JPH0291756A - Temporary file managing system for loose coupling multiprocessor system - Google Patents

Abstract

PURPOSE:To constitute the title system so that when a system falls into a long-term stop, job which is operated by its system can be operated by the other system by reallocating a temporary file by the other system in a loose coupling multiprocessor system. CONSTITUTION:By information of a temporary file real managing system corresponding memory 100 updated by a temporary file real managing system corresponding memory updating means 40, a reallocation of a common secondary storage device 5 is executed to a system having a system ID which coincides with a real managing system. Accordingly, a temporary file can be reallocated by other system than the system which is stopped. In such a way, when one system falls into a long-term stop in a loose coupling multi-processor system, a job can be restarted quickly by the other system.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention is directed to a loosely coupled multi-system consisting of a plurality of systems connected via a shared secondary storage device. This paper relates to a temporary file management method for a loosely coupled multiprocessor system in which a job is restarted from the start.

(Prior Art) In conventional loosely coupled multiprocessor systems, when one system in a loosely coupled multiprocessor system stops, the job is restarted in its own system and the job is not restarted from another system. There is a system that restarts a job on another system, starting from the beginning of the job without having to reallocate temporary files.

(Problems to be Solved by the Invention) However, in conventional loosely coupled multiprocessor systems, when one of the systems that make up the loosely coupled multiprocessor system stops, jobs are restarted by the own system. When the two systems are stopped for a long time, the entire loosely coupled multiprocessor system is stopped for a long time, resulting in a problem that the reliability of the system is low.

When restarting a job on another system, it is configured to restart from the beginning of the job, and it is not possible to restart from the middle of a job step during job execution or from a specific checkpoint within a job step. The drawback is that the processing time up to step 2 is lower than the restart level in a single system.

The present invention has been made with a focus on the above-mentioned problems, and provides a temporary file for a loosely coupled multiprocessor system, which is a highly reliable system that does not stop the system for a long time and can be restarted from the middle of a job step by another system. The purpose is to provide a management method.

(Means for Solving the Problem) The problem described above is that in a loosely coupled multiprocessor system consisting of a plurality of systems connected via a shared secondary storage device, when a job is restarted, a temporary file is stored on the shared secondary storage device. A temporary file management method for a loosely coupled multiprocessor system in which reassignment of files can be performed from a system other than the stopped system among the systems, wherein the system start-up count of each system is a startup number management memory for storing a startup number in correspondence with a system ID; a system startup number identification means for reading and identifying the startup number of each system from the startup number management memory; A temporary file management memory that stores the system ID and the number of startups at that time when a secondary storage device is allocated, and a temporary file allocation that identifies and allocates a temporary file management destination based on the information in the temporary file management memory. means, a memory corresponding to a temporary file real management system that stores an ID of a real management system that actually manages temporary files in correspondence with a pair of the system ID and startup number; Temporary file real management system memory initial setting means for setting and inputting the system ID of a system as the real management system ID, and when one system goes down, the real management system ID of the temporary file real management system compatible memory is input to the system that is down. A temporary file real management system compatible memory update means updates the system ID of the system to be migrated to a system ID other than the temporary file real management system compatible memory update means, and the temporary file real management system compatible memory update means updates the temporary file real management system compatible memory information based on the information of the temporary file real management system compatible memory update means. This problem can be solved by providing temporary file reallocation means for allocating the shared secondary storage device to a system having a system ID that matches the management system ID.

(Operation) According to the present invention, the information in the temporary file real management system compatible memory updated by the temporary file real managing system compatible memory updating means allows the system having the system ID that matches the real management system to use the shared secondary storage. Devices can be reassigned. This allows temporary files to be reallocated on other systems than the one that stopped them.

As a result, if one system in a loosely coupled multiprocessor system goes into a long-term suspension, the job can be quickly restarted on another system.

Thereby, system reliability in a loosely coupled multiprocessor system can be improved.

(Example) An example of the present invention will be described in detail with reference to the drawings.

As shown in FIG. 1, the present invention corresponds to systems 1 and 2 that constitute a loosely coupled multiprocessor system, job 4 running within system 2, and job 4 running on system 1. a temporary file 6 in the shared secondary storage device 5 used by the job 3 and the job 4, a temporary file actual management system compatible memory 100, and a system startup number identification means 10 that operates in each system of the loosely coupled multiprocessor system. , temporary file allocation means 2
0, temporary file actual management system compatible memory initial setting means 30, temporary file actual management system compatible memory updating means 40, temporary file reallocation means 5G, start-up count management memory 101, and temporary file management memory 102. Note that 8.9 indicates a program.

The details of the processing flow will be described according to FIG.

System 1 and system 2 constitute a loosely coupled multiprocessor system and share a secondary storage device 5.

For each system 1.2, the system startup number is ■ when it is started for the first time, becomes ■ when it is started after the system is stopped, and so on, and is incremented by 1 each time the system is started. The start-up number is set in the start-up number management table 101, which serves as a start-up number management memory of the present invention. Incidentally, in the startup number management table 101 of FIG. 2, the system startup number of system 1 is ■, and the system startup number of system 2 is ■.

In each system, the system startup number of each system 1.2 of the loosely coupled multiprocessor system can be obtained from the startup number management table 101 by the system startup number identifying means 10.

When a temporary file allocation request occurs on system 2,
The temporary file allocation means 20 obtains the system start-up number of its own system from the system start-up number identification means 10. If the allocation is successful thereafter, the temporary file allocation means 20 uses the temporary file management table 102 as the temporary file management memory of the present invention.
The system ID (■) assigned in accordance with the file description information of the temporary file 6 and the system start-up number (■) mentioned above are set in . Note that this temporary file management table 102 is normally placed in a shared secondary storage device.

