JP2009098715A - Redundant system device, job execution method in redundant system device, and execution program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform switching of a system when a failure occurs in a redundant system in a short period of time. <P>SOLUTION: A main system server 10 and a standby system server 20 are provided with execution period DB11 and 21 with periods to execute jobs stored therein and execution orders DB 13 and 23 with the execution orders of jobs which have dependence with each other stored therein. The standby server 20 is provided with an unexecuted job DB22 with jobs which have not been executed by the main system server 10 stored therein. In a normal time, the main system server 10 executes the jobs by referring to the execution period DB11 and the execution order DB13, and notifies the standby system server 20 of executed jobs, and the standby system server 20 updates the unexecuted job DB22 on the basis of the notification. In an abnormal time, the standby system server 20 refers to the execution period DB21, the execution order DB23 and the unexecuted job DB22, and executes the jobs which have not been executed by the main system server 10, that is, the jobs having dependence with jobs have been already executed. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a redundant system device and a job execution method and execution program in the redundant system device, and more particularly to a redundant system device and a job in the redundant system device that improve reliability by multiplexing with a primary server and a standby server. The present invention relates to an execution method and an execution program.

  Today, various information systems are the basis of business. Therefore, even when the system goes down due to a factor such as a disaster, it is required to restore the system in a short time.

  In view of this, a technique is known in which a system is made redundant to improve resistance to system down. For example, in a system having a primary server (production system) and a standby server (standby system), if the primary server fails, the job execution subject is switched from the primary server to the standby server. Thus, the job can be continuously executed.

  As a typical system including a main server and a standby server, a hot standby system (hot spare system) and a cold standby system (cold spare system) are known.

  The standby server in the hot standby system always performs the same operation as that of the main server operating normally (that is, in the mirroring state). Therefore, when an abnormality occurs in the primary server, the standby server can immediately take over the processing.

  On the other hand, the standby server in the cold standby system is started after confirming the occurrence of a failure in the main server, and the job execution is taken over. In addition, a method (warm standby method) that realizes faster process switching than the cold standby method by starting up only the OS in the standby server is also known. The cold standby method is less expensive than the hot standby method, but is less resistant to system down.

  In Patent Document 1, as an example of the hot standby method, two computers simultaneously perform online business processing, and increase the reliability of the entire system by combining the processing results of each other (parallel redundancy method, dual operation method) Is disclosed.

  Also, Patent Document 2 discloses a hot standby system that enables quick continuous operation even if a failure occurs in the active server between the active server and the standby server that are automatically operated.

  Note that Patent Document 3 discloses a method that enables re-execution from a processed job step without correcting a description in a job control language when a failure occurs in an offline system.

JP-A-60-043772 JP 11-259326 A Japanese Patent Laid-Open No. 06-103078

  The following analysis was made by the present inventors.

  In the conventional cold standby method, it is difficult to switch between the primary server and the standby server in a short time. This is because it takes time to check the execution status of the job before the system down among the scheduled jobs and to set the location where the job is continuously executed after switching the execution system.

  Also, in the conventional cold standby system, when synchronizing the job execution status, it is necessary to exchange data between the primary server and the standby server, and both systems are installed at remote locations. If so, it took time and effort to exchange data.

  On the other hand, in the hot standby system disclosed in Patent Documents 1 and 2, the standby server executes a job in the same manner as the main server even when it is normal, so that there is a problem that the cost becomes high.

  Therefore, it is an object to provide a redundant system apparatus that can perform a system switching operation in a short time when a failure occurs in a redundant system apparatus that does not execute a job in a standby server during normal operation.

  The redundant system apparatus according to the first aspect of the present invention includes an execution time database (hereinafter, referred to as DB) storing job execution times and an execution order storing execution orders of a plurality of jobs that are mutually dependent. A redundant system device comprising: a main server having a DB; and a standby server having a non-executed job DB storing jobs that have not been executed in the execution time DB, the execution order DB, and the main server. In the normal state, the primary server executes the job with reference to the execution time DB and the execution order DB, notifies the standby server of the executed job, and the standby server The non-executed job DB is updated based on the notified executed job, and when an abnormality occurs, the standby server is connected to the execution time DB. The execution order DB and the unexecuted job DB are referred to, and a job that has not been executed on the primary server and that has been executed for a job having a dependency relationship with the job is executed. It is characterized by that.

  A job execution method in a redundant system apparatus according to a second aspect of the present invention includes an execution timing database (hereinafter referred to as DB) storing job execution timing and execution of a plurality of jobs that are dependent on each other. A primary server having an execution order DB storing an order; and a standby server having an execution time DB, an execution order DB, and an unexecuted job DB storing jobs not executed on the primary server. A job execution method in a redundant system device comprising: a step of executing a job by referring to the execution time DB and the execution order DB when the main server is in a normal state; A step of notifying the standby server, and a step of the standby server updating the unexecuted job DB based on the notified executed job The standby server refers to the execution time DB, the execution order DB, and the unexecuted job DB, and is not executed on the primary server, and the job It includes a step of executing a job having a dependency relationship that has been executed.

