JPS62285171A - Control system for switching system constitution - Google Patents

Abstract

PURPOSE:To improve the availability in switching control of a system having plural information processors by performing early the switch control of the system constitution in case the retry processing time of a processor may possibly exceed the set value when the processor has a fault. CONSTITUTION:A processor 1A is provided together with a spare processor 1B. When the processing carried out by the processor 1A has abnormality, the retry processing is performed within the processor 1A. If this retry processing succeeds within a prescribed time, an abnormality report signal 8A is not transmitted. However the signal 8A is sent to a constitution switch controller 2 from a retry processing control logic 9A of the processor 1A in case the retry processing is impossible or delayed. Thus the controller 2 decides the abnormality of the processor 1A and informs this abnormality to the processor 1B. Then the processor 1B gives a command to the controller 2 to switch the changeover switches 31-3n to the processor 1B. Thus the controller 2 switches the connection to the processor 1B.

Description

Detailed Description of the Invention 3. Detailed Description of the Invention [Field of Industrial Application] The present invention relates to a connection switching control method for component devices in an information processing system having a plurality of processing devices, that is, system configuration switching control. 0 [Prior Art] Conventionally, in configuration switching control in an information processing system having multiple processing devices, various methods have been used to shorten down time when changing the configuration and improve system operation rate. There is a system configuration switching control method.

The conventional system is a configuration switching control device as described in Japanese Patent Publication No. 19540/1983. There is something installed in the system. The configuration switching control device grasps the status of the processing equipment in operation, and when it detects an abnormality in one processing equipment, it notifies the abnormality to other processing equipment in a normal state, and switches the processing equipment issued from this processing equipment. The configuration of the system is changed by operating each switch on the configuration switching device according to the command. Here, in order for the configuration switching control device to grasp the status of the processing device in operation, the configuration switching control device issues an inquiry signal to the processing device at regular intervals, and receives a response signal from the processing device over time. This is done by monitoring. That is, if an abnormality occurs in the processing device and the processing device is unable to return a response to the arrangement from the configuration switching control device within a predetermined time, the processing device is considered to be abnormal.

In order to improve the operating rate of the system, it is necessary to detect abnormalities 2 earlier, and therefore there is a demand for shortening the time interval between inquiries. Now, recent processing devices perform retry processing as one way to improve reliability. Retry is
When a temporary failure occurs in the processing device, the processing device will try to execute it again. If the retry processing time becomes longer, a problem arises in that a response to the above-mentioned arrangement cannot be returned. The retry processing time varies depending on the type of failure, and increases proportionately as the number of retries increases. Therefore, when monitoring the status of the processing device by the configuration switching control device, it is necessary to take into consideration the processing time of retry performed by the processing device.

[Problem that the invention seeks to solve]

The above conventional technology has a problem in that no consideration is given to retry processing when a failure occurs in the processing device.

If the IJ) rye processing is successful, the processing device in which the failure has occurred is considered to be normal, and if the IJ) rye processing fails, the processing device is considered to be abnormal. However, it should be noted here that if the retry process takes too long, the monitoring time in the configuration switching control device will be exceeded, and a response to the arrangement from the configuration switching control device may not be returned. be. To solve this problem, the processing device
J) A possible method is to issue a signal to the configuration switching control device to suppress abnormality detection while the lie process is being executed.

However, in this solution, if the retry process ends in failure, system configuration switching control will start from the point at which the retry process is unsuccessful, compared to the case where no retry process is performed. This results in a drawback that the switching time actually increases.

The purpose of the present invention is to promptly declare to the configuration switching control device that the self-processing device is abnormal when there is a possibility that the processing time exceeds a preset value; The objective is to hasten the start of system configuration switching control.

[Means for solving problems]

The above object is achieved by estimating the retry processing time to be performed in the processing device when a failure occurs in the processing device.

