JPH1069390A - Device and method for task stall - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain the device and method for task stall monitoring which monitor the stall state of an event-driven process task and also monitors a state wherein its driving is disabled for some reason. SOLUTION: An external timer means 1 issues a task call start request to process tasks 61-6n by a repetitive task call means 3 at constant intervals. A task status holding means 4 registers the answer contents of a task call answering means 7 in task status control information 40. The task call means 3 monitors the end of a current process when obtaining an 'in-process' answer, monitors driving by a next start request when there is a call answer without any sign of driving, and considers that a task has a task stall or can not be driven when the task makes no 'call answer' or 'in-process' does not end even after making a call as many times as a set limit time by process tasks, thereby driving a fault recovering means 5.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a task stall detection of a computer, and more particularly to a task stall detection apparatus and method for detecting when a task cannot be started and cannot be completed and when the task cannot be started.

[0002]

2. Description of the Related Art In an online system operated continuously for 24 hours by a computer, when a task stall occurs due to logical inconsistency such as a bug in a program,
If the task cannot be started for some reason, the computer will not operate properly and the online system will become abnormal. When such a state occurs, a method of detecting a task stall is provided as a means for automatically taking a recovery means immediately.

Conventionally, this type of task stall detection method is a method used in an online system or the like to create a trigger for detecting and recovering from a system abnormality. An example of a conventional stall detection method is disclosed in Japanese Patent Laid-Open No. 5-26580.
No. 7 is described. The task stall detection method described in this publication issues a stall monitoring request at the start or end of execution of a task, and when a stall monitoring external timer that receives this request detects a timeout while setting the timer, The notification is made to the system interruption notifying means.

[0004]

SUMMARY OF THE INVENTION The above-mentioned JP-A-5-26
The first problem with the task stall detection method disclosed in Japanese Patent Publication No. 507 is that, after a task is started, it can be monitored that the task has entered a permanent loop state and has entered a stalled state. If it cannot be started for some reason, stall monitoring cannot be performed. The reason is that when the processing task is started, the stall monitoring request is made and the stall monitoring is performed at the start or end of execution of the task. However, if the processing task is not started, the stall monitoring request cannot be made. Is not performed.

An object of the present invention is to realize a method of detecting that a plurality of processing tasks of a computer have become stalled due to a bug or the like and detecting that a plurality of processing tasks cannot be started for some reason. Things.

[0006]

According to the present invention, when a computer has a plurality of processing tasks under its control and controls the activation of a processing task corresponding to an event to occur, the processing task is activated and enters a stalled state. In the task stall detection device of the computer that monitors whether there is,
External timer means for generating a signal for periodically activating the monitoring of the processing task, task call means for receiving the signal and calling each processing task, and "processing", "no response", and "no response" to the processing task With the three responses of `` with response '', the number of calls and the limit value of the number of calls, the task status control information for managing the state of the processing task, the task status control information is updated by the roll call response,
Task status holding means for detecting an abnormality of the processing task and notifying the task call means, and failure recovery means for receiving a notification of the processing task abnormality from the task call means and performing failure processing of the abnormal processing task. In response to the roll call of the task roll call means, the processing task responds to the task status holding means whether its state is "processing" or "call roll response". And roll call response means.

The task status control information includes:
Processing task name and "processing" or "no roll call response"
Or a task status of “with a roll call response”, and the number of times that the “in process” and “without a roll call response” were continuously responded,
And a value of a limit number of times of the continuous response.

Further, when the computer has a plurality of processing tasks under its control and controls the activation of the processing task corresponding to the event that occurs, the computer monitors whether the processing task is activated and is in a stalled state. In the task stall detection method of the computer, each processing task is periodically rolled, and the state of the processing task itself is monitored by three responses of “processing”, “no call response”, and “call response”, When the response of "under processing" and the state of "no roll call response" continue for a predetermined number of roll calls, the processing task is determined to be in a stall state or an unstartable state, and when responding to the roll call with "response", The processing task is determined to be normal.

The task call means for monitoring the stall is periodically started by an external timer means. The started task roll call means issues a start request to sequentially call each of the processing tasks started by the occurrence of the event. If each processing task is normal, it is sequentially activated by an activation request by the task call means. In order to respond to the task call means, the task call response means located in each processing task requests the task status holding means to register its own status in the roll call response. The status holding unit checks the registered roll call response status, and counts the number of roll calls for the processing task of “no call response” or the processing task that has responded “under processing”. When the number of times of roll calls reaches the limit value by counting the number of times of roll calls, system stall is detected as task stall or task inability to be activated, and fault recovery means is activated. The failure recovery means executes recovery processing of the processing task.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described with reference to the drawings.

Referring to FIG. 1 showing the configuration of one embodiment of the present invention, a task stall detecting device is a computer 2
Task call monitoring means 3 for monitoring task stall, task call response means 7 located in a plurality of processing tasks 61 to 6n, task status holding means 4, failure recovery means 5, external timer means 1, , Is provided.

The task call means 3 for task stall monitoring started by the external timer means 1 is set to a higher running level than the processing tasks 61 to 6n started by the occurrence of an event. The external timer means 1 periodically starts the task stall monitoring means 3 for task stall monitoring. The task roll call means 3 transmits a response to the previous task roll call request to the task status control information 40 of the task status holding means 4,
Check in 2. If the response is "call roll response", a task roll call activation request is made to tasks 61 to 6n again. If the task status 42 is “processing” or “no call response” and reaches the limit value 44 due to the update of the task call count 43, it is regarded as a failure due to task stall detection or task activation abnormality, and failure recovery is performed. The means 5 is activated. The fault recovery means 5 executes a fault recovery process of the fault task.

