JP2012194790A - Fault detection method, control device and multiprocessor system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem in which detecting a fault that occurs on a communication path in a device whose hardware is divided into multiple partitions is time consuming since it is difficult to detect the location of the fault and necessary to identify a faulty part by manually changing parts.SOLUTION: A control device in this invention includes: communication means that communicates with an external device using a first interface; detection means that detects the occurrence of a communication fault; first count means that counts the number of faults detected by the detection means; and count control means that controls the first count means to count up and controls count means in the external device to count up using a second interface, when the detection means detects the fault.

Description

  The present invention relates to a failure detection method for appropriately processing a failure location that occurs in a system in which a plurality of components or devices are connected, a control device used in the method, and a system thereof.

  In recent years, the shift from single processor systems to multiprocessor systems has been accelerated in order to improve the execution speed of applications through parallel processing. In addition, the number of processor modules installed in a multiprocessor system is also increasing, and the paths in the system are complicated.

  If the path becomes complicated in this way, it becomes difficult to specify the location of the failure when a failure occurs. Until now, various techniques for facilitating the identification of a fault occurrence location have been developed. For example, Patent Document 1 describes whether a failure occurs between DCE and DCE when a failure occurs in a system in which a plurality of data terminal devices (DTE) are connected via line termination devices (DCE). A technique for easily specifying whether a failure has occurred between DTE and DCE is disclosed.

  Further, in Patent Document 2, in a system in which a central control unit accesses a storage device group by a plurality of paths, a path in which a failure has occurred is not blocked based only on the number of occurrences of the failure, but the remaining usable paths A technique for changing the blocking condition according to the number is disclosed.

  Patent Document 3 discloses a technology for sharing a device in a plurality of processor systems that operate with different operation systems as a related technology. In general, in such a shared device system, each processor system individually has failure information regarding a failure that occurs when the shared device is used. If a failure occurs, only the failure information is used. Determine the suspicion of the failure location and take action.

JP 2003-18329 A JP 2000-148655 A Japanese Patent Laid-Open No. 2004-246779

  Conventionally, an apparatus that is hardware-divided into a plurality of partitions determines a suspicion only from failure information such as an access error log acquired in each partition. In such a failure processing method, if a failure occurs on the path from the partition to the shared device, it is difficult to identify the location of the failure, and parts are identified (isolated) by manual parts replacement. Because it was necessary, it took time to make a suspicion. Also, in a shared device system with a large number of partitions and a complicated path, the conventional countermeasures cannot identify the location of the failure in some cases, and all parts on the suspected path There was also a situation that had to be replaced.

  For such a situation, even if the technique of the above-mentioned patent document is used, it is not an effective solution. Therefore, a new failure detection method and a control device therefor have been demanded.

  An object of the present invention is to provide a failure detection method capable of easily narrowing down suspected devices when a failure occurs on a communication path, a control device used in the failure detection method, and a multiprocessor system.

  The multiprocessor system of the present invention is a multiprocessor system including a plurality of hardware-divided partitions and a shared device accessed from the plurality of partitions, wherein the plurality of partitions and the shared device are The plurality of partitions and the shared device are connected by a first interface used for data communication and a second interface used for failure detection, and the plurality of partitions comprise a counting means for counting failures, and the plurality of partitions are And a control means for performing control to change the count value of the counting means included in the partition related to the data communication and the shared device when a failure occurs in the data communication.

  The control device according to the present invention includes a communication unit that communicates with an external device using the first interface, a detection unit that detects that a failure has occurred in the communication, and a failure detected by the detection unit. A first counting means for counting and a control for counting up the first counting means when a failure is detected by the detecting means, and a control for counting up the counting means of the external device; And a count control means which performs using

  In addition, the control device of the present invention receives communication information from another control device, and based on the input communication failure information, a communication unit that communicates with an external device using the first interface; When a failure is detected in a communication performed by the communication unit based on the information on the communication failure, and a detection unit that detects that a failure has occurred in the communication performed by the communication unit based on the information on the communication failure, When the control for counting up the counting unit of the external device is performed using the second interface, and no failure is detected in the communication performed by the communication unit based on the information regarding the communication failure, the external device Count control means for performing control to count down the count means using the second interface; That.