  An execution program for a job in a redundant system apparatus according to a third aspect of the present invention includes an execution time database (hereinafter referred to as DB) that stores the time at which a job is executed and the execution of a plurality of jobs that are mutually dependent. A primary server having an execution order DB storing an order; and a standby server having an execution time DB, an execution order DB, and an unexecuted job DB storing jobs not executed on the primary server. A job execution program in a redundant system apparatus comprising: a process for causing the main server to execute a job with reference to the execution time DB and the execution order DB in a normal state; and A process for notifying the standby server of the executed job, and based on the executed job notified to the standby server. The process of updating the unexecuted job DB is executed by a computer, and in the event of an abnormality, the standby server is referred to the execution time DB, the execution order DB, and the unexecuted job DB, and the main server The computer is caused to execute a process for executing a job that has not been executed in step 1 and a job having a dependency relationship with the job has been executed.

  In the redundant system device according to the first development mode, the standby server refers to the execution time DB, the execution order DB, and the non-executed job DB and is not executed in the primary server when an abnormality occurs. If a job that has a dependency relationship with the job has not been executed, the job is preferably executed after the job having the dependency relationship is executed.

  The job execution method in the redundant system apparatus according to the second development mode is such that the standby server refers to the execution time DB, the execution order DB, and the unexecuted job DB when an abnormality occurs, and the main system When a job that has not been executed in the server and a job having a dependency relationship with the job has not been executed, it is preferable to include a step of executing the job after executing the job having a dependency relationship.

  With the redundant system apparatus according to the present invention, it is possible to switch from the primary server to the standby server in a short time. This is because the standby server makes the schedule determination in the same way as the primary server, and it is easy to set the point for continued execution of the job because the completed job can be determined from the unexecuted job DB and job end information. is there.

  Further, the redundant system apparatus according to the present invention can automatically determine the job dependency in the standby server and execute only those that can be submitted. This is because the operation status of a preceding job having a dependency relationship is accumulated, and a job that can be submitted can be determined based on the operation status.

  Further, the redundant system apparatus according to the present invention can synchronize the job execution status without imposing a load on the primary server. This is because a message is transmitted only at the end of the job, and synchronization can be achieved with a small amount of data.

  A redundant system apparatus according to an embodiment of the present invention will be described with reference to the drawings.

  Referring to FIG. 1, the redundant system device includes a main server 10 and a standby server 20.

  The main server 10 includes an execution time DB 11 and an execution order DB 13, and further includes a schedule determination unit 15, a server determination unit 16, a dependency relationship determination unit 17, and a job execution unit 18.

  The standby server 20 includes an execution time DB 21, an unexecuted job DB 22, an execution order DB 23, a schedule determination unit 25, a server determination unit 26, a dependency relationship determination unit 27, a job execution unit 28, and a job execution status synchronization unit 29.

  Under normal conditions, the main server 10 operates as follows. That is, for a job submitted based on the execution time DB 11, the dependency is checked by the execution order DB 13, and then the job is executed.

  Only the end message of each job is transmitted as information indicating the job execution status from the job execution unit 18 of the primary server 10 to the job execution status synchronization unit 29 of the standby server 20.

  The standby server 20 operates as follows. That is, the job submission is requested based on the same execution time DB 21 as the execution time DB 11 of the main server 10. However, in the standby server 20, when normal, actual job submission is not performed, and jobs that have not been submitted to the unexecuted job DB 22 are recorded. The job execution status synchronization unit 29 of the standby server 20 periodically acquires the job completion status from the job execution unit 18 of the primary server 10 and performs synchronization processing related to the job execution status.

  When a switch occurs between the primary server 10 and the standby server 20 at the time of an abnormality, the situation from the point where the previous synchronization processing was performed to the present is synchronized, and the standby server 20 Input is made. Here, it is determined by the unexecuted job DB 22 and the execution order DB 23 whether there is any unexecuted preceding job having a dependency relationship with the requested job. If there is an unexecuted predecessor job that has a dependency relationship with the requested job, the requested job is not executed. Therefore, the standby server 20 executes only jobs that do not depend on the preceding job.

  Next, a first embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 2 is a block diagram of the redundant system apparatus according to the first embodiment of the present invention.

  Referring to FIG. 2, the redundant system device according to this embodiment includes a main server 10, a standby server 20, and an execution status storage server 40.

