JPH1153333A - Information processing device that can be reconfigured during operation - Google Patents

Abstract

(57) [Summary] In a multiprocessor information processing apparatus, a processor that is not incorporated in the configuration at the time of initial startup of the information processing system is incorporated in a configuration in which the information processing system is operating. SOLUTION: By adding mounting reservation information to apparatus configuration initial state management information, further providing an online preparation state flag in the apparatus configuration state management information, and performing initial startup processing of a single processor in operation confirmation processing prior to online processing. Incorporate the processor addition procedure into the offline and online procedures of conventional operating systems.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION In order to maintain the reliability that an information processing apparatus can withstand nonstop operation, it is essential to improve the availability as an information processing system. As a method to improve the availability of the information processing system, maintenance and diagnosis of the information processing device are performed while the information processing system is operating, and even if a failure of the information processing device occurs, the affected area of the failure is limited to the failed part, There is a method of replacing a component in a failed part without stopping the processing system to recover the failed part of the information processing apparatus. In addition, in order to secure a scalable performance range in a large-sized information processing device, the configuration of the information processing device can be increased for each processor. However, even in the case where the processor of the information processing apparatus is added, it is necessary to add the processor while the information processing system is operating without stopping in order to improve the availability. The present invention relates to a method for changing or adding a configuration for each processor during operation of an information processing system in an information processing apparatus having a multiprocessor configuration in order to improve the availability.

[0002]

2. Description of the Related Art Conventionally, in a basic information processing system comprising a general multiprocessor information processing device, various input / output devices, and an operating system, a configuration change per processor during the operation of the information processing system, that is, an information processing system Is realized by a disconnection operation and a connection operation for each processor. Hereinafter, this disconnection operation is called an offline operation, and the connection operation is called an online operation. The processor which is separated from the operating information processing system by the offline operation and is brought into the offline state can be maintained / diagnosed and replaced by the processor alone. Then, when maintenance / diagnosis or replacement is completed, the processor is connected to the information processing system again by online operation.