[Effect]

In order to solve the problems in the prior art, the operation of IJ try processing performed by the processing device will be described below. Classifying retry processing in terms of IJ) retry processing time is 3.
classified into types. The first case is IJ). In this case, the processing device in which the failure has occurred does not perform IJ) processing, but sends an abnormality signal to the configuration switching control device. Second, although it is impossible to know whether the retry process will succeed or fail until the retry process is actually performed, the time it takes to find out the result of the first (first) IJ) retry process is A predetermined time? This is a case where there is a possibility of exceeding the limit. In this case, an abnormality signal is sent to the configuration control device, immediately giving up on retry processing in the processing device in which the failure has occurred. Thirdly, as in the second case, it cannot be determined until the IJ) lie processing is actually performed, but since the 'J) lie processing time per 1ti21 is short, the predetermined within the time,
This is a case where the IJ) lie process is performed multiple times until it is successful. In this case, if T is successful, no abnormality signal is sent to the configuration control device and the processing in the processing device continues. Also, if the IJ) lie processing within a predetermined period of time is unsuccessful, an abnormality signal is sent to the configuration controller.

Embodiment] Hereinafter, an embodiment of the present invention will be described with reference to FIGS. 1 and 2. In FIG. 1, 1A and 11B are processing devices, respectively. 4fi4. are input/output devices, respectively, and 031 to 33 connected to each processing device IA and IB by the input/output interface 5 are input/output interface changeover switches of each of the input/output devices 41 to 4rL.

2 is a configuration switching control device that performs switching control of the system configuration, and is connected to the changeover switches 31 to 3n via changeover switch control signals 1!6 to changeover control interfaces 7A, 7.
Each processing device through B! IA, IB are connected to each processing device IA, IB via an abnormality report signal IR8BI. A retry processing control logic 9A that controls retry processing performed by the processing device when a failure occurs.

9B exist in the processing bag Wt1 person 1B, and are respectively connected to the above-mentioned abnormality report signal 118 person 8B.

Now, one processor is in use and the other processor is 1B.
is used as a spare machine. When each processing device 1B enters the operating state, it sends a start command to W or the switching control device 2 via the switching control interfaces 7A and 7B. By receiving this start command, the configuration switching control device 2 recognizes that the processing bags WIA, IB have entered the operating state, and from then on, each processing device IA.

The status of 1B (operating, stopped) is determined by the other processing device IA.

1B at fixed time intervals. Based on this status report, the preliminary processing device 1B recognizes the status of the current processing device tIA and the current processing bag [IA is the status of the preliminary processing device 1B], and at the same time recognizes the status of the current processing device tIA and the status of each processing device IA.

1B sends a response within a predetermined time after receiving the status report, and notifies the configuration switching control device 2 via the switching control interfaces 7A and 7B that the processing devices 1A and 1B are operating normally. In other words, the configuration switching control device 112
monitors the state of the processing device 1lIA by monitoring the response time from one processing device, and also monitors the state of the preprocessing device 1B in the same manner.

After recognizing that one processing device has entered the operating state, the configuration switching control device i12 switches the changeover switch 3 according to the switching command of the changeover switches 31 to 3rL from the one processing device.
．． ~3rL is operated to connect each input/output device 4l-4rL to the processing bag @IA, and the system is operated with one processing device as the active machine.

Processing Bag [If an abnormality occurs in the process being executed by IA,
Retry processing is performed within one processing device.