    The failure processing method of the present invention includes a communication step of communicating with an external device using the first interface, a detection step of detecting that a failure has occurred in the communication, and a failure detected in the detection step And a count control step of performing control to count up the counting means of the external device using the second interface when the failure is detected.

  According to the present invention, it is possible to provide a failure detection method capable of easily narrowing down suspected devices when a failure occurs on a communication path, a control device used in the failure detection method, and a multiprocessor system.

1 is a block diagram showing a configuration of a communication system according to a first exemplary embodiment. FIG. 3 is a block diagram showing a configuration of a multiprocessor system according to a second exemplary embodiment. FIG. 10 is a sequence diagram illustrating a failure handling operation of the multiprocessor system according to the second embodiment. FIG. 6 is a block diagram showing a configuration of a modification of the multiprocessor system according to the second exemplary embodiment;

(Embodiment 1)
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a communication system having a failure handling function according to the present invention.

  The communication system according to Embodiment 1 of the present invention is roughly composed of three parts: a partition 100, a partition 200, and a partition 300.

  The partition 100 is one block in which functional modules to be described later are gathered, and may be a device or a part, or may be a partition divided by hardware. The partition 100 includes a communication module 101, a failure detection module 102, a count control module 103, and a count module 104.

  The communication module 101 performs communication such as transmission and reception of signals and data with the communication module 301 in the partition 300 using the control / data interface 801 which is the first interface. Further, communication such as transmission and reception of signals and data is performed with the communication module 201 in the partition 200 as necessary.

  The failure detection module 102 detects that a failure has occurred in communication performed by the communication module 101. As a failure detection method, for example, when failure information is sent from a communication destination partition or when a reply to a request is not returned for a certain period of time, it is determined that a communication failure has occurred, and the count control module 103 has failed. Notify that has occurred.

  When receiving a notification from the failure detection module 102 that a failure has occurred in the communication performed by the communication module 101, the count control module 103 performs control to count up the count modules belonging to the partition related to the communication. For example, when a failure occurs between the communication module 101 and the communication module 301, the partitions related to the communication are the partition 100 and the partition 300. Accordingly, the count control module 103 performs control to count up the count module 104 belonging to the partition 100 and the count module 302 belonging to the partition 300. Here, the control is performed using a diagnostic interface 802 provided separately from the control / data interface 801 in which the communication is performed. A dedicated interface may be separately provided between the count module 104 and the count control module 103, and the count-up control in the same partition and the count-up control for the external partition may be performed using different interfaces. Including the case where the dedicated interface is provided in the same partition, an interface for performing count-up control is defined as a second interface, and an interface for performing communication is defined as a first interface.

  The count module 104 counts the communication failure. The count in the count module 104 is performed based on count-up control performed by a count control module 103 described later. Further, when a failure occurs in communication from the partition 200 to the partition 100, the count in the count module 104 may be configured to be performed by the count control module 203 belonging to the partition 200.

  Next, the partition 200 will be described. The partition 200 has the same configuration as the partition 100. The partition 200 includes a communication module 201, a failure detection module 202, a count control module 203, and a count module 204.

  The communication module 201 communicates with the communication module 301 in the partition 300 using the control / data interface 801 that is the first interface. Further, communication is performed with the communication module 101 in the partition 100 as necessary.

  The failure detection module 202 detects that a failure has occurred in communication performed by the communication module 201. As a failure detection method, the same method as the failure detection module 102 can be used.

  The count control module 203 performs control to count up the count modules belonging to the partition related to the communication when a failure occurs in the communication performed by the communication module 201. Here, like the count control module 103, this control is performed using a diagnostic interface 802 provided separately from the control / data interface 801 through which the communication is performed. As in the case of the partition 100, a dedicated interface is separately provided between the count module 204 and the count control module 203, and count-up control in the same partition and count-up control for the external partition are performed using different interfaces. Also good.