  The primary server 10 stores an execution time DB 11 that describes the date and time when the job should be executed, an unexecuted job DB 12 that records whether or not the job is actually executed based on the execution time DB, and job dependencies necessary for flow control. The execution order DB 13, job definition information 14 storing information necessary for job execution, a schedule determination unit 15, a server determination unit 16, a dependency relationship determination unit 17, and a job execution unit 18.

  The standby server 20 stores an execution time DB 21 that describes the date and time when the job should be executed, an unexecuted job DB 22 that records whether or not the job is actually executed based on the execution time DB, and job dependencies necessary for flow control. The execution order DB 23, job definition information 24 storing information necessary for job execution, a schedule determination unit 25, a server determination unit 26, a dependency relationship determination unit 27, a job execution unit 28, and a job execution status synchronization unit 29.

  The execution status storage server 40 includes job end information 41.

  Each of these units operates as follows.

  The schedule determination unit 15 of the primary server 10 reads the execution time DB 11 and makes a job submission request to the server determination unit 16 when the date and time to be executed has come. Similarly, the schedule determination unit 25 of the standby server 20 reads the execution time DB 21 and makes a job submission request to the server determination unit 26 when the date to be executed has arrived.

  The server determination unit 16 of the primary server 10 transmits a job submission request to the dependency determination unit 17. The server determination unit 26 of the standby server 20 discards the job submission request and records in the unexecuted job DB 22 that the job has not been executed.

  In the primary server 10, since no data exists in the unexecuted job DB 12, the dependency determination unit 17 transmits a job submission request to the job execution unit 18.

  When there is data in the unexecuted job DB 22, the dependency determination unit 27 of the standby server 20 determines the dependency of the requested job based on the information in the execution order DB 23. A job submission request is transmitted to the execution unit 28.

  The job execution unit 18 of the primary server 10 executes the job based on the job submission request, and records the job end information 41 in the execution status storage server 40 when the end of the job is detected.

  The job execution status synchronization unit 29 of the standby server 20 acquires the job end information 41 from the execution status storage server 40 and deletes the job included in the job end information 41 among the jobs included in the unexecuted job DB 22.

  When the server determination unit 26 of the standby server 20 receives a notification of system switching by a system administrator or the like, the standby server 20 replaces the primary server 10 and performs job execution.

  The server determination unit 26 discards the job submission request when it is normal. However, after receiving the switching notification, the server determination unit 26 transmits a job submission request to the dependency determination unit 27.

  Next, the operation of the redundant system apparatus according to the present embodiment will be described in detail with reference to the drawings.

  FIG. 3 is a flowchart of the operations of the schedule determination units 15 and 25.

  The schedule determination units 15 and 25 read the execution time DBs 11 and 21 (step S10), and determine whether there is a job to be activated based on the defined date and time information (step S11). If there is a job to be activated, a job submission is requested (step S12).

  FIG. 4 is a flowchart of the operation of the server determination units 16 and 26.

  The server determination units 16 and 26 read information related to the main system or the standby system (step S20). It is determined whether or not it is the primary server 10 (step S21). If it is the primary server 10, a job submission request is transmitted to the dependency determination unit 17 (step S22). If the server 20 is a standby server 20, the fact that it has not been executed is recorded in the unexecuted job DB 22 (step S23).

  FIG. 5 is a flowchart of the operations of the dependency determination units 17 and 27 and the job execution units 18 and 28.

  The dependency determination units 17 and 27 determine whether there is a job in the unexecuted job DB (step S30). If there is no job in the unexecuted job DB (No in step S30), the job execution unit 18 , 28 execute the job (step S34). When there is a job in the unexecuted job DB (Yes in step S30), the dependency determination unit 27 reads the execution order DB 23 (step S31), and displays the execution status of the job having the dependency to be executed first. Confirm (step S32). It is determined whether or not the job having the dependency relationship has been executed (step S33). If the job has been executed (Yes in step S33), the job execution unit 28 executes the job (step S34). If it has not been executed (No in step S33), the job is terminated without executing it.

  When the job execution unit 18 detects the end of the executed job (Yes in step S35), the job execution unit 18 transmits the job end information to the execution status storage server 40, and the job status storage server 40 records the received job end information 41. (Step S36).

  FIG. 6 is a flowchart of the operation of the job execution status synchronization unit 29.

  The job end information 41 is acquired from the execution status storage server 40 (step S40), and it is determined whether or not the job recorded in the unexecuted job DB 22 is included (step S41). In (Yes in step S41), the completed job is deleted from the unexecuted job DB 22 (step S42).

  A redundant system apparatus according to a second embodiment of the present invention will be described with reference to the drawings.

  Referring to FIG. 7, the standby server 50 in this embodiment further includes an unexecuted job determination unit 30 in the standby server 20 in the redundant system device according to the first embodiment.

