instruction After the alternative program is updated, the management module processor 31 accesses the alternative management processor 32 and instructs to read the updated alternative program.

Step S14: Path Switching The management module processor 31 accesses the second switching device 50 so that the replacement management processor 32 and the update server 2 can access the path 36 and the management processor replacement path. 41 is connected.

Step S15: Execution of Substitute Program Upon receiving the read instruction, the substitution management processor 32 reads the updated substitution program from the program storage unit 33 and executes it. Further, the replacement management processor 32 accesses the first switching device 21 to connect the server management path 23 and the management processor replacement path 41. As a result, the active device 25 of the update server 2 is managed by the replacement management processor 32.

Step S16: Update Management Program The management module processor 31 is designated to the server management processor 22 of the update server 2 while the active device 25 of the update server 2 is managed by the replacement management processor 32. To update the management program from the new location. The server management processor 22 receives the instruction and starts updating.

  When the update of the management program 22-1 is completed, the management module processor 31 notifies the replacement management processor 32 of the update completion. The replacement management processor 32 that has received the update completion notification ends the replacement management.

Step S17; Switching of Paths In the update server 2, the server management processor 22 accesses the first switching device 21 to perform normal management so that the active device 25 is managed by the server management processor 22. The server path 24 and the server management path 23 are connected. Further, the management module processor 31 accesses the second switching device 50 and disconnects the connection between the path 36 and the management processor replacement path 41.

  Through the operations in steps S11 to S17 described above, both the alternative program in the program storage unit 33 and the management program of the server management processor 22 can be updated without stopping the operation of the update server 2.

  According to the present embodiment, even when the server management processor 22 becomes unusable, such as when a failure occurs or when an update is performed, the target server can be replaced and managed by the replacement management processor 32. Server operation can be continued.

Usually, in a computer system having a plurality of servers such as a blade server, a management module device for managing the entire system is provided. Here, when the management processor on the server side fails, it is conceivable to replace the function from the management module device side. However, not all of the installed management programs are the same in each of the plurality of servers. Therefore, simply performing substitution management from the management module device side, the functions that can be substituted are limited to basic processing to the extent that the power is safely turned off, and the operation of the server cannot be continued. It is also conceivable that management of one of a plurality of servers is replaced by another server. However, in this case as well, the management programs are not necessarily the same for a plurality of servers, so the functions that can be replaced are limited to basic processing.
On the other hand, according to the present embodiment, the replacement management processor 32 reads and executes the same replacement program as that of the management processor, so that the same function as that of the server management processor 22 can be replaced. Therefore, the operation of the server can be continued as it is.

In addition, in order to be able to continue server management even in the event of a failure or update, it is conceivable to duplicate the management processor in each server. However, in order to duplicate the management processor in each server, two management processors are required for one server, which increases the cost.
On the other hand, according to the present embodiment, it is only necessary to prepare one replacement management processor for the plurality of servers 2. Therefore, the cost can be greatly reduced as compared with the case where the management processor is duplicated in each server.

  In the present embodiment, an alternative management processor 32 is provided separately from the management module processor 31. Therefore, it is possible to continue the management of the plurality of servers 2 by the management module processor 31 even while the replacement management processor 32 performs the replacement management of the server. That is, it is possible to perform alternative management of a server without stopping the function as the management module device 3.

  Further, the replacement management processor 32 may be configured to play an auxiliary role of the management module processor 31 during normal operation. In this way, resources can be used more effectively than when the function of the substitute processor 32 is limited to substitution management.

  In the present embodiment, the case where the second switching device 50 is provided in the management module device 3 has been described. However, the second switching device 50 is not necessarily provided in the management module device 3. For example, as shown in the modification of FIG. 4, the second switching device 50 may be provided on the backplane 4. Even when configured as in the modified example of FIG. 4, the same operation as in the present embodiment can be achieved. In addition, the amount of wiring between the backplane 4 and the management module device 3 can be reduced.

  Further, in the present embodiment, a case has been described in which the second switching device 50 selects which server among the plurality of servers 2 to validate the management processor replacement path 41. However, the management processor replacement path 41 to be valid may be selected from the plurality of servers 2 by another method. For example, in the modification shown in FIG. 5, as the second switching device 50, a management processor replacement path 41 corresponding to all servers is connected to the bus. A unique address is set for this bus. The replacement management processor 32 can access the server 2 to be replaced and managed by designating the bus address. Even with a configuration such as that of the modification shown in FIG. 5, the same operation as in the present embodiment can be achieved.

