JPH0756761A - Computer device - Google Patents

Abstract

PURPOSE:To minimize the influence exerted on a system by halting system management at the time of diagnosing the generation of a self-fault and outputting a command instructing a stand-by service processor(SVP) to start system management. CONSTITUTION:Each SVP previously selects a substitutive SVP to be a system SVP to be used when a fault occurs in the SVP itself from normal other SVPs. The SVP executes self-diagnosis by a self-diagnosis part 19, and when no fault occurs, continues a series of processing. If a fault occurs, a diagnostic interface 21 informs a fault processing executing part 15 of the occurance of the fault to halt system management and a substitutive SVP is recognized by referring to a substitutive system SVP management table and the substitutive system SVP is informed of a function substituting request through a transmission processing part 22.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer system capable of controlling system connection of a central processing unit having a self-diagnosis function.

[0002]

2. Description of the Related Art FIG. 13 shows, for example, JP-A-64-88677.
It is a block diagram which shows the conventional computer apparatus shown by Unexamined-Japanese-Patent No. 1-N in the figure is N CPU of multiprocessor structure via the system bus 4.
(Central processing unit) has a self-diagnosis function for diagnosing whether or not it is operating normally, and a function for transmitting a failure notification signal when a failure occurs as a result of self-diagnosis. Have Reference numeral 2 denotes a self-diagnosis unit that diagnoses whether or not it is operating normally, and 3 stops various processing of the CPU when the self-diagnosis unit 2 diagnoses that a failure has occurred, and a failure notification signal. Send
Further, it is a diagnostic interface for receiving various signals from the service processor 6 described later.

Reference numeral 4 denotes a computer system bus, and 5
Is a computer diagnostic bus, and 6 is CPUs 1-1 to 1-N
When a failure notification signal is received from the SVP (service processor), the CPU that sends the failure notification signal is disconnected from the system, and 7 is failure information of the CPUs 1-1 to 1-N managed by the SVP 6 and the current system. It is a memory that stores the configuration and the like.

Next, the operation will be described. CPU1-1
Each of 1 to 1-N performs various kinds of processing according to its own program, but when there are no waiting tasks to be executed during operation, or periodically, a macro instruction for performing self-diagnosis by the self-diagnosis unit 2. Is executed and self-diagnosis is performed. If, as a result of the self-diagnosis, it is diagnosed as normal, the various processes are continued as before, but if it is diagnosed that a failure has occurred, the various processes are stopped,
The diagnostic interface 3 transmits a fault notification signal as an interrupt signal to the SVP 6 via the diagnostic bus 5. For convenience of explanation, assume that a failure occurs in the CPU 1-1 and a failure notification signal is transmitted from the diagnosis interface 3 of the CPU 1-1.

Next, the operation of the SVP 6 will be described with reference to the flowchart of FIG. First, the SVP 6 manages the CPUs 1-1 to 1-N based on the stored contents of the memory 7, but as described above, when a failure occurs in the CPU 1-1, a failure notification signal is transmitted, so that the SVP 6 is notified to the SVP 6. Interrupt is generated (step ST1). As a result, the SVP 6 determines whether it is an interrupt from the CPU 1-1 to 1-N or an interrupt from another device (step ST2). If the interrupt is from another device, a predetermined process is executed (step ST3), and the interrupt is the CPU 1-1 to 1-N.
If it is, the contents of the interrupt notification (fault notification signal) are analyzed in detail and the process is branched (step ST
4).

[0006] For example, as the type of notification content, the above-mentioned C
In addition to the self-diagnosis error during execution of processing in the PU, there are self-diagnosis errors when the power is turned on, trap processing cannot continue, etc. In these cases, although not described in detail, each processing is performed depending on the cause of interrupt. (Steps ST5-7). On the other hand, as a result of detailed analysis of the interrupt notification content, when it is determined that the self-diagnosis error is occurring during execution of processing in the CPU 1-1, the SVP 6 performs processing for disconnecting the CPU 1-1 from the system.

That is, the CPU 1-1, as described above,
Since various processes have already been stopped, the CPU 1-1 is disconnected from the system configuration stored in the memory 7 (step ST8), and the other CPUs 1-2 to 1-N in which no failure has occurred On the other hand, the CPU 1-1 is disconnected from the system by notifying the CPU 1-1 that a failure has occurred through the diagnostic bus 5 (step ST9), and the CPU 1-1 is disconnected from the system. .