Next, a method of changing the configuration for each processor while the information processing system is operating by the offline operation and the online operation will be described. The information processing apparatus shown in FIG. 2 is a general multiprocessor information processing apparatus including a plurality of instruction processors and an input / output processor. Here, the instruction processors 10, 11, 1N are connected to the system controller 30 by instruction processor connection ports 20, 21, 2N.
Connected to meet. Optional instruction processor 10,
11 and 1N can control the connection state with the system controller 30 by performing the connection operation of the corresponding instruction processor connection ports 20, 21 and 2N, respectively. The input / output processors 40, 41, and 4N are connected to the system controller 30.
And the input / output processor connection ports 50, 51, and 5N. Optional input / output processor 40, 4
1, 4N can control the connection state with the system controller 30 by performing the connection operation of the corresponding input / output processor connection ports 50, 51, 5N. The connection operation between the instruction processor connection ports 20, 21, 2N and the input / output processor connection ports 50, 51, 5N is performed under the control of the service processor 80. The service processor 80 stores the device configuration status initialized by the operator from the service console 90 in the device configuration initial status management file 100. FIG. 3 shows the format of the device configuration initial state management file 100. The device configuration initial state management file 100 contains instruction processors 10, 11, 1N and input / output processors 50, 51, 5N, respectively.
Mounting status flag 31 indicating whether or not it is mounted
0, a permission state flag 320 indicating whether or not to permit the incorporation into the information processing system, and a connection state flag 330 indicating whether the apparatus is connected to the information processing system.
Are provided for each processor. Here, FIG. 7 shows an initial startup procedure of the information processing system. When the information processing system is initially started, a device configuration status management file 110 is created based on the device configuration initial status management file 100 in the first step. FIG. 4 shows the format of the device configuration status management file 110.
Shown in The device configuration status management file 110 includes an online status flag 430 indicating whether or not it is incorporated in the information processing system, in addition to the mounting status flag 410 and the permission status flag 420, similarly to the device configuration initial status management file 100. Prepare for. The mounting status flag 410 and the permission status flag 420 of the device configuration status management file 110 are changed only in the first step at the time of the initial startup of the information processing system, and are not changed during the operation of the information processing system. In the second step, the instruction processors 10, 11, 1N and the input / output
Initial microprogram loading and initialization processing of 0, 41, and 4N are performed. In the third step, the instruction processors 10 and 1 in the mounted state and the permitted state in the device configuration state management file 110 that has been initialized in the second step
1, 1N and the instruction processor connection ports 20, 21, 2N of the input / output processors 40, 41, 4N and the input / output processor connection ports 50, 51, 5N are connected, and the online status flag 4 of the device configuration status management file 110 is set.
Bring 30 online. In the fourth step, the operating system is loaded and started. At this time, the operating system reads the mounting status flag 410 and the online status flag 430 of the device configuration status management file 110 from the service processor 90 by a service command, and reads the mounting status flag 410 from the mounting status flag 410. The maximum configuration of the information processing system is recognized, and the available processor is recognized from the online status flag 430. At this time, the device configuration state management file 110 and the instruction processor connection port 20,
21, 2N, input / output processor connection ports 50, 51,
Table 1 shows the relationship among the 5N, the instruction processors 10, 11, 1N, the input / output processors 40, 41, 4N, and the processor state management table of the operating system. In Table 1, as an example, the instruction processor 1 in the non-permitted state
The relationship between X and the input / output processor 4X, and the relationship between the unmounted instruction processor 1Y and the input / output processor 4Y are also shown. Here, the processor that is not in the mounted state does not exist in the processor state management table of the operating system. After the initial startup of the information processing system, a specific processor is operated offline by an operator during the operation of the information processing system to disconnect the processor from the information processing system. The processor can be incorporated into the information processing system by the online operation by the operation of (1). FIG. 8 shows the procedure of the offline operation. In the first step, in response to the processor offline instruction to the operating system of the operator, the operating system recognizes the processor as an unusable processor and releases it from the information processing system being executed, and sets the processor state management table of the processor to the offline state. After that, a service instruction for instructing the processor to execute an offline operation is issued. Hereinafter, the service instruction instructing execution of the offline operation for the processor is referred to as an offline instruction. In the second step, the service processor executes an off-line instruction for the processor, and after stopping the processor, disconnects the processor connection port and sets the online status flag 430 of the device configuration status management file 110 to the offline status. I do. In the offline state, the operating system cannot use the processor unless an online operation is performed, and the operating system can read out the device configuration state management file 110 by a service instruction to determine that the processor is in the offline state. At this time, the device configuration state management file 110, the instruction processor connection ports 20, 21, 2N, the input / output processor connection ports 50, 51, 5N and the instruction processor 1
Table 2 shows the relationship among 0, 11, 1N, the input / output processors 40, 41, 4N, and the processor state management table of the operating system. Table 2 shows, as an example, a relationship when the instruction processor 11 and the input / output processor 41 are set to an offline state by an offline operation. When the processor is in the offline state, maintenance, diagnosis, and replacement of the processor can be performed independently of the information processing system that is operating the information processing. Next, the procedure of the online operation is shown in FIG. In the first step, a service instruction for instructing the processor to perform an online operation is issued by the processor online instruction to the operating system of the operator. Hereinafter, the service instruction instructing the processor to execute the online operation is referred to as an online instruction. In the second step, the service processor executes an online instruction for the processor, performs an initial microprogram load of the processor and performs an initialization process, and then connects the processor connection port to a connected state to store the online state of the device configuration state management file 110. The flag 430 is brought online. In the third step, the operating system recognizes the processor as an available processor upon completion of the online instruction for the processor in the second step, sets the processor state management table of the processor to the online state, and operates the information processing system. And the information processing in the processor is started. At this time, the device configuration state management file 110 and the instruction processor connection ports 20 and 21,
The relationship between the 2N, the input / output processor connection ports 50, 51, 5N, the instruction processors 10, 11, 1N, the input / output processors 40, 41, 4N and the processor status management table of the operating system is as shown in Table 1 before execution of the offline operation. Return. Thereafter, the operating system issues a device configuration status management file 1 by a service command.
By reading 10, it can be seen that the processor is online.