FIG. 6A shows a modification of the present embodiment. In this modification, in the first switching device 21 portion, the normal management path 24, the management processor replacement path 41, and the server management path 23 are connected via a bus. A unique address is assigned to this bus. The replacement management processor 32 controls the validity / invalidity of the management processor replacement path 41 by switching the bus via the second switching device 50. Even in the configuration as in the modified example of FIG. 6A, the same operation as in the present embodiment can be achieved. Further, in the modified example of FIG. 6A, the second switching device 50 may be arranged on the backplane 4 as in the modified example shown in FIG.
As still another modification of the first switching device 21 portion, the first switching device 21 itself detects a failure of the server management processor 22, and the first switching device 21 automatically switches the path. You can also In this case, the server management processor 22 and the first switching device 21 may be connected by a signal line whose level changes when the server management processor fails.
FIG. 6B shows still another modification regarding the first switching device 21 portion. In this modification, the first switching device 21 and the management module processor 31 are connected by a path 40 that does not go through the server management processor 22. The first switching device 21 is switched not by the replacement management processor 32 but by the management module processor 31 via the path 40. Even with the configuration as shown in FIG. 6B, the same operation as in the present embodiment can be achieved.

  FIG. 7 shows still another modification of the present embodiment. In the modification shown in FIG. 7, the server management processor 22 is attached to the server 2 in a hot-swappable manner. Similarly, the replacement management processor 32 is also attached to the management module device 3 in a hot-swappable manner. According to the configuration of the modified example of FIG. 7, when the server management processor 22 is out of order, replacement management is performed by the replacement management processor 32, and the failed server management processor 22 is replaced during that time. can do. In other words, the server management processor 22 can be replaced without stopping the computer system 1 and without stopping the management function of the server 2.

  Note that the modification examples shown in FIGS. 1 and 4 to 7 can be combined within a consistent range.

It is a schematic block diagram of embodiment of this invention. It is a flowchart at the time of failure occurrence. It is a flowchart at the time of an update. It is a schematic block diagram which shows the modification of embodiment of this invention. It is a schematic block diagram which shows the other modification of embodiment of this invention. It is a schematic block diagram which shows the other modification of embodiment of this invention. It is a schematic block diagram which shows the other modification of embodiment of this invention. It is a schematic block diagram which shows the other modification of embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Server system 2 Server 3 Management module apparatus 4 Backplane 21 1st switching part 22 Server management processor 23 Server management path 24 Normal management path 25 Operation system device 31 Management module processor 32 Replacement processor 33 Program storage part 34 Program Table 35 path 36 path 37 path 38 path 39 path 41 Management processor alternative path 42 System management path 50 Second switching section

Claims (20)