[0008]

Since the conventional computer system is configured as described above, when a CPU fails, the CPU with the failure can be disconnected from the system, and therefore, it is in operation. Although it is possible to minimize the adverse effect on the system, if a failure occurs in the SVP, no measures are taken and the SVP in which the failure occurs may have a great adverse effect on the system. There was a problem.

The invention of claim 1 has been made to solve the above-mentioned problems, and provides a computer device capable of minimizing the adverse effect on the system even when a failure occurs in the SVP. The purpose is to

The inventions of claims 2 and 3 are defined by claim 1.
Another object of the present invention is to provide a computer device that can more surely minimize the adverse effects on the system.

[0011]

According to a first aspect of the present invention, there is provided a computer system in which, when the self-diagnosis unit diagnoses that a failure has occurred in the self-system management, the system management is performed. Is canceled and a command to the effect that system management should be started is output to the standby service processor.

Further, the computer system according to the second aspect of the present invention outputs a command for self-diagnosis to each standby service processor when the system is managed by itself. , Check the diagnosis result of each standby service processor based on the presence / absence of a normal response signal from each standby service processor, and stop the system management when the self-diagnosis unit diagnoses that a self-failure has occurred. At the same time, based on the diagnosis result, a command to start system management is output to the normal service processor for standby.

Further, the computer apparatus according to the third aspect of the present invention is such that when the load condition measured by the load factor measuring unit exceeds a predetermined value, it is considered that a failure has occurred in itself. is there.

[0014]

In the computer system according to the present invention, when the self-diagnosis unit diagnoses that a failure has occurred in the self-system management, the computer system suspends the system management and waits. By providing a switching control unit that outputs a command to start system management to another service processor, the faulty service processor is disconnected from the system and the standby service processor performs system management. become.

Further, the computer device according to the second aspect of the present invention outputs a command for performing self-diagnosis to each standby service processor when the computer device itself manages the system. The diagnosis result of each standby service processor is known based on the presence or absence of a normal response signal from each standby service processor, and system management is stopped when the self-diagnosis unit diagnoses that a failure has occurred in itself. In addition, by providing a switching control unit that outputs a command to start system management to the normal standby service processor based on the diagnosis result, the faulty service processor is disconnected from the system. , The service processor for normal standby comes to perform system management.

Further, the computer system according to the third aspect of the present invention considers that a fault has occurred in itself when the load condition measured by the load factor measuring unit exceeds a predetermined value. Only service processors with low operating rates are selected as alternative service processors.

[0017]

【Example】

Example 1. An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a computer apparatus according to an embodiment of the present invention. In the figure, the same reference numerals as those of the conventional one indicate the same or corresponding portions, and the explanation thereof will be omitted. 8-1
8 to N have a self-diagnosis function for diagnosing whether or not the self is normally operating, and as a result of the self-diagnosis, a CPU which transmits a normal response signal when the normal operation is performed.
(Central processing unit), 9-1 to 9-M are CPUs 8-1 to 8-1.
If there is a CPU that outputs a command for self-diagnosis to 8-N and does not send a normal response signal within a predetermined time, it is determined that a failure has occurred in that CPU, and the C
SVP that manages the system by disconnecting the PU from the system
(Service processor).

FIG. 2 is a block diagram showing the detailed structure of the SVPs 9-1 to 9-M. In FIG. 2, 11 is a system bus interface which is an interface with the system bus 4, and 12 is a diagnostic bus 5. Diagnostic bus interface serving as an interface of the SVP, 13 is another SVP or CP via the diagnostic bus interface 12.
The reception processing unit that receives various notifications from U8-1 to U-N, 14 analyzes various notifications received by the reception processing unit 13 and, for example, if the received notification is failure information, a failure has occurred. The reception information analysis unit 15 for identifying the SVP or CPU disconnects the SVP or CPU in which the failure has occurred from the system based on the analysis result of the reception information analysis unit 14 and the failure processing information registered in the failure processing table 16 (see below). The diagnostic interface 21 is instructed to update the system configuration registered in the system configuration table 20), and when the diagnostic interface 21 receives a command indicating that a failure has occurred, the system management is stopped. It is a process execution unit (switch control unit).

