System for automatically avoiding frequent occurrence of deadlock

Description: TECHNICAL FIELD [Detailed description of the invention]

【0001】

[Industrial application field]

The present invention relates to a deadlock-prone automatic avoidance method, and particularly in an information processing system in which a plurality of jobs simultaneously use a shared resource such as a database and perform processing in parallel while receiving exclusive control, dead within one processing section. Regarding the deadlock frequent occurrence automatic avoidance method that automatically avoids the frequent occurrence of locks.

【0002】

[Conventional technology]

In an information processing system such as a database system in which two or more jobs use the same resource exclusively and process in parallel, the two jobs request each other's resources and each other. A deadlock phenomenon occurs in the waiting state. As a measure to avoid this, when a deadlock occurs, generally, the deadlock occurrence job on the side that generated the deadlock state by requesting a resource later is already executed in the processing section currently being executed. The completed processing is canceled and re-executed from the start point of the processing section. With this measure, the immediate deadlock can usually be avoided.

【0003】

As described above, the deadlock that occurred at a certain point in the processing interval is avoided by re-executing the deadlock occurrence job from the start point, and the deadlock target job that became the deadlock partner is the deadlock target job at the deadlock occurrence point. You can continue processing from the processing of, but if the processing interval is long and complicated, the deadlock phenomenon will occur again at other points in the same processing interval, and this time if you are a deadlock occurrence job, processing Cannot be continued and will be returned to the start point of the processing section. Therefore, if two jobs alternately cause deadlocks within one processing section, both jobs alternately start processing from the processing section start point (for example, database update processing) and deadlock. A phenomenon (called multiple deadlocks) occurs in which the target process cannot be completed simply by repeating the update cancel process by locking.

【0004】

When such a multiple deadlock occurs, conventionally, a method has been used in which one of the jobs is abnormally terminated to continue the processing of the other job.

【0005】

[Problems to be Solved by the Invention]

In the conventional method of avoiding multiple deadlocks described above, when deadlocks occur frequently within one processing section, one of the jobs is forcibly terminated abnormally and execution is stopped, so processing of both jobs is performed. Could not be continued, and operator intervention was required to execute the job that ended abnormally.

【0006】

An object of the present invention is to provide a deadlock frequent occurrence automatic avoidance method capable of continuously processing both jobs without abnormal termination of jobs when deadlocks occur frequently within one processing section.

【0007】

[Means for solving problems]

The deadlock-prone automatic avoidance method of claim 1 is a dead that occurs in the processing section being executed for each job in an information processing system in which a plurality of jobs exclusively use shared resources to process in parallel. A deadlock occurrence count storage table that holds the number of lock occurrences, and a deadlock occurrence count storage table that corresponds to a job that has been returned to the start point of the processing section and instructed to re-execute when a deadlock occurs. A job pause means that suspends the job when the updated value reaches a predetermined limit value and a job that is stopped by the job pause means. A stop time monitoring means that monitors the pause time and outputs an end notification when the specified time elapses, and a deadlock of the corresponding job in the deadlock occurrence count storage table by canceling the pause of the corresponding job in response to the end notification. A job stop release means for returning the number of occurrences to the initial state and a job processing section end processing means for returning the number of deadlock occurrences of the corresponding job in the deadlock occurrence number storage table to the initial state each time the processing of each job processing section is completed. It is configured to prepare.

【0008】

The deadlock frequent occurrence automatic avoidance method according to claim 2 is the deadlock frequent occurrence automatic avoidance method according to claim 1. In the deadlock frequent occurrence automatic avoidance method, the limit value for the job suspension means to suspend the job is limited to the number of deadlock occurrences for each job. The feature is that it is set in the value table.

【0009】

The deadlock frequent occurrence automatic avoidance method according to claim 3 is the deadlock frequent occurrence automatic avoidance method according to claim 1 or 2, wherein the specified time until the stop time monitoring means releases the temporary stop of the job is a job. The feature is that it is set in the job stop setting time table for each job.

【0010】

The deadlock frequent occurrence automatic avoidance method according to claim 4 is the deadlock frequent occurrence automatic avoidance method according to claim 2 or 3, wherein at least the limit value for each job in the deadlock occurrence number limit value table is set for each job. The feature is that it is set by the job control statement at startup.

【0011】

[Example]

Next, examples of the present invention will be described with reference to the drawings.

【0012】

FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention.