  The count module 204 counts the communication failure. Counting in the count module 204 is performed based on count-up control performed by the count control module 203. Further, when a failure occurs in communication from the partition 100 to the partition 200, the count in the count module 204 may be performed by the count control module 103 belonging to the partition 100.

  Next, the partition 300 will be described. The partition 300 is one block in which functional modules to be described later are collected, and may be a shared device, for example. The partition 300 includes a communication module 301 and a count module 302.

  The communication module 301 communicates with the communication module 101 belonging to the partition 100 and the communication module 201 belonging to the partition 200 using the control / data interface 801.

  The count module 302 counts the communication failure based on count-up control performed by the count control module included in the external partition using the diagnostic interface 802. Further, the counted value is read to the count control module based on read control performed by the count control module as necessary.

  According to this configuration, when a communication failure occurs in communication performed between these multiple partitions, the count value of the count module belonging to the partition causing the failure is greater than the count value of the count module belonging to another partition. Will also grow. Therefore, it is possible to easily identify the location where the failure has occurred.

  In the above description, the case where the count value of the count module is counted up when a failure occurs has been described. However, when the failure does not occur, the count value of the count module may be counted down. .

(Embodiment 2)
The second embodiment relates to a multiprocessor system in which the failure handling function according to the first embodiment is introduced. Hereinafter, it demonstrates using drawing. A part of the description of the overlapping parts is omitted.

  FIG. 2 shows the configuration of the multiprocessor system 1000 according to the second embodiment. The multiprocessor system 1000 is hardware-divided into two partitions, a partition 100 and a partition 200. The multiprocessor system 1000 is roughly composed of these two partitions 100, 200, a relay device 410, and an optical drive 310. The partition 100 and the partition 200 correspond to the partition 100 and the partition 200 in FIG. 1, respectively, and the optical drive 310 corresponds to the partition 300 in FIG. The relay device 410 is a device for relaying communication between the partition 100 and the partition 200 and the partition 300, and is arranged on the motherboard 400. Hereinafter, each block will be described in detail.

  The partition 100 includes a processor module 110 and a service processor 120. The processor module 110 includes a processor that executes and controls instructions and a main memory that stores data. A request signal is transmitted to the optical drive 310 via the service processor 120 based on the request executed by the processor module 110.

  The service processor 120 is a processor for assisting the processor module 110, and performs a failure detection process and a startup process of the processor module 110, which will be described later. The service processor 120 includes a communication unit 121, a failure detection unit 122, a count control unit 123, and a counter 124 inside.

  The communication unit 121 communicates with a communication unit of an external device. The communication includes communication transfer from the processor module 110 in addition to communication to an external device executed by the service processor 120 itself. The communication unit 121 inputs a request signal sent from the processor in the processor module 110 and transfers the request signal to the request destination apparatus.

  The failure detection unit 122 detects whether a failure has occurred in the communication performed by the communication unit 121. The failure detection unit 122 receives an access error packet, which is one piece of failure information, from the communication unit 121, analyzes the packet, and identifies on which communication path the failure has occurred. In addition, if the communication unit 121 measures the time after outputting the request signal and the reply signal is not returned from the output destination of the request signal even after a predetermined time has elapsed, a failure has occurred in the communication path. Judge. When the failure detection unit 122 detects that a failure has occurred in communication performed by the communication unit 121 by the above method or the like, the failure detection unit 122 notifies the count control unit 123 that a communication failure has occurred. The information notified from the failure detection unit 122 to the count control unit 123 includes the failure occurrence time, the number of failure occurrences, the fatality of the failure, information on the communication path in which the failure has occurred, and the device that needs to perform count-up control. Necessary necessary information is selected from the address information of the counter and notified.