Reference numeral 16 indicates in advance the initial setting stage of the system.
A failure processing table in which failure processing information as a system is registered, and when the reception information analysis unit 14 analyzes that a CPU has failed, the failure processing execution unit 15 displays the analysis result of the reception information analysis unit 14. Faulty CPU to register
When the reception information analysis unit 14 analyzes the management unit 18 that a failure has occurred in the SVP, the failure processing execution unit 15 registers the analysis result of the reception information analysis unit 14 and the SVP management unit 19, 19 A self-diagnosis unit for diagnosing whether or not it is operating normally, 20 is an SVP existing in this system
And a system configuration table for storing system configuration information of the CPU, 21 updates the storage contents of the system configuration table 20 in accordance with an instruction from the failure processing execution unit 15, and the self-diagnosis unit 19 diagnoses that a self-failure has occurred. Then, the failure processing execution unit 15 is instructed that a failure has occurred, and another SVP to be replaced is selected based on the registered contents of the SVP management unit 18, and the S
A diagnostic interface (switch control unit) for transmitting a command to the VP to start system management (hereinafter referred to as a function substitution request) via the transmission processing unit 22, and 22 is a function substitution transmitted from the diagnostic interface 21. While transmitting the request, the failure processing executing unit 15 instructs the CPU and other SVPs in which no failure has occurred to
A transmission processing unit that transmits a signal indicating that a failure has occurred in a certain CPU. It should be noted that, as described above, the failure processing execution unit 15 and the diagnosis interface 21 constitute a switching control unit.

FIG. 3 is a block diagram showing a detailed configuration of the CPUs 8-1 to 8-M. In the figure, 31 is a system bus interface which is an interface with the system bus 4, and 32 is a diagnostic bus 5. A diagnostic bus interface serving as an interface of the SVP 33, and a reception processing unit 33 for receiving information on an SVP or a CPU having a failure from an SVP currently performing system management (hereinafter referred to as a system SVP) via the diagnostic bus interface 38. , 34 analyzes the information received by the reception processing unit 33. For example, if the received information is CPU failure information, the information is registered in the failed CPU management unit 35 and the system SVP is changed. If so, the information is registered in the system SVP registration table 36, and the system SVP registration table 36 is registered.
If it is an instruction to self-diagnose, the reception information analysis section gives the instruction to the diagnostic interface 38.

Reference numeral 35 is a faulty CPU management section for managing which CPU has a failure, and 36 is the current system SV.
System SVP that registers the SVP operating as P
A registration table, 37 is a self-diagnosis unit for diagnosing whether or not the self is normally operating, and 38 is a self-diagnosis unit 3 when a command to perform self-diagnosis is given from the consultation information analysis unit 34.
7 performs a self-diagnosis, and the diagnostic result of the self-diagnosis unit 37 is transmitted to the system SVP by referring to the system SVP registration table 36. 39 is a diagnostic result of the self-diagnosis unit 37 according to the instruction of the diagnostic interface 38. To the system SVP.

Next, the operation will be described. First, the operation of the system SVP will be described with reference to the flowchart of FIG. 6, but for convenience of explanation, it is assumed that the SVP 9-1 is the system SVP.

First, the SVP 9-1 recognizes the SVP which is the system SVP by referring to the SVP management unit 18 at the time of startup (step ST1).
1), it determines whether or not it is the system SVP (step ST12). That is, the SVP management unit 18 has a system SVP registration table 18c as its component, and the system SVP registration table 18c includes the system SVs in the current system as shown in FIG.
P is registered (incidentally, flag 1 is SVP9-1,
The flag 2 is SVP9-2, ..., The flag M is SVP9.
(Corresponding to -M), the SVP 9-1 can recognize whether or not it is the system SVP by referring to the SVP management unit 18.

As described above, since the SVP 9-1 is registered as the system SVP, the following steps ST
13 and subsequent processes are performed, but other SVPs are system SVPs.
However, the processing in FIG. 6 ends. And
Since the SVP 9-1 is a system SVP, the diagnostic interface 21 refers to the system configuration table 20, and thus the SV connected to the current system.
The P and CPU are recognized, and a command to the effect that self-diagnosis should be performed is transmitted to the connected SVP and CPU via the transmission processing unit 22.