【0013】

In the deadlock frequent occurrence automatic avoidance method of this embodiment, as shown in FIG. 1, information that two jobs, a deadlock occurrence job 8 and a deadlock target job 9, update by exclusively using the database resource 10. In the processing system, the deadlock occurrence count storage table 5 that holds the number of deadlock occurrences that occurred in the processing interval during execution of each job and the deadlock occurrence count return to the start point of the processing interval when a deadlock occurs. The deadlock occurrence count storage table 5 corresponding to the job instructed to be re-executed is updated by adding 1 to the deadlock occurrence count limit value. Job pause means 1 that suspends a job, and stop time monitoring means 2 that monitors the pause time of a stopped job and outputs a end notification when the specified time set in the job stop set time table 7 elapses. , The job stop release means 3 that receives the end notification and cancels the suspension of the corresponding job and returns the deadlock occurrence count of the corresponding job in the deadlock occurrence count storage table 5 to the initial state, and each time the processing of each job processing section ends. It is configured to include a job processing section end processing means 4 that initializes the deadlock occurrence count of the corresponding job in the deadlock occurrence count storage table 5.

【0014】

When two jobs running in the system, deadlock occurrence job 8 and deadlock target job 9 are both updating database resource 10, deadlock occurrence job 8 is deadlocked against deadlock target job 9. When a lock occurs, the deadlock occurrence job 8 is canceled by the system up to the start point of the processing section, and the job suspension means 1 is notified of the deadlock occurrence.

【0015】

The job suspension means 1 updates the deadlock occurrence number of the deadlock occurrence job 8 in the deadlock occurrence count storage table 5 by adding 1 to the updated occurrence count and the deadlock occurrence count limit value table 6. Compare with the limit number of deadlock occurrence job 8 of. If the number of occurrences has not reached the limit, the deadlock occurrence job 8 is made to re-execute the update processing of the database resource 10 from the start point of the processing section. When the number of occurrences has reached the limit, the job suspension means 1 suspends the processing of the deadlock occurrence job 8 and notifies the stop time monitoring means 2 that the deadlock occurrence job 8 has been stopped. ..

【0016】

The stop time monitoring means 2 monitors the stop time of the deadlock occurrence job 8, and when the stop time exceeds the specified time of the deadlock occurrence job 8 set in the job stop setting time table 7, the job stop release means 3 is set. Notice.

【0017】

The job stop release means 3 releases the temporary stop of the deadlock occurrence job 8 and changes the deadlock occurrence count of the deadlock occurrence job 8 in the deadlock occurrence count storage table 5 to the initial state (0). The number of deadlock occurrences of each job in the deadlock occurrence count storage table 5 is changed to the initial state by the job processing section end processing means 4 each time the processing section of each job ends.

【0018】

Although the deadlock occurrence job 8 has been mainly described above, the same processing is performed for the deadlock target job 9. In this way, when deadlocks occur frequently within one processing section and the number of occurrences reaches the specified value, the job processing is temporarily stopped and the processing section start timing is shifted from the other job to create two jobs. It is possible to avoid multiple deadlocks in which processing of the same processing section is alternately repeated halfway without the intervention of an operator, and to continue both jobs.

【0019】

In the above-described embodiment, the deadlock occurrence number limit value table 6 and the stop time job stop setting time table 7 are provided, and the limit value and the specified time for each job are set for each, and the job pause means. 1 and the stop time monitoring means 2 shall notify the pause and release by referring to the limit value and the specified time corresponding to each job, but instead of setting the limit value and the specified time for each job, The same effect can be expected even if the same value is applied uniformly to all jobs without providing a table.

【0020】

If a table is provided as in the above embodiment so that different values can be set for each job, the set values can be changed according to the length of the processing section and the importance of the job, so more efficient operation is possible. There is an advantage that becomes. Only the setting values for specific jobs may be registered in the table, and uniform values (for example, a small limit value and a long specified time) may be applied to unregistered jobs. The table setting values may be set separately, but if they are registered by the job control statement when the job is started, effective operation can be performed without requiring extra setting man-hours or memory area.

【0021】

[Effect of the invention]

As described above, in the deadlock frequent occurrence automatic avoidance method of the present invention, when a plurality of jobs exclusively use the same resource at the same time as in database update processing, both jobs alternately have the same processing section. It is possible to automatically avoid the occurrence of the multiple deadlock phenomenon in which the deadlock is repeated and the job does not end without the intervention of the operator, and there is an effect that both jobs can be continuously processed without abnormal termination.

[Simple explanation of drawings]

[Figure 1]

It is a block diagram which shows the structure of one Example of this invention.

[Explanation of symbols]

1 Job suspension means 2 Downtime monitoring means 3 Job stop release means 4 Job processing section end processing means 5 Deadlock occurrence count storage table 6 Deadlock occurrence limit value table 7 Job stop setting time table 8 Deadlock occurrence job 9 Deadlock target job 10 Database resources