As described above, when the information processing system is initially started, the configuration of the information processing apparatus is determined according to the initial setting of the operator, and during the information processing operation after the initial startup of the information processing apparatus, the information processing apparatus is brought into the online state at the time of the initial startup. The processor can be placed in an off-line state by an off-line operation to perform maintenance, diagnosis, and replacement independently of the operating information processing system, and can be incorporated into the on-line information processing system again by the on-line operation.

[0005] Even in an information processing system using a computer resource logical partitioning mechanism, the on-line operation and the off-line operation of the physical processor are performed by the above-described method, whereby the physical processor is operated during the operation of the information processing system. Maintenance, diagnosis and replacement can be realized.

[0006]

According to the prior art,
For processors that are online when the information processing system is initially started, perform offline operation for each processor, perform maintenance, diagnosis, and replacement independently of the information processing system that is operating the information processing system. Could be incorporated into the information processing system inside.

However, when an information processing system stop failure occurs due to an unrecoverable failure of a single processor, the failed processor is set as a non-permission state at the time of setting the initial configuration of the information processing system as an emergency measure, and the information processing system is re-executed. It is conceivable to perform recovery from a stop failure of the information processing system by performing the initial startup. In this case, the performance of the entire information processing system is reduced because the faulty processor is not incorporated in the information processing system, and the processing capability required for performing the normal information processing operation is reduced, and some information processing operations cannot be performed. Can be considered. Therefore, it is necessary to promptly replace the failed processor with a normal product and restore the information processing system to a state where all information processing operations can be performed. However, in the conventional information processing system, the processor which has been set to the non-permission state at the time of setting the initial configuration of the information processing system goes offline at the initial startup of the information processing system, and is in the offline state from this initial startup during the information processing operation thereafter. The processor could not be brought online and incorporated into the information processing system. For this reason, even if the failed processor is replaced with a normal one while the information processing system is operating, the information processing system is stopped again and the processor is permitted again in the initial configuration setting in order to incorporate the processor into the information processing system again. Then, the information processing system had to be started again.

Further, when the processing performance of the information processing system becomes insufficient due to the expansion of the information processing business, a scalable performance upgrade method by adding a processor is prepared in a large-scale information processing system. An information processing system adapted to the scale of the processing operation can be constructed. However, when the information processing system is initially started up, the operating system sets the maximum configuration of the processor based on the mounting state. Necessary initial setting is not performed, and there is no means for newly installing this processor in the information processing system even if the processor is added and newly mounted during the operation of the information processing system. Therefore, in order to add a processor for upgrading the processing performance of the information processing system, it is necessary to stop the information processing system, set the processor to the mounting state and the enabled state in the initial configuration setting, and then restart the information processing system again. Need to do fit.

[0009] As described above, stopping the system for recovering the information processing system from the stop failure or for upgrading the performance is significantly contrary to the improvement of the availability of the information processing system. For this reason, in the present invention, means for re-installing a processor that is not in the permitted state into the information processing system without stopping the information processing system in the operating information processing system, or newly incorporating a processor that is not in the mounted state into the information processing system. The purpose is to realize the means. Further, when the processing method related to the management of the information processing apparatus by the operating system is changed to realize this object, the implementation of the means of the present invention is newly applied to the information processing system not implementing the means of the present invention. It is necessary to stop the information processing system and replace the operating system. Therefore, an object of the present invention is to realize the above means without changing the operating system.