Then, the SVP 9-1 waits for a response of the self-diagnosis result from the SVP and the CPU, which have transmitted the command for the self-diagnosis, and receives it when the response signal is transmitted. The processing unit 13 receives the signal, and the reception information analysis unit 14 analyzes the response signal (normal response signal or abnormal response signal), but does not transmit the normal response signal indicating normality within a predetermined time. Regarding the SVP and CPU, it is determined that a failure has occurred.

If the S connected to the current system
If at least one of the VPs and CPUs has a failure, the processing for disconnecting the failed SVP or CPU from the system is performed (steps ST15, 16, 1).
7). In case of SVP failure (for example, SV
Assuming that a failure has occurred in P9-3), as shown in FIG. 5, the flag 3 of the failure management table 18a in the SVP management unit 18 of the SVP 9-1 is turned on, and the SVP 9-
It is made possible to identify that a failure has occurred in step 3 (step ST15). Then, the failure processing execution unit 15 of the SVP 9-1 refers to the failure processing table 16,
The diagnostic interface 21 is instructed to update the system configuration registered in the system configuration table 20, and the diagnostic interface 21 selects the SVP 9-3 from among the system configurations registered in the system configuration table 20. It is deleted (step ST16). In addition, the diagnostic interface 21 is used for other normal SVP9.
-2, 9-4 to 9-M and CPUs 8-1 to 8-N are notified that the SVP 9-3 has failed (step ST17).

On the other hand, if the CPU fails (for example, if the CPU 8-3 fails), the flag 3 of the failed CPU management unit 17 of the SVP 9-1 is set as shown in FIG. It is turned on so that it can be identified that a failure has occurred in the CPU 8-3 (step ST
15). Then, similarly to the case of SVP, the system configuration table 20 is updated (step ST16), and other normal SVPs 9-2 to 9-M and CPUs 8-1 and 8-
The CPU 8-3 is notified that the failure has occurred in the CPUs 8-3 to 2, 8-4 to 8-N (step ST17).

Next, the SVP 9-1 is a substitute SV that becomes a system SVP when a failure occurs in itself from the normal SVP in advance, in case the failure occurs in the SVP 9-1.
P is selected (step ST18). That is, by referring to the failure management table 18a of the SVP management unit 18, a normal SVP is recognized (in this case, all other than SVP3 are recognized as normal), and a certain rule, for example, an operation program of the diagnostic interface 21 is recognized. , The priority order of the alternative system SVP candidates for each SVP is defined, and the alternative system SVP is selected accordingly (step ST18).

Here, the alternative system SVP is SVP9.
If the setting result of −2 is obtained, the selection result is shown in FIG.
It is registered in the management table 18b (step ST19).

Next, the SVP 9-1 carries out self-diagnosis by the self-diagnosis unit 19 (step ST20). If no fault has occurred, the process returns to step ST13 to continue the series of processes, but a fault has occurred. In this case, the diagnosis interface 21 notifies the failure processing execution unit 15 of the occurrence of a failure, thereby stopping the system management and referring to the alternative system SVP management table 18b to perform the replacement. The system SVP is recognized, and the substitute system SVP (SVP9-2) is notified of the function substitution request via the transmission processing unit 22. Further, the other SVPs 9-3 to 9-M are notified of the update request via the transmission processing unit 22 (step ST2).
2).

Next, the operation of the standby SVP will be described with reference to the flow chart of FIG. First, SVP9
-2-9-M, the reception processing unit 13 receives the interrupt signal via the diagnostic bus interface 12 (step S
T31), the interrupt signal is the system SVP (SVP9
The reception information processing unit 14 analyzes whether or not the interrupt signal is transmitted from -1) (step ST32).

If the interrupt signal is transmitted from a device other than SVP 9-1, the process according to the interrupt signal is performed (step ST33), which will not be described in detail. On the other hand, if the interrupt signal is transmitted from the SVP 9-1, further analysis is performed and the following processing is performed.

That is, if the interrupt signal indicates that a failure has occurred in a certain CPU (step ST3).
3), failure processing execution unit 1 in SVP 9-2 to 9-M
5 is a CPU in which a failure has occurred in each failed CPU management unit 17.
Is registered (step ST34). As described above, in this example, the CPU 8-3 is assumed to have failed, so the flag 3 of each failed CPU management unit 17 is turned on. Then, the failure processing execution unit 1 of the SVPs 9-2 to 9-M
5 gives an instruction to the diagnostic interface 21 to update the system configuration registered in the system configuration table 20 by referring to each failure processing table 16, and the diagnostic interface 21 registers in the system configuration table 20. The CPU 8-3 is deleted from the existing system configurations (step ST35).