If the temporary file 6 is already allocated in the system 2, the system will stop. Then, when the system palindrome startup of system 2 is performed and the job is restarted, the temporary file 6 is
reallocation is performed. That is, if it is verified from the temporary file management table 102 that the temporary file was allocated in the own system before the system was stopped, reallocation is permitted. The temporary file area cannot be used from other systems until the re-startup of system 2 is completed and the reallocation is completed, or the job of system 2 is not restarted and the temporary file is subsequently released. Assignments are no longer allowed.

The temporary file actual management system correspondence table 100 as a memory compatible with the temporary file actual management system of the present invention includes a pair of system ID and system startup number, and management of the temporary file 6 allocated at the time of the startup number. This is a table that associates real management system IDs to which rights have been granted. The temporary file actual management system correspondence table 100 is usually placed in a shared secondary storage device.

The temporary file actual management system correspondence table initial setting means 30 as a temporary file actual management system compatible memory initial setting means of the present invention sets the actual management system ID having management rights as the system ID of its own system as an initial state. be. By the way, in this explanation, as shown in Figure 2, the actual management system ID ■ is used for the pair of system ID ■ and startup number, and the real management system ID is used for the combination of system ID ■ and startup number ■. ■Set.

The temporary file actual management system correspondence table initial setting means 30
is called each time each system starts up, and initializes the entry for each system.If system 1 or 2 is stopped for a long time, as described above, the temporary memory allocated by the stopped system is File 6 cannot be allocated in the song system, so a job running on the stopped system cannot be restarted on another system.
In such a case, the temporary file real management system correspondence table updating means 40 as the temporary file real management system corresponding memory updating means of the present invention changes the management right of the temporary file of the stopped system to another system ID.

That is, in the example of FIG. 2, if the system 2 is stopped for a long time, the temporary file real management system correspondence table updating means 40 updates the real management system ID of the entry corresponding to the system ID ■ and the system startup number ■ to ■. do.

By updating this real management system ID, the management authority for temporary file allocation at the stage of system startup number 5 of system 2 is transferred to system 1. The updated temporary file actual management system correspondence table 100 is as shown in FIG.

Next, after changing the actual management system described above, if you want to restart job 4 that was running on system 2 on system 1, the job 4 that was running on system 2 will be restarted on system 2. The temporary file 6 that was previously assigned to job 4 is reallocated.

That is, in the temporary file reallocation means 50,
Referring to the updated temporary file actual management system correspondence table 100 as shown in FIG. 3, it is confirmed that the management right of system startup number 5 of system 2 has been transferred to system 1. As a result, reassignment to the resumed job in system I is permitted only when migration is confirmed. Furthermore, the reallocated resources are updated as having been allocated by system 1. That is, the entry of the corresponding temporary file 6 in the temporary file management table 102 is updated as assigned by the system ID ■ and the system startup number ■, as shown in the temporary file management table 102 shown in FIG.

(Effects of the Invention) As described above, according to the present invention, temporary files can be reallocated to other systems in a loosely coupled multiprocessor system. As a result, in a loosely coupled multiprocessor system, if one system goes down for a long period of time, it is possible to run jobs that were running on that system on another system from the perspective of reallocating temporary files. . This improves system reliability in loosely coupled multiprocessor systems.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the configuration of an embodiment of the present invention, FIG. 2 is a diagram for explaining the constituent elements of the embodiment and the contents of each related daple, and FIG. 3 is a diagram showing the configuration of an embodiment of the present invention. It is a figure which shows the content after an update of a related table. 1...System 1 of a loosely coupled multiprocessor system
．． 2... System of loosely coupled multiprocessor system 2.3... Program running on system 1, 4...
- Program running on system 2, 5... Shared secondary storage device, 6... Temporary file on shared secondary storage device, 10... System startup number identification means, 20...
Temporary file allocation means, 30...Temporary file real management system correspondence table initial setting means, 40...Temporary file real management system correspondence table updating means, 50...Temporary file reallocation means, 100...Temporary File actual management system correspondence table, 101... Launch number management table, 102... Temporary file management table.

Claims (1)

[Claims] In a loosely coupled multiprocessor system consisting of a plurality of systems connected via a shared secondary storage device, reallocation of temporary files onto the shared secondary storage device is stopped among the systems when a job is restarted. A temporary file management method for a loosely coupled multiprocessor system that can be performed from a system other than the system, wherein the number of system startups of each system is stored as a startup number in association with each system ID for each system. system startup number identification means for reading and identifying the startup number of each system from the startup frequency management memory; and a system ID when the shared secondary storage device is allocated by each of the systems.
a temporary file management memory for storing and the number of startups at that time; a temporary file allocation means for identifying and allocating a management destination of the temporary file based on the information in the temporary file management memory; and a temporary file allocation means for actually managing the temporary file. A temporary file real management system compatible memory that stores the real management system ID in association with a pair of the system ID and startup number, and a temporary file real management system compatible memory that stores the system ID of each system in association with the real management system ID. Temporary file actual management system compatible memory initial setting means that inputs settings as ``temporary file actual management system compatible memory initial setting means'', and when one system goes down, the actual management system ID of the temporary file actual management system compatible memory is set to the system ID of the system to be migrated other than the down system. A system ID that matches the real management system ID based on the information of the temporary file real management system compatible memory update means to be updated and the temporary file real management system compatible memory updated by the temporary file real management system compatible memory update means. A temporary file management method for a loosely coupled multiprocessor system, comprising: temporary file reallocation means for allocating the shared secondary storage device to the system.