  When the count control unit 123 receives a notification regarding a communication failure from the failure detection unit 122, the count control unit 123 performs control to count up a counter belonging to the partition or external device related to the communication. For example, when a failure occurs between the service processor 120 and the optical drive 310, the partition and the external device related to the communication are the partition 100, the relay device 410, and the optical drive 310. Therefore, the count control unit 123 performs control to count up the counter 124 included in the service processor 120, and performs control to count up the counter 313 and the counter 412 included in a device outside the service processor 120 using the diagnosis interface 802. Do. In addition, the count control unit 123 inputs a read instruction from a control unit (not shown), and performs control to read a value stored in the counter from a counter connected via the diagnosis interface 802. The read counter values are presented to the user via a user interface (not shown) as necessary. In addition, the count control unit 123 performs control to count down a counter of a device existing on the communication path as necessary.

  The counter 124 holds a count value counted for the failure detected by the failure detection unit 122. The counter 124 is configured by, for example, a nonvolatile ROM. The value held by the counter 124 is incremented based on the count up control from the count control unit 123.

  Next, the partition 200 will be described. The partition 200 includes a processor module 210 and a service processor 220. Since the partition 200 has the same configuration as the partition 100, the description thereof is omitted. That is, the processor module 210 has the same configuration as the processor module 110, and the service processor 220 has the same configuration as the service processor 120.

  Next, the optical drive 310 will be described. The optical drive 310 is a shared device that receives access from a plurality of partitions, and includes a communication unit 311, a failure detection unit 312, and a counter 313.

  The communication unit 311 communicates with the communication unit of the external device according to a predetermined protocol. For example, a request signal related to data read / write is input from an external device, and a control unit (not shown) performs control related to data read / write with respect to an address specified by the request signal. The communication unit 311 returns a read signal indicating that the read data or access has been completed to the request signal transmission source.

  The failure detection unit 312 detects a failure that has occurred in the optical drive 310 itself. The failure detection unit 312 detects a communication error that has occurred in the communication unit 311 or a failure such as read / write that occurs in a control unit (not shown) as a failure, and includes failure information that summarizes the failure type, failure occurrence time, and the like. create. The failure information created by the failure detection unit 312 is transmitted to a service processor installed in a partition that is divided into hardware as necessary.

  The counter 313 is connected to the diagnostic interface 802 and increments a value to be stored based on count-up control from a count control unit existing outside the optical drive 310. Further, the stored value is decremented by one based on the countdown control from the count control unit. The counter 313 is used for suspicious judgment regarding the failure of the optical drive 310.

  Next, the relay device 410 will be described. The relay device is a device arranged on the mother board 400 in order to relay communication between each partition and the optical drive. The relay device 410 includes a switching control unit 411 and a counter 412.

  The switching control unit 411 exclusively controls access from the service processor 120 or the service processor 220 to the optical drive 310. That is, the switching control unit 411 performs exclusive control so that only one of the partitions can use the optical drive by switching the connection destination of the optical drive 310.

  The counter 412 is connected to the diagnosis interface 802 and increments a value to be stored based on count-up control from a count control unit existing outside the relay device 410. Further, the stored value is decremented by one based on the countdown control from the count control unit. The counter 412 is used for suspicion determination regarding the failure of the relay device 410.

  Next, the operation of the multiprocessor system of the present invention will be described with reference to the drawings. FIG. 3 is a sequence diagram showing the flow of processing in the multiprocessor system of the present invention when a failure occurs in the path from the service processor 120 to the optical drive 310.

  The service processor 120 belonging to the partition 100 executes a request (read or write) to the optical drive 310 and outputs a request signal (S101). The request signal is output from the communication unit 121 in the service processor 120. Here, the service processor 120 enters a reply waiting state for the request (S102).

  The optical drive 310 having received the request signal performs control and processing according to the content of the request, and returns a reply signal (S103). If there is no failure on the path from the service processor 120 to the optical drive 310, a reply to the request from the service processor 120 is returned from the optical drive 310 to the service processor 120 via the relay device 410. On the other hand, if there is a failure on the path from the service processor 120 to the optical drive 310, the request does not reach the optical drive 310, and therefore no reply is returned for the request.