Although the initial microprogram load processing time of a processor varies greatly depending on the type and scale of the processor, it generally takes several seconds to several tens of seconds in a large general-purpose computer. In this case, in the conventional on-line instruction procedure, the time for performing the initial microprogram load processing of the processor becomes dominant with respect to the entire processing time. The instruction execution time of several seconds to several tens of seconds is orders of magnitude longer than the execution time of general instructions, which makes it difficult to design a time monitoring abnormality detection mechanism of the information processing system. Therefore, an object of the present invention is to realize means for reviewing the online processing procedure and drastically improving the execution time of the online instruction.

[0011]

According to the present invention, in an information processing system using a multiprocessor information processing apparatus, when some processors are in a non-permitted state in an initial configuration setting at the time of initial startup, an offline A processor that has not been incorporated into an information processing system that has been left in a state is set to the enabled state while the information processing system is operating, then the initial microprogram loading and initialization processing are performed in the single processor initial startup processing, and then online processing is performed And an online preparation state flag indicating that the processor is performing an initial startup process in the single processor initial startup process, and a single processor of the processor. Until the initial startup process is completed, this online Means for controlling whether to execute online processing by setting the ready state flag is provided.

Further, in the information processing system using the multiprocessor information processing apparatus, the operating system can recognize the addition of the processor which is not mounted at the time of the initial startup of the information processing system. In order to make the processor in a mounting reserved state and offline state, a mounting reservation state flag is provided, and by setting this mounting reservation state flag at the time of initial startup of the information processing system, the processor becomes offline state, and thereafter, While the processing system is running, the processor is added, the mounting reservation state is released, the single processor is initially activated, and online processing can be executed, and then the online processing of the processor is executed from the operating system to execute the information processing system. To provide a procedure that can be incorporated.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS When some processors are in a non-permission state in an initial configuration setting of an information processing apparatus at the time of initial startup of an information processing system, a non-permission of the information processing apparatus in which the new means according to the present invention is implemented. A method of incorporating the processor in the state into the operating information processing system will be described. FIG. 1 shows a configuration of a multiprocessor information processing system according to the present invention. The processor which was in the mounted state and the non-permitted state in the initial configuration setting of the information processing apparatus reads out the mounting information from the service processor 90 when the operating system issues a service instruction at the initial startup of the information processing system. Since it is in the mounted state, it is incorporated in the initial configuration information of the information processing system, and in the non-permitted state, the processor is regarded as offline (Table 1). The processor in the non-permitted state in the initial configuration setting of the information processing apparatus has not performed the initial microprogram loading and initialization processing of the processor required for starting the information processing system in the initial startup of the information processing apparatus. For this reason, even if an online operation is performed on the processor that is offline during the operation of the information processing system,
Since the service processor is in the non-permission state, the service processor informs the operating system that the apparatus is in an unusable state as a result of executing the online instruction. As a result, the processor is not incorporated in the information processing system. Therefore, in order to incorporate the processor in the off-line state into the information processing system, in order to make the processor usable, it is necessary to carry out a procedure for initial microprogram loading and initialization processing of the processor. FIG. 10 shows a procedure for performing the initial startup of the processor, and the procedure will be described in the following order. In the first step, an instruction for initial activation of the processor is input from the console of the service processor by the operator. In the second step, the service processor sets the processor state in the device configuration state management file 110 to the permitted state in order to perform the initial startup of the processor. Conventionally, the permission status flag of the device configuration status management file 110 has not been changed during the operation of the information processing system. Make new changes. Furthermore, even if the processor is enabled at this stage,
Even if an online command is issued from the operating system without waiting for the completion of the next third step, the operating system must be informed that the device is unusable as an execution result. Provide a flag. In the embodiment shown in FIG. 1, as the online preparation state flag, new online preparation state flag registers 120, 121, and 12N are prepared in the system controller for the instruction processor, and the service processor is prepared for the input / output processor. A new online preparation status flag 540 is prepared in the device configuration status management file 110 of FIG. FIG. 5 shows a new format of the device configuration status management file 110 according to the embodiment of the present invention. Then, when the processor is in the permission state at this stage, the online preparation state flag register 120 or the online preparation state flag 540 is set so that it can be recognized that the single processor initial startup processing of the processor is being performed. In the third step, the service processor performs the initial microprogram loading and initialization processing of the processor, and the initial startup processing of the processor is completed. Here, since the single processor initial startup process of the processor is completed, the setting of the online preparation state flag (online preparation state flag register 120, online preparation state flag 540) is released. As described above, the processor becomes online-enabled in the offline state, and if an online operation is performed on the processor, an online command is issued from the operating system, and the service processor 90 can bring the processor into the online state. This online procedure is shown in FIG. Compared to the on-line procedure according to the prior art (FIG. 9), these steps can be omitted because the initial microprogram loading and initialization processing of the relevant processor have been completed, and the initial microprogram loading need not be performed. Speed up online processing. The procedure for performing the single processor initial startup process of the processor is required when the single processor initial startup process of the processor has not been performed yet, that is, when the information processing system is initially booted. Either the processor is in the non-permission state in the initial configuration setting of the apparatus, or the processor is placed in the offline state during information processing operation, and maintenance, diagnosis, and replacement are performed. In this case, it is necessary to perform a diagnosis before incorporating the failure into the information processing system before incorporating the failure into the information processing system. There is a need to do. Therefore, the initial microprogram loading process and the initialization process of the processor are completed by performing the above-described single processor initial startup process when the maintenance worker performs the single diagnosis of the processor by the service processor during the maintenance work. Then, when the processor is brought online with an online instruction, since the initial microprogram loading process and the initialization process have already been completed,
The single processor initial startup processing of the processor during the online processing procedure is unnecessary. If a specific processor is offline-operated from the operating system during normal information processing system operation, and then online operation is performed, the initial microprogram loading process and initialization process of the processor are already performed at the initial startup of the information processing system. At this time, the single processor initial startup processing of the processor during the online processing procedure is unnecessary.