Next, if the interrupt signal is a notification indicating that a failure has occurred in a certain SVP (step ST3).
6), failure processing execution unit 1 in SVP 9-3 to 9-M
5 registers the faulty SVP in the fault management table 18a of each SVP management unit 18 (step ST3).
7). As described above, in this example, since it is assumed that the SVP 9-3 has a failure, the failure management table 1
The flag 3 of 8a is turned on. In addition, SVP9-3 is
Since the SVP has a failure, no such notification is given. Also in this case, the failure processing execution unit 15 of the SVPs 9-2 to 9-M refers to the respective failure processing tables 16 so that the system configuration registered in the system configuration table 20 is set to the diagnostic interface 21. Diagnostic interface 21 with instructions to update
Deletes SVP9-3 from the system configurations registered in the system configuration table 20 (step ST35).

Next, the interrupt signal is the system SVP.
If it is a notification regarding the update request (step ST3
8), failure processing execution unit 1 in SVP 9-2 to 9-M
5 registers the alternative system SVP in the alternative system SVP management table 18b of each SVP management unit 18 (step ST39). As mentioned above, in this example, SV
Since P9-2 is selected as the alternative system SVP, flag 2 of the alternative system SVP management table 18b is selected.
While turning on the SVP which is the current system SVP
9-1 is no longer the system SVP, so the flag 1 of the alternative system SVP management table 18b is turned off.

Next, the interrupt signal is the system SVP.
If it is a notification regarding the function substitution request of (step ST4
0), failure processing execution unit 1 in SVP 9-2 to 9-M
5 notifies each diagnostic interface 21 that there has been a function replacement request for the system SVP. As a result, each diagnostic interface 21 refers to the alternative system SVP management table 18b, and the alternative system SVP is the SVP.
9-2 is recognized (step ST41). Here, each diagnostic interface 21 is replaced by the alternative system S.
As a result of referring to the VP management table 18b, only when the self recognizes that it is the alternative system SVP, the process proceeds to step ST12 in FIG. 6 and system management is started as the system SVP. In this example, only the SVP 9-2 shifts to the processing of step ST12 in FIG.

Finally, the interrupt signal indicates that the system SV
If it is a self-diagnosis request from P (step ST42),
The failure processing execution unit 15 in the SVPs 9-2 to 9-M notifies each diagnostic interface 21 of the self-diagnosis request. As a result, each diagnostic interface 21
Causes the self-diagnosis unit 19 to perform self-diagnosis (step ST43). If a failure has occurred as a result of the self-diagnosis, an abnormal response signal is transmitted to the system SVP (step ST45), and if no failure has occurred, a normal response signal is transmitted to the system SVP (step ST45). S
T46).

Next, the operation of the CPU will be described with reference to the flowchart of FIG. First, the CPUs 8-1 to 8
-N, the reception processing unit 33 receives the interrupt signal through the diagnostic bus interface 32 (step ST5).
1) If the interrupt signal is a self-diagnosis request from the system SVP (step ST52), the reception information analysis unit 3
4 notifies each diagnostic interface 38 that there is a self-diagnosis request. Thereby, each diagnosis interface 38 causes the self-diagnosis unit 37 to perform self-diagnosis (step ST53).

If a failure has occurred as a result of self-diagnosis,
Failed CPU management unit 35 (SVP failed CPU management unit 17
The fact that a failure has occurred is registered in the same configuration (step ST55). In this example, since the failure occurs in the CPU 8-3, only the CPU 8-3 performs the process of step ST55, and the faulty CPU management unit 3
The flag 3 of 5 is turned on.

Then, each diagnostic interface 38 is
A message corresponding to the diagnosis result is transmitted to the system SVP (step ST56). Specifically, if a failure has occurred, an abnormal response signal is transmitted, and if no failure has occurred, a normal response signal is transmitted.