  The failure detection unit 122 in the service processor 120 determines whether a failure has occurred in the communication performed by the communication unit 121 (S104). Specifically, when a reply to the request executed by the communication unit 121 is not returned within a predetermined time, it is determined that a failure has occurred on the communication path from the service processor 120 to the optical drive 310. On the other hand, if the reply is returned within a predetermined time, it is determined that no failure has occurred. When it is determined that a failure has occurred, the failure detection unit 122 notifies the count control unit 123 that a failure has occurred. On the other hand, when it is determined that no failure has occurred, the failure detection unit 122 prepares for failure detection in the next communication without performing special processing.

  When the failure detection unit 122 determines that the failure has occurred, the count control unit 123 in the service processor 120 performs count-up control of the counter 124 managed by itself (S105). The count control unit 123 further performs count-up control of the counter 313 in the optical drive 310 and the counter 412 in the relay device 410 using the diagnostic interface 802 (S106). As a result, the values of the counter 124, the counter 313, and the counter 412 are each “1”. Note that the values of these counters indicate the number of failures that have occurred on the path through the device or component to which the counter belongs.

  Next, the service processor 120 notifies the service processor 220 of the partition 200 of the failure in the path from the partition 100 side to which the service processor 120 belongs and the value of the counter 124 of the service processor 120 (S107). . More specifically, the count control unit 123 reads a value stored in the counter 124 based on a read instruction from a control unit (not shown). The count control unit 123 outputs the read count value to the control unit. The control unit collectively generates diagnosis information including the value of the counter 124 input from the count control unit 123, the content of the request, information regarding which route or device the failure has occurred in, and the like. The diagnostic information is information related to a communication failure that has occurred on the partition 100 side. When receiving the diagnostic information from the control unit, the communication unit 121 outputs the diagnosis information to the partition 200 using the control / data interface.

  The service processor 220 of the partition 200 that has received the diagnostic information executes a request for the optical drive 310 (S108). More specifically, the communication unit 221 of the service processor 220 inputs the diagnostic information output from the communication unit 121 of the service processor 120, and outputs the diagnostic information to a control unit (not shown). The control unit identifies the request destination and request content from the received diagnostic information, and executes the request. By executing the request, a request signal is output from the communication unit 221 to the optical drive 310. Thereafter, the service processor 220 enters a reply waiting state for the request (S109).

  The optical drive 310 having received the request signal performs control and processing according to the request content, and returns a reply signal (S110). If there is no failure on the path from the service processor 220 to the optical drive 310, a reply to the request executed by the service processor 220 is returned from the optical drive 310 to the service processor 220 via the relay device 410. On the other hand, if there is a failure on the path from the service processor 220 to the optical drive 310, the request does not reach the optical drive 310, and therefore no reply is returned for the request.

  The failure detection unit 222 in the service processor 220 determines whether a failure has occurred in communication performed by the communication unit 221 based on the diagnostic information (S111). Specifically, if the reply signal for the request signal output by the communication unit 221 is not returned within a predetermined time, it is determined that a failure has occurred on the communication path from the service processor 220 to the optical drive 310. . On the other hand, when the reply signal for the request signal output from the communication unit 221 is returned within a predetermined time, it is determined that there is no failure on the communication path from the service processor 220 to the optical drive 310. The failure detection unit 222 generates determination information that summarizes the determination result and the address of the counter of the external device related to the communication, and outputs the determination information to the count control unit 223.

  The count control unit 223 performs count-up / count-down control of the counter based on the input determination information. Specifically, when the determination result included in the determination information indicates that there is no failure on the communication path, the count control unit 223 counts down the counter value of the device related to the communication. Control is performed (S112). Here, the count control unit 223 performs control to count down the counter 313 belonging to the optical drive 310 and the counter 412 belonging to the relay apparatus 410 using the diagnosis interface 802.

  Next, the count control unit 223 performs read control for reading values from the counter 313 and the counter 412. The count control unit 223 reads a value from the counter 224 (S113). The read counter values are output to a control unit (not shown).