  The redundant system apparatus according to the first embodiment is configured not to execute the requested job when there is no execution result of the dependent job for the requested job.

  The unexecuted job determination unit 30 according to the present embodiment checks the unexecuted job having the dependency relationship based on the information extracted by the dependency relationship determination unit 27, and if there is an unexecuted job, the job This is executed by the execution unit 28. The job execution unit 28 executes the requested job after the unexecuted job having the dependency relationship is terminated.

  Although the above description has been made based on examples, the present invention is not limited to the above examples.

  In a batch job management system, it can be applied to a system switching when a disaster occurs, and a backup center can be easily constructed.

1 is a block diagram of a redundant system device according to an embodiment of the present invention. 1 is a block diagram of a redundant system device according to a first example of the present invention. FIG. It is a flowchart of operation | movement of the schedule determination part in the redundant system apparatus based on 1st Example of this invention. It is a flowchart of operation | movement of the server determination part in the redundant system apparatus based on 1st Example of this invention. It is a flowchart of operation | movement of the dependence determination part and the job execution part in the redundant system apparatus based on 1st Example of this invention. It is a flowchart of operation | movement of the job execution condition synchronization part in the redundant system apparatus based on 1st Example of this invention. It is a block diagram of the redundant system apparatus which concerns on 2nd Example of this invention.

Explanation of symbols

10 Main servers 11, 21 Execution time database (DB)
12, 22 Unexecuted job DB
13, 23 Execution order DB
14, 24 Job definition information 15, 25 Schedule determination unit 16, 26 Server determination unit 17, 27 Dependency determination unit 18, 28 Job execution unit 20, 50 Standby server 29 Job execution status synchronization unit 30 Unexecuted job determination unit 40 Execution status storage server 41 Job end information

Claims (5)

A main server having an execution time database (hereinafter referred to as DB) storing job execution times and an execution order DB storing execution orders of a plurality of jobs that are mutually dependent;
A redundant system device comprising: a standby server having an execution time DB, an execution order DB, and an unexecuted job DB storing a job not executed in the primary server;
In a normal state, the primary server refers to the execution time DB and the execution order DB, executes the job, notifies the standby server of the executed job, and the standby server is notified. Configured to update the unexecuted job DB based on the executed job,
At the time of abnormality, the standby server refers to the execution time DB, the execution order DB, and the unexecuted job DB, and is a job that is not executed on the primary server, and has a dependency relationship with the job. A redundant system device configured to execute a job that has been executed.   At the time of abnormality, the standby server refers to the execution time DB, the execution order DB, and the unexecuted job DB, and is a job that is not executed on the primary server and has a dependency relationship with the job. The redundant system apparatus according to claim 1, wherein when a job has not been executed, the job is executed after a job having a dependency relationship is executed. A main server having an execution time database (hereinafter referred to as DB) storing job execution times and an execution order DB storing execution orders of a plurality of jobs that are mutually dependent;
A job execution method in a redundant system device including the execution time DB, the execution order DB, and a standby server including an unexecuted job DB storing a job that is not executed in the primary server,
A step in which the main server executes a job with reference to the execution time DB and the execution order DB at a normal time;
Notifying the standby server of the executed job;
The standby server updates the unexecuted job DB based on the notified executed job, and
At the time of abnormality, the standby server refers to the execution time DB, the execution order DB, and the unexecuted job DB, and is a job that is not executed on the primary server, and has a dependency relationship with the job. A method for executing a job in a redundant system apparatus, comprising a step of executing a job that has been executed.   At the time of abnormality, the standby server refers to the execution time DB, the execution order DB, and the unexecuted job DB, and is a job that is not executed on the primary server and has a dependency relationship with the job. 4. The method for executing a job in a redundant system apparatus according to claim 3, further comprising a step of executing the job after execution of the job having dependency relation when the job has not been executed. A main server having an execution time database (hereinafter referred to as DB) storing job execution times and an execution order DB storing execution orders of a plurality of jobs that are mutually dependent;
A job execution program in a redundant system device including the execution time DB, the execution order DB, and a standby server including an unexecuted job DB storing a job that is not executed in the primary server,
A process of causing the primary server to execute a job with reference to the execution time DB and the execution order DB at a normal time;
A process for causing the primary server to notify the standby server of an executed job;
Causing the standby server to execute a process of updating the unexecuted job DB based on the notified executed job,
At the time of abnormality, the standby server is referred to the execution time DB, the execution order DB, and the unexecuted job DB, and is a job that is not executed on the primary server and has a dependency relationship with the job. A job execution program in a redundant system apparatus, characterized in that a computer executes a process for executing a job that has already been executed.

Images