The operation of this embodiment will be described with reference to FIGS. 2, 3 and 4. In FIG. 2, when the task call means 3 for stall monitoring started by the external timer means 1 is started, a task status 4 corresponding to the first task 61 is sent from the task status holding means 4.
2 is taken out (step 20). If the retrieved task status 42 is "call response" to the task roll request (YES in step 21), the task 6
1 normally operates, resets the task status 42 of the task status holding means 4 and the number of times of task call 43 (step 29), and issues a call call activation request to the next processing task 62 (step 25). Task roll call means 3
The task status 42 of the next processing task 62 is checked (step 26).

In step 21, if the content of the task status 42 is not "call response" but "no call response" (NO in step 21), the task 61 which requested the roll activation is in process or for some reason. Indicates that it could not be started. Therefore, the task call number 43
Is updated (step 22), and when the result reaches the task call limit value 24 (YES in step 23), it is regarded as task stall detection and the failure recovery means 5 is activated (step 28). If the content of the task status 42 is "processing", "processing" is registered to wait until the processing of the processing task 61 is completed (YES in step 24), and the task status of the next task 62 is registered. 42 is extracted (step 26), and the process returns to the stall monitoring check as in the case of the first processing task 61 (N in step 27).
O). Check the above processing tasks for the last processing task 6
After the execution up to n (YES in step 27), the task call means 3 for stall monitoring ends.

In FIG. 3, when the processing task 61 is activated, the task roll call responding means 7 checks whether the task 61 itself has been activated by the activation request by the task roll call means 3 or activated by the occurrence of an event (see FIG. 3). Step 31). In the case of activation by the former (Step 3
(1) (YES), a registration request 22 indicating "calling response" is made in the task status 42 of the task status holding means 4, and the process is terminated (step 34). In the case of activation by the latter (NO in step 31), the task status holding unit 4
A request is made to register "processing in progress" in the task status 42 (step 32), and the processing process of the task 61 is executed (step 33). When the execution of the processing process is completed, a request is made to register "call roll response" in the task status 42 of the task status holding means 4, and all the processes are completed (step 34).

FIG. 4 shows status control information 40 defined by the number of tasks, with a processing task name 41, a task status 42, a task call count 43, and a task call limit value 44 as one component. ing. The task status 42 is an area for registering the status of "call roll response", "no call response", and "processing". The task roll call frequency 43 is an area counted by the task roll call means 3 when the content of the task status 42 is "processing" or "no roll call response". The task call limit value 44 is used by the task call means 3 to count the number 23 of task call calls and to compare whether the limit value 44 has been reached. When this value is reached, whether the task 61 is in a stalled state, Or processing task 6
1 detects that it cannot be started for some reason. The limit value 44 can be set variably for each of the processing tasks 61 to 6n, and the stall monitoring time can be adjusted by the set value of the task call limit value 44.

[0017]

As described above, the first effect of the present invention is that the task call means monitors whether or not the processing task can be activated simultaneously with the detection of the stall of the processing task. Thus, a failure due to a factor other than a stall during the task processing can be detected. The reason is that a task activation request is issued to the task for a task roll call, and a response result by the activation is periodically monitored.

[Brief description of the drawings]

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

FIG. 2 is a flowchart illustrating a process of a task call unit in FIG. 1;

FIG. 3 is a flowchart illustrating a task call response means process of FIG. 1;

FIG. 4 is a diagram showing items of status control information of a task status holding unit.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 External timer means 2 Computer 3 Task call means 4 Task status holding means 5 Failure recovery means 7 Task call response means 61 to 6n Processing tasks

Claims (3)

[Claims] When a computer has a plurality of processing tasks under its control and controls activation of a processing task corresponding to an event to occur, the computer monitors whether a processing task is activated and is in a stalled state. In the task stall detecting device of the computer, an external timer means for generating a signal for periodically starting monitoring of the processing task; a task call means for receiving the signal to call each processing task; , "No response", "Response", task count control information for managing the state of the processing task by the number of calls and the limit value of the number of calls, and the task status control by the call response. A task status holding unit for updating information, detecting an abnormality in a processing task, and notifying the task call unit, Upon receiving the notification of the processing tasks abnormality call means, and fault recovery means for the failure processing of the abnormality processing task,
The processing task, upon receiving the roll call of the task roll call means, responds to the task status holding means whether its own state is "processing" or "call roll response". A task stall detecting means, comprising: 2. The task status control information includes a processing task name, “processing” or “no call response”.
Or a task status of “with a roll call response”, and the number of times that the “in process” and “without a roll call response” were continuously responded,
The task stall detection device according to claim 1, further comprising a value of a limit number of times of the continuous response. 3. When a computer has a plurality of processing tasks under its control and controls activation of a processing task corresponding to an event to occur, the computer monitors whether the processing task is activated and is in a stalled state. In the task stall detection method of the computer, each processing task is periodically rolled, and the state of the processing task itself is monitored by three responses of “processing”, “no call response”, and “call response”, When the response of "under processing" and the state of "no roll call response" continue for a predetermined number of roll calls, the processing task is determined to be in a stall state or an unstartable state, and when responding to the roll call with "response", A task stall detection method, wherein the processing task is determined to be normal.