  The control unit compares the values of these counters with the value of the counter 124 included in the diagnostic information received from the service processor 120 in S107, thereby identifying a suspected part related to a failure that has occurred inside the multiprocessor system 1000 ( S114). Specifically, as a result of the above series of processing, the counter values of the respective devices included in the multiprocessor system 1000 are “1” for the counter 124, “0” for the counter 224, and “0” for the counter 313, respectively. “The counter 412 is set to“ 0 ”. Accordingly, the service processor 120 to which the counter 124 having the largest counter value belongs becomes a suspected part in communication failure.

On the other hand, if the determination result included in the determination information indicates that a failure has occurred on the communication path, the count control unit 223 counts to increment the counter value of the device related to the communication. Up control is performed (S115). Here, the count control unit 223 performs control for counting up the counter 313 belonging to the optical drive 310 and the counter 412 belonging to the relay apparatus 410 using the diagnosis interface 802.

  Next, the count control unit 223 performs read control for reading values from the counter 313 and the counter 412. Further, the count control unit 223 reads a value from the counter 224 (S116). The read counter values are output to a control unit (not shown). The control unit compares the values of these counters with the value of the counter 124 included in the diagnostic information received from the service processor 120 in S107, thereby identifying a suspected part related to a failure that has occurred inside the multiprocessor system 1000 ( S117). Specifically, as a result of the above series of processing, the counter values of the respective devices included in the multiprocessor system 1000 are “1” for the counter 124, “1” for the counter 224, and “2” for the counter 313, respectively. “The counter 412 is set to“ 2 ”. Therefore, the optical drive 310 to which the counter 313 having the largest counter value belongs or the relay device 410 to which the counter 412 belongs becomes a suspected part in communication failure. Which of these is a suspicious part may be specified by comprehensively determining it in combination with an access error log or other failure information acquired separately, or these two suspicious parts may be replaced.

  With the above-described configuration, it is possible to easily narrow down the suspected parts and suspected devices that may have failed, so that it is possible to reduce the number of replacement parts and the system stop time associated with parts replacement.

  In the above description, the countdown is performed according to the reply to the request made by the service processor 220 that has received the diagnostic information. However, the present invention is not limited to this. The countdown may be performed every time the communication performed by the communication unit 221 is successful, or the countdown may not be performed.

  The count control unit 223 is configured to decrement or increment the counter value of the communication-related device connected to the diagnostic interface 802 in S112 or S115, but is not limited thereto. The count control unit 223 may perform control to decrement or increment the value of the counter 224 belonging to the service processor 220 as well. Further, the count control unit 223 may be configured to specify a counter of a device to be counted down based on diagnostic information from a control unit (not shown) and to count down the specified counter. That is, the count control unit 223 may be configured to count down the counters belonging to only the devices in the overlapping portion of the route from the partition 100 and the route from the partition 200.

  In the above description, the configuration in which the value of the counter 124 in the service processor 120 is included in the diagnostic information sent from the service processor 120 to the service processor 220 is not limited to this. When performing control to read the count value from the counter 313 and the counter 412 in S113 or S116, it may be read from the counter 124 together.

  In the above description, the configuration in which the counter belonging to each device is directly connected to the diagnosis interface 802 is shown, but the present invention is not limited to this. For example, a configuration may be adopted in which each device includes a control unit that counts up / down a counter value belonging to the own device, and the control unit is connected to a diagnostic interface. In this case, based on the count-up / count-down instruction signal sent from the count control unit 123 or the count control unit 223 using the diagnostic interface, these in-device control units may increment / decrement the counter value.

  In the above description, a configuration is shown in which counters are arranged in the optical drive and relay device, which are external devices, but the present invention is not limited to this. A configuration in which a storage unit for storing the count value of each device existing in the system may be separately provided. The storage unit is connected to a diagnostic interface and is divided into areas for storing count values of the respective devices. And the structure which can perform control which changes the count value of the applicable apparatus in the said memory | storage part using the said diagnostic control interface may be sufficient.

  The shared device is not limited to the optical drive, and various devices such as an auxiliary storage device and a shared interface can be used.

  In the above description, the configuration in which the count control unit increments or decrements the value of the counter has been described. However, the present invention is not limited to this. Each counter may be a flag storage unit that stores a value of “0” or “1”. In this case, the count control unit becomes a flag control unit, and the flag control unit raises or lowers the flag of the flag storage unit existing in the communication path based on the detection result in the failure detection unit. Such a configuration may be adopted.