IJ) If the lie processing is successful within the predetermined time, the abnormality report signal 8A is not sent. Retry process 3: If the retry process is not possible without the need to perform it, if there is no possibility that the retry process will finish within the specified time, or if the retry process is repeated several times but does not succeed within the specified time, an error occurs. An abnormal signal is detected at the reporting signal 8A, and the processing bag [I in IA]
J) Sent from the lie processing control logic 9A to the configuration switching control device 2. The configuration switching control device 2 that has received the abnormality signal regards the one processing device as abnormal, stops monitoring of this one processing device, and then... ? Processing device (processing device in ffIA state) processes
Informs about tIA abnormality. When the processing device 1B recognizes an abnormality in one of the processing devices, it issues a switching command to the configuration switching control device 2 to switch the changeover switches 31 to 5n to the processing bag @IB side. When this switching command is issued, the switching control device Wt2 operates the changeover switches 31 to 3rL based on the signal on the switching command signal 6 to switch the connection to the processing device 1B side. When this switching is completed, the configuration switching control device 2 reports this completion to the processing device 1B. Thereafter processing device 1
B will continue to operate the system as the active machine. Note that if an abnormality signal cannot be sent to the configuration control device 2 using the abnormality report signal i8A due to an abnormality in one processing device, a response to the status report on the switching control interface 7A will be output within a predetermined time. It will no longer be done. When there is no response from one processing device, the configuration switching control device 2 considers that one processing device to be abnormal, and performs subsequent operations as if the abnormality report signal 8A had been sent to the configuration switching control device. conduct.

FIG. 2 shows a flowchart of the IJ) processing carried out by the processing device. In this figure, the predicted time for one IJ) lie process is the time T0 from the occurrence of a T''N abnormality until the abnormality is detected, and the retry processing time allowed on the processing device side is TmαX. ( Here, Tmaz <
Ta holds true. ) In FIG. 2, the number shown on the right shoulder of each step indicates the step number.

If an error occurs in step 20, it is determined in step 21 whether the error is a retryable error or not, and if it is an error that cannot be retried, the abnormality is notified to the configuration switching control device in step 28. Thereafter, the configuration switching control device performs configuration switching, and processing is performed by the backup processing device. If it is determined in step 20 that retry is possible, the parameter N is set to the value )' in step 22, and it is checked in step 23 whether Tr>Tmax.

If the predicted time Tr for one retry process exceeds Tmaz, the process branches to step 2. If the test is unsuccessful in step 23, a retry process is performed in step 24. As a result of the retry process, if it is determined in step 25 that the retry process is successful, the process branches to step 26 and the process in the current machine is continued. If it is determined in step 25 that the retry process is unsuccessful, the process branches to step 27 and the value of the variable N is incremented by 1, and the predicted retry process time is checked again in step 23. Thereafter, the processing continues until the loop of steps 23, 24, 25 and 27 is exited.

As described above, according to the embodiment of the present invention, even if the processing device detects an abnormality in the equipment and performs a retry process, the time required for the retry process does not extend beyond a predetermined value. You can save time and improve system availability.

The above embodiment is a system with two processing devices,
It goes without saying that the present invention can be similarly implemented in a system having three or more processing devices.0 [Effects of the Invention] According to the present invention, when a failure occurs in a processing device constituting the system, the processing device's ' J) Since the increase in down time due to the extension of the lie processing time can be suppressed by judgment in the processing device, there is an effect of improving the system operation rate.

[Brief explanation of the drawing]

Fig. 1 is a system configuration diagram showing one embodiment of the present invention;
The figure is a flowchart showing the IJ) processing in the processing device. ...I/O devices aA, aB...Abnormality report signal line 9 people 9B...Retry processing control logic 1st side

Claims (1)

[Claims] 1. In an information processing system consisting of a plurality of processing devices and a plurality of input/output devices having input/output interface changeover switches between each of the processing devices, the configuration of the system is configured by operating each of the changeover switches. A configuration switching control device is provided to switch the configuration, and when an abnormality is detected in one processing device and retry processing is possible in that processing device, the system switches the configuration within a preset allowable time for retry processing. The processing equipment repeats retries, and if the retry is successful within the permissible time, the processing equipment continues to operate, but if it is predicted that the permissible time will be exceeded, or if retry processing is not possible with the processing equipment. In such a case, a means for stopping the retry processing is provided, and an abnormality signal is sent to the above-mentioned configuration switching control device, and the configuration switching control device operates the above-mentioned changeover switch according to a predetermined procedure to restart the system. A configuration switching control method characterized by changing the configuration.