Next, if the interrupt signal is a notification indicating that a failure has occurred in a certain CPU (step ST5).
7), the reception information analysis unit 3 in the CPUs 8-1 to 9-N
4 is a CPU in which a failure has occurred in each failed CPU management unit 35
Is registered (step ST58). As described above, in this example, it is assumed that the CPU 8-3 has failed, so the flag 3 of each failed CPU management unit 57 is turned on. Since the CPU 8-3 is the CPU in which the failure has occurred, such notification is not issued.

Finally, the interrupt signal indicates that the system SV
If it is a notification regarding the update request of P (step ST5
9), the reception information analysis unit 3 in the CPUs 8-1 to 8-N
4 registers the alternative system SVP in each system SVP registration table 36 (the same configuration as the alternative system SVP management table 18b of the SVP) (step ST6).
0). As described above, in this example, since the SVP 9-2 is selected as the alternative system SVP, while the flag 2 of the system SVP registration table 36 is turned on,
The current system SVP SVP9-1 is the system S
Since it is no longer a VP, the system SVP registration table 36
The flag 1 of is turned off.

If the interrupt signal does not correspond to any of the above, details will not be described, but processing corresponding to the interrupt signal is performed (step ST61).

As is apparent from the above description, according to the first embodiment, when a certain CPU fails, it has not only a function of disconnecting the failed CPU from the system, but also a failure of the system SVP. If a system SVP occurs, it has a function of replacing the system SVP with another SVP, so that stable system operation is possible without adversely affecting the system due to the occurrence of a system SVP failure. .

Example 2. FIG. 9 is a configuration diagram showing a detailed configuration of an SVP of a computer apparatus according to another embodiment of the present invention. In the figure, 40 is a load factor measuring unit for measuring its own load condition, and 41 is a normal operation of the self. It is a self-diagnosis unit that diagnoses whether or not the failure occurs, and at the time of the diagnosis, if the load condition measured by the load factor measurement unit 40 exceeds a predetermined value, the self-diagnosis unit determines that a failure has occurred. .

Next, the operation will be described. First, the load factor measuring unit 40 measures the time during which the SVP is operating within a predetermined time, and calculates the ratio of the operating time (load condition). On the other hand, the self-diagnosis unit 41 diagnoses whether or not the self is operating normally in the same manner as the self-diagnosis unit 19 in the first embodiment, and at the time of the diagnosis, it is measured by the load factor measurement unit 40. If the load status exceeds a predetermined value, it is considered that a failure has occurred even if the failure has not occurred. For example, if the load condition of the self exceeds 50% of the operating capacity of the SVP, it is considered that the self has a failure. This allows the diagnostic interface 21 to replace the system S
When selecting a VP, the load condition exceeds the specified value SV
Since P is excluded from the selection candidates, only SVPs having a low load condition at present, that is, SVPs that do not cause much burden even if the function of the system SVP is added, become the selection candidates for the system SVP. Therefore, the alternative system SV
Since P can operate reliably,
Further, it is possible to surely minimize the adverse effect on the system. Note that, except for the load factor measurement unit 40 and the self-diagnosis unit 41, the description is omitted because it is the same as in the first embodiment.

Example 3. In the above embodiment, in each SVP,
Failure processing table 16, failed CPU management unit 17, SVP
The management unit 18 and the system configuration table 20 are arranged, and each CPU has a failed CPU management unit 35 and a system S.
Although the arrangement of the VP registration table 36 is shown, as shown in FIG. 10, each of these tables and the management unit is configured as a shared memory 42 including, for example, a battery backup memory, and the shared memory 4 is also provided.
It is also possible to provide a maintenance tool 43 for managing 2 and perform unified management of information online. This further improves maintainability and flexibility. The detailed configuration of each SVP and CPU in the third embodiment is shown in FIGS. 11 and 12.

[0048]

As described above, according to the first aspect of the present invention, when the system is managed by itself, if the self-diagnosis unit diagnoses that a failure has occurred, the system management is performed. Since it is configured to output the command to start system management to the standby service processor when it is stopped, the faulty service processor is disconnected from the system, and the standby service processor is managed by the system. With such a configuration, even if a failure occurs in the service processor, the adverse effect on the system can be minimized.