  Further, the suspected device and the suspected component in the above description indicate a failure occurrence location that is estimated as a cause of failure when a failure occurs.

  The above description is only an example of a multiprocessor system, and various modifications can be made. FIG. 4 is a block diagram showing a modification of the multiprocessor system according to the present invention. In the multiprocessor system of FIG. 4, the hardware is divided into N partitions from the partition 100 -A to the partition 100 -N. Each partition includes a plurality of processor modules, and further includes a service processor that assists these processor modules. Further, M relay devices from the relay device 410-A to the relay device 410-M are relayed and connected to L shared devices from the shared device 310-A to the shared device 310-L. Here, the solid line indicates the control / data interface 801, and the alternate long and short dash line indicates the diagnosis interface 802. It may be expanded in this way.

  In addition, the present invention is not limited to the above-described embodiment, and can be modified as appropriate without departing from the spirit of the present invention. For example, the following configuration can be adopted.

(1) A multiprocessor system including a plurality of hardware-divided partitions and a shared device accessed from the plurality of partitions, wherein the plurality of partitions and the shared device are used for data communication. The plurality of partitions and the shared device are provided with counting means for counting failures, and the plurality of partitions are connected in the data communication. A multiprocessor system, comprising: a control unit that performs control to change a count value of a count unit included in the partition related to the data communication and the shared device when a failure occurs.
(2) Communication means for communicating with an external device using the first interface, detection means for detecting that a failure has occurred in the communication, and first count means for counting the failure detected by the detection means And, when a failure is detected by the detection means, performs control to count up the first count means, and performs control to count up the count means of the external device using a second interface. And a control device.
(3) The control according to (2), wherein the count control means performs control to count down the count means of the external device using the second interface when no failure is detected by the detection means. apparatus.
(4) The communication unit communicates with the external device via a relay device, and the count control unit performs control to count up the first count unit when a failure is detected by the detection unit. The control device according to (2), wherein the control is performed using a second interface to count up the counting unit of the external device and the relay device.
(5) Communication information input from another control device, communication means for communicating with an external device using the first interface based on the input communication failure information, and information related to the communication failure Detecting means for detecting that a failure has occurred in the communication performed by the communication means, and counting means of the external device when a failure is detected in the communication performed by the communication means based on the information relating to the communication failure Control for counting up the count means of the external device when no fault is detected in the communication performed by the communication means based on the information relating to the communication fault. And a count control means for performing the above operation using the second interface.
(6) The apparatus further includes a determination unit that determines the suspected device, wherein the count control unit further performs control to read a count value from the count unit connected to the second interface, and the determination unit includes the count unit. The control device according to (5), wherein the suspicious device is determined based on the count value read by the control means.
(7) The information related to the communication failure includes a count value counted by the counting unit of the other control device, and the determination unit includes the count value included in the information related to the communication failure and the count control unit. The control device according to (6), wherein the suspicious device is determined by comparing with the count value read by.
(8) The communication unit communicates with the external device by outputting a request signal and a reply signal to the request signal, and the detection unit is within a predetermined time after the request signal is output. The control device according to any one of (2) to (7), wherein it is determined that a failure has occurred in the communication when a reply signal to the request signal is not input.
(9) a communication step of communicating with an external device using the first interface, a detection step of detecting that a failure has occurred in the communication, a counting step of counting the failure detected in the detection step, And a count control step of performing control to count up the counting means of the external device using a second interface when the fault is detected.
(10) A reading step for reading a count value from the counting means connected to the second interface, and a suspected device are determined based on the count value read in the reading step after the count control step. The failure detection method according to (9), further comprising a determination step.
(11) The communication unit inputs information regarding a failure detected by the detection unit included in the external device itself from the external device using the first interface, and the determination unit includes the information regarding the failure and the information about the failure The control device according to (6) or (7), wherein the suspicious device is determined from the read count value.
(12) An input step for inputting information on communication failure from another communication device, and communication for communicating with a predetermined external device using the first interface based on the information on communication failure input in the input step A determination step of determining whether a failure has occurred in the communication, and a counting means of the external device involved in the communication is counted up when it is determined in the determination step that a failure has occurred in the communication Control is performed using the second interface, and when it is determined in the determination step that no failure has occurred in the communication, control for counting down the counting means of the external device related to the communication is performed using the second interface. And a count control step.
(13) Subsequent to the count control step, a read step for reading the count value from the count means connected to the second interface, and determining the suspected device based on the count value read in the read step The failure processing method according to (12), further comprising a suspected device determination step.
(14) The control device according to any one of (2) to (7), wherein the communication unit relays a request signal from an external processor.