Next, a description will be given of a method of adding a processor, which was not mounted at the time of the initial startup of the information processing system, to the information processing apparatus in which the new means according to the present invention is implemented, and incorporating the processor into the information processing system while the information processing system is operating. I do. The processor which was not mounted in the initial configuration setting of the information processing apparatus reads out the mounting state information from the service processor 90 when the operating system issues a service instruction at the time of initial startup of the information processing system.
It is not included in the initial configuration information of the information processing system because it is in the mounting state (Table 1). Therefore, in order to incorporate the processor into the information processing system even if a processor which was not mounted in the initial configuration setting of the information processing apparatus is conventionally added, it is necessary to set the initial configuration of the information processing apparatus after stopping the operating information processing system. In this case, the processor must be mounted and allowed in the device configuration initial state management file 100, and the information processing system must be started again. For this reason, in the present invention, as shown in FIG. 6, a mounting reservation status flag 600 is added to the device configuration initial state management file 100, and even if the processor is not mounted when the information processing system is initially started, the device configuration initial state management is performed. If the file 100 is in the mounting reservation state, the operating system issues a service instruction by setting the mounting state flag of the device configuration state management file 110 to the mounting state, and the service processor 9
When the mounting information is read from 0, it is incorporated in the initial configuration information of the information processing system because of the mounting state. At this time, the device configuration state management file 110, the instruction processor connection ports 20, 21, 2N, the input / output processor connection ports 50, 51, 5N and the instruction processors 10, 11, 1N, the input / output processors 40, 41, 4
Table 3 shows the relationship between N and the processor state management table of the operating system. During the operation of the information processing system, the service processor 90 remains online even if an online instruction is issued to the processor even though an online instruction is issued to the processor even though the processor is not mounted in the mounting reservation state even though the processor is not mounted. Notifying the operating system that the processor is in a failure state without completing the instruction normally, so that the processor is not incorporated in the information processing system. When adding the relevant processor while the information processing system is operating, mount the relevant processor during the offline state,
After the mounting is completed, the mounting reservation state of the device configuration initial state management file 100 is released, and the service processor 90 is set to a state where the online instruction can be executed. In this state, the processor is in the mounted state and the non-permitted state in the device configuration state management file 110. If the processor in the non-permitted state is installed in a running information processing system by performing online processing on the processor, Can be incorporated into the information processing system while the information processing system is operating.