Multiple servers,
A management module device that collectively manages each of the plurality of servers,
Each of the plurality of servers is
A server management processor for managing each of the servers;
The management module device includes:
A program storage unit for storing a management program for the server management processor for each server;
When the server management processor fails or is updated, the management program corresponding to the server management processor is read from the program storage unit and the read management program is executed. A computer management system comprising: a replacement management processor that replaces the function of the server management processor until the update is completed and continues management of the servers. A computer management system according to claim 1,
Each of the servers
An operational device for actually executing the processing requested by each server;
Always connected to the active device via a server management path, normally connected to the server management processor via a normal management path, and a management processor replacement path when the server management processor fails or is updated And a first switching device connected to the replacement management processor via the computer management system. A computer management system according to claim 2 ,
The management module device includes:
A management module connected to the server management processor of each server via a system management path, connected to the program storage unit via a first path, and connected to the replacement management processor via a second path Further comprising a processor;
The management module processor selects a server equipped with the server management processor in which a failure is detected from the plurality of servers as a failure server, and the server management processor of the failed server is executing The management program is specified, the program storage unit is accessed through the first path, the alternative program corresponding to the specified management program is searched, and the replacement management processor is accessed through the second path. And a computer management system that instructs to read the corresponding alternative program by the search. A computer management system according to claim 3,
A third path is connected to the management module processor via a third path, is connected to the replacement management processor via a fourth path, and is connected to the server management processor of each server via a management processor replacement path. 2 further includes a switching device,
The management module processor accesses the second switching device via the third path, validates the management processor replacement path connected to the failed server, and is connected to the fourth path and the failed server. A computer management system for connecting the management processor replacement path to enable access to the replacement management processor and the server management processor of the failed server. A computer management system according to claim 4,
The management module processor overwrites the management program of the server management processor of the failed server while the management of the failed server is performed by the replacement management processor, and manages the server of the failed server When a recovery is attempted by restarting the processor, the first switching device and the second switching device are switched back to normal management and the failed server is managed by the server management processor. Computer management system. A computer management system according to claim 5, wherein
The management module processor selects a server to be updated from among the plurality of servers as an update server, accesses the program storage unit via the first path, and corresponds to the management program of the update server A computer management system that updates an alternative program and, after updating the alternative program, accesses the alternative management processor via a second path to instruct to read the updated alternative program. A computer management system according to claim 6, comprising:
The management module processor accesses the second switching device via the third path, validates the management processor replacement path connected to the update server, and is connected to the fourth path and the update server. A computer management system that connects the management processor replacement path to enable access to the replacement management processor and the server management processor of the update server. A computer management system according to claim 7,
The management module processor is configured to send the management program from a specified location to the server management processor of the update server while the operational device of the update server is being managed by the replacement management processor. of instructing the update, an update of the management program of the management processor for the server is complete, perform the update completion notification to the alternative for the management process Tsu support,
When the replacement management processor receives the update completion notification, the replacement management processor ends the replacement management of the active device of the update server. A computer management system according to any one of claims 4 to 8, comprising:
The second switching device is a computer management system provided on a backplane existing between each server and the management module device. A computer management system according to any one of claims 1 to 9,
A computer management system in which at least one of the server management processor and the replacement management processor is mounted to be hot-swappable.   The computer management system as described in any one of Claims 1 thru | or 10 WHEREIN: The computer used as each of several servers or a management module apparatus. A computer system management method including a plurality of servers and a management module device that collectively manages each of the plurality of servers,
Storing a management program for a server management processor for managing each server in each server in a program storage unit in the management module device;
The replacement management processor in the management module device reads the management program corresponding to the server management processor from the program storage unit when the server management processor fails or is updated, and executes the read management program Then, replacing the server management processor until the replacement or update of the server is completed, replacing the function of the server management processor and continuing the management of each server. A computer system management method according to claim 12, comprising:
The operational device in each server executes a process requested by each server;
A first switching device in each of the servers,
Constantly connecting to the active device via a server management path;
Normally connecting to the server management processor via a normal management path;
A management method for a computer system, further comprising: connecting to the replacement management processor via a management processor replacement path when the server management processor fails or is updated. A computer system management method according to claim 13 , comprising:
A management module processor in the management module device;
Connecting to the server management processor of each server via a system management path;
Connecting to the program storage unit via a first path;
Connecting to the alternative management processor via a second path;
Selecting a server equipped with the server management processor in which a failure is detected as the failed server among the plurality of servers;
Identifying the management program being executed by the server management processor of the failed server;
Accessing the program storage unit via the first path and searching for an alternative program corresponding to the identified management program;
A method of managing a computer system, further comprising the step of accessing the replacement management processor via a second path and instructing to read the corresponding replacement program by a search. A method for managing a computer system according to claim 14, comprising:
The second switching device
Connecting to the management module processor via a third path;
Connecting to the alternative management processor via a fourth path;
Connecting to the server management processor of each server via a management processor alternative path;
The management module processor is
Access the second switching device via the third path, validate the management processor replacement path connected to the failed server, and replace the management processor connected to the fourth path and the failed server A computer system management method, further comprising: connecting a path to make the replacement management processor accessible to the server management processor of the failed server. A computer system management method according to claim 15, comprising:
The management module processor is
Overwriting the management program of the server management processor of the failed server while the failure server is being managed by the replacement management processor;
Restarting the server management processor of the failed server and attempting recovery;
A step of switching the first switching device and the second switching device to return to normal management when recovery is performed, so that management of the failed server is performed by the server management processor; Computer system management method. A computer system management method according to claim 16, comprising:
The management module processor is
Selecting a server to be updated from among the plurality of servers as an update server;
Accessing the program storage unit via the first path to update an alternative program corresponding to the management program of the update server;
A method of managing a computer system, further comprising the step of accessing the replacement management processor via a second path and instructing to read the updated replacement program after updating the replacement program. A computer system management method according to claim 17, comprising:
The management module processor accesses the second switching device via the third path, validates the management processor replacement path connected to the update server, and is connected to the fourth path and the update server. The management system management method further comprising the step of: connecting the management processor replacement path to enable access to the replacement management processor and the server management processor of the update server. A computer system management method according to claim 18, comprising:
The management module processor is
Instructing the server management processor of the update server to update the management program from a specified location while the operational device of the update server is being managed by the replacement management processor When,
When the update of the management program of the server management processor is completed, an update completion notification is sent to the replacement management processor;
The replacement management processor comprises:
A management method for a computer system, further comprising: upon receiving the update completion notification, terminating substitution management of the active device of the update server.   A management program for causing a computer to execute the computer system management method according to any one of claims 12 to 19.

Images