100 partition 101 communication module 102 fault detection module 103 count control module 104 count module 110 processor module 120 service processor 121 communication unit 122 fault detection unit 123 count control unit 124 counter 200 partition 201 communication module 202 fault detection module 203 count control module 204 count Module 210 Processor module 220 Service processor 221 Communication unit 222 Fault detection unit 223 Count control unit 224 Counter 300 Partition 301 Communication module 302 Count module 310 Optical drive 311 Communication unit 312 Fault detection unit 313 Counter 400 Motherboard 410 Relay device 411 Switching control unit 412 Counter 8 1 data interface 802 diagnostic interface 1000 multiprocessor system

Claims (10)

Multiple partitions partitioned by hardware,
A shared device accessed from the plurality of partitions;
A multiprocessor system comprising:
The plurality of partitions and the shared device are connected by a first interface used for data communication and a second interface used for failure detection,
The plurality of partitions and the shared device comprise counting means for counting failures,
The plurality of partitions include a control unit that performs control to change a count value of a partition related to the data communication and a counting unit included in the shared device when a failure occurs in the data communication.
Multiprocessor system. A communication means for communicating with an external device using the first interface;
Detecting means for detecting that a failure has occurred in the communication;
First counting means for counting faults detected by the detecting means;
A count control means for performing control to count up the first count means when a failure is detected by the detection means, and performing control to count up the count means of the external device using a second interface; ,
A control device comprising: The count control means performs control to count down the count means of the external device using the second interface when no failure is detected by the detection means.
The control device according to claim 2. The communication means communicates with the external device via a relay device,
The count control means performs control to count up the first count means when a failure is detected by the detection means, and performs control to count up the count means of the external device and the relay device. Using the interface of
The control device according to claim 2. Communication means for inputting information related to a communication failure from another control device, and communicating with an external device using the first interface based on the information related to the input communication failure;
Detecting means for detecting that a failure has occurred in communication performed by the communication means based on information on the communication failure;
When a failure is detected in communication performed by the communication means based on the information related to the communication failure, control for counting up the counting means of the external device is performed using the second interface, and the information related to the communication failure A count control means for performing control to count down the count means of the external device using the second interface when no failure is detected in communication performed by the communication means based on
A control device comprising: It further comprises determination means for determining the suspected device,
The count control means further performs control to read a count value from the count means connected to the second interface,
The determination means determines the suspected device based on the count value read by the count control means;
The control device according to claim 5. The information related to the communication failure includes a count value counted by the counting means of the other control device,
The determination unit determines the suspected device by comparing the count value included in the information regarding the communication failure and the count value read by the count control unit;
The control device according to claim 6. The communication means outputs a request signal and communicates with the external device by inputting a reply signal for the request signal,
The detection means determines that a failure has occurred in the communication when a reply signal to the request signal is not input within a predetermined time after the request signal is output.
The control device according to claim 2. A communication step of communicating with an external device using the first interface;
A detecting step for detecting that a failure has occurred in the communication;
A counting step for counting faults detected in the detection step;
A count control step of performing control to count up the counting means of the external device using the second interface when the failure is detected;
A fault detection method comprising: A reading step of reading a count value from the counting means connected to the second interface after the count control step;
A determination step of determining the suspected device based on the count value read in the reading step;
The fault detection method according to claim 9, further comprising:

Images