[0015]

According to the present invention, a processor which has not been incorporated into an information processing system in a non-permitted state at the initial startup of the information processing system and which has not been incorporated in the information processing system is brought online while the information processing system is operating. It becomes possible to incorporate.

Further, it is possible to add a processor which has not been mounted at the time of initial startup of the information processing system and add it to the information processing system while the information processing system is operating.

In any of the above cases, the information processing apparatus according to the present invention is simply used for the existing information processing system without adding any new processing to the operating system. It can be incorporated when the processing system is running.

[Table 1]

[Table 2]

[Table 3]

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a multiprocessor information processing apparatus according to the present invention.

FIG. 2 is a configuration diagram of a conventional multiprocessor information processing device.

FIG. 3 is a diagram showing a conventional apparatus configuration initial state management file.

FIG. 4 is a diagram showing a conventional device configuration state management file.

FIG. 5 is a diagram showing an apparatus configuration status management file according to the present invention.

FIG. 6 is a diagram showing an apparatus configuration initial state management file according to the present invention.

FIG. 7 is a diagram showing an initial startup processing procedure of the multiprocessor information processing system.

FIG. 8 is a diagram illustrating an off-line processing procedure of a processor in a multiprocessor information processing system.

FIG. 9 is a diagram showing an online processing procedure of a conventional processor.

FIG. 10 is a diagram showing a single processor initial startup procedure according to the present invention.

FIG. 11 is a diagram showing an online processing procedure of a processor according to the present invention.

[Explanation of symbols]

10,11,1n… Instruction processor 20,21,2n… Instruction processor connection port 30… System controller 40,41,4n… I / O processor 50,51,5n… I / O processor connection port 60… Main storage device 70… Main storage device connection port 80: service processor 90: service console 100: device configuration initial state management file 110: device configuration initial management file 120: online preparation status flag register 310: mounting status flag 320 in the conventional device configuration initial status management file … Permission status flag in the conventional device configuration initial state management file 410… mounting status flag in the conventional device configuration initial status management file 420… permission status flag in the conventional device configuration status management file 430… in the conventional device configuration status management file Online status flag 510… Conventional device configuration status management flag Mounting status flag 520 in the device configuration status management file according to the present invention permission status flag 530 in the device configuration status management file according to the present invention online status flag 540 in the device configuration status management file according to the present invention 600 Implementation mounting status flag 610 in the device configuration initial state management file according to the present invention. Mounting status flag 620 in the device configuration initial state management file according to the present invention.

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Kazuo Hibi, 1st Horiyamashita, Hadano-shi, Kanagawa Prefecture Nichi Information Technology Co., Ltd. General-purpose computer division

Claims (3)

[Claims] In an information processing system using an information processing apparatus including a plurality of processors, information processing is performed on a processor which is not incorporated in the configuration at the time of initial startup of the information processing system and which is in an offline state without performing an initial startup process. An information processing apparatus comprising: means for independently performing an initial startup process during operation of a processing system; and incorporating the processor into the information processing system by online operation thereafter. 2. The information processing apparatus according to claim 1, further comprising means for storing information indicating that the single unit initial startup processing of the processor has not been completed, and means for notifying the operating system of the information. apparatus. 3. The information processing apparatus according to claim 2, further comprising: means for storing mounting reservation information for an unmounted processor; and means for notifying the operating system of the information.