According to the second aspect of the present invention, when the system is managed by itself, by outputting a command to the self-diagnosis to each standby service processor, The diagnosis result of each standby service processor is known based on the presence / absence of a normal response signal from the standby service processor. When the self-diagnosis unit diagnoses that a failure has occurred, system management is stopped. Since it is configured to output a command to start system management to the normal standby service processor based on the diagnosis result, the faulty service processor is disconnected from the system and the normal standby The service processor for use in the system now manages the system, and more reliably than the invention of claim 1, adversely affects the system. There is an effect, such as can be kept to a minimum.

Further, according to the invention of claim 3, when the load condition measured by the load factor measuring unit exceeds a predetermined value, it is considered that a failure has occurred in itself, so that the operation is performed. Only service processors with low rates are selected as alternative service processors,
According to the invention of claim 1, there is an effect that the adverse effect on the system can be more surely suppressed to the minimum.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a computer apparatus according to an embodiment of the present invention.

FIG. 2 is a configuration diagram showing a detailed configuration of SVPs 9-1 to 9-M.

FIG. 3 is a configuration diagram showing a detailed configuration of CPUs 8-1 to 8-M.

FIG. 4 is a configuration diagram showing a configuration of a failed CPU management unit.

FIG. 5 is a configuration diagram showing a configuration of an SVP management unit.

FIG. 6 is a flowchart showing the operation of the system SVP.

FIG. 7 is a flowchart showing the operation of a standby SVP.

FIG. 8 is a flowchart showing the operation of the CPU.

FIG. 9 is an SV of a computer device according to another embodiment of the present invention.
It is a block diagram which shows the detailed structure of P.

FIG. 10 is a block diagram showing a computer system according to another embodiment of the present invention.

11 is a configuration diagram showing a detailed configuration of an SVP of the computer device of FIG.

12 is a configuration diagram showing a detailed configuration of a CPU of the computer device of FIG.

FIG. 13 is a configuration diagram showing a conventional computer apparatus.

FIG. 14 is a flowchart showing an operation of a conventional computer device.

[Explanation of symbols]

 8-1 to 8-N CPU (Central Processing Unit) 9-1 to 9-M SVP (Service Processor) 15 Failure Processing Execution Unit (Switch Control Unit) 19, 41 Self-Diagnosis Unit 21 Diagnostic Interface (Switch Control Unit) 40 Load factor measurement section

Claims (3)

[Claims] 1. A self-diagnosis function for diagnosing whether or not the self is operating normally, and a self-diagnosis result,
When operating normally, a plurality of central processing units that send normal response signals and a command to self-diagnose each of the above central processing units are output, and normal response signals are sent within a predetermined time. If there is a central processing unit that does not send, it is determined that a failure has occurred in the central processing unit, and the central processing unit is disconnected from the system. In addition to the standby service processor, a self-diagnosis unit that diagnoses whether or not it is operating normally, and a self-diagnosis unit that causes self-diagnosis when the system is managed by itself If it is diagnosed that the system management is stopped, the system management is stopped and the standby service processor is instructed to start the system management. On the other hand, each service processor is equipped with a switching controller that starts system management when it receives a command from the service processor that is performing system management to start system management A computer device characterized by the above. 2. A self-diagnosis function for diagnosing whether or not the self is operating normally, and a self-diagnosis result,
When operating normally, a plurality of central processing units that send normal response signals and a command to self-diagnose each of the above central processing units are output, and normal response signals are sent within a predetermined time. If there is a central processing unit that does not send, it is determined that a failure has occurred in the central processing unit, and the central processing unit is disconnected from the system. In addition to multiple standby service processors, a self-diagnosis unit for diagnosing whether or not the self is operating normally, and a self-diagnosis service processor for each of the above standby service processors when the system is managed by itself. By outputting a command to the effect that self-diagnosis should be performed, the standby processor Over previously know the service processor diagnostic result, with the self failure is diagnosed as occurring to stop the system managed by the self-diagnosis unit,
Based on the above diagnosis result, it outputs a command to the normal service processor for standby to start system management. On the other hand, when it is in standby, the service processor performing system management performs system management. A computer apparatus, wherein each service processor is provided with a switching control unit that starts system management when receiving a command to start the service. 3. Each of the service processors is provided with a load factor measuring unit that measures its own load condition, and the self-diagnosis unit is configured to determine whether the load condition measured by the load factor measuring unit exceeds a predetermined value. 3. The computer system according to claim 1, wherein the computer is regarded as having a failure.