JP2003296140A - Object monitoring system, object monitoring method and object monitoring program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a structure for improving resistance to fault in a system for remotely monitoring the state of a plurality of objects to be monitored present on a computer. <P>SOLUTION: This object monitoring system, in which a monitoring side computer monitors objects to be monitored in a computer to be monitored, is characterized in that the computer to be monitored is provided with object monitoring program parts present by the number of the objects to be monitored, each of which collects the state of one object to be monitored predefined not to overlap from the objects to be monitored at fixed time intervals, and reports state information including the collected state and the time of collection; and an agent program part for acquiring the state information from each of the object monitoring program parts followed by storing, and collectively reporting, in response to an inquiry from the observer computer, the stored state information of the objects to be monitored. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an object monitoring system, an object monitoring method and an object monitoring program, and more particularly to an object monitoring system, an object monitoring method and an object monitoring program which can improve fault tolerance.

[0002]

2. Description of the Related Art An example of a conventional computer monitoring system is disclosed in Japanese Patent Laid-Open No.
01-005691. This computer monitoring system is composed of a manager program section, a communication path, an agent program section, and a monitoring target. The manager program unit of the conventional computer monitoring system having such a configuration inquires of the agent program unit about the state of the monitoring target operating on the computer, and the agent program unit receives the request, collects information, and responds. To do.

[0003]

However, this conventional technique has the following problems. The first problem is
If the load on the monitored computer increases, the agent program part will also be affected, so even if a request from the manager program part is received, the monitored information cannot be collected, and even if it is collected, it will respond to the manager program part. The point is that you cannot do it. The reason is that the agent program unit itself requires a lot of resources to operate. The second problem is that, when a failure occurs in any of the monitoring targets, the agent program unit that tries to collect the information of the monitoring target that has failed cannot continue its operation. It is not possible to judge whether the operation of the agent program part cannot be continued due to the cause. The reason is that the process of collecting the information to be monitored and the process of responding the information to the manager program unit are integrated. It is an object of the present invention to provide a system for remotely monitoring the states of a plurality of monitoring target objects existing on a computer and providing a configuration for improving fault tolerance.

[0004]

An object monitoring system of the present invention is an object monitoring system in which a monitoring source computer monitors a monitoring target object in a monitoring target computer connected to the monitoring source computer via a communication path. The monitoring target computer exists in the number of the monitoring target objects, and each of the monitoring target objects has a predetermined non-overlapping state of one monitoring target object that is collected at a constant time interval and the collected state And an object monitoring program unit for notifying the state information including the collected time, and acquiring and storing the state information from each of the object monitoring program unit,
When an inquiry is received from the monitoring source computer, an agent program unit that collectively notifies the monitoring source computer of the stored state information of the monitoring target object, and each of the object monitoring program units, The object monitoring program unit operates independently of the monitoring target objects other than the monitoring target objects monitored by the object monitoring program unit, and the agent program unit is not affected by the operation of the object monitoring program unit and the failure of the monitoring target object. It is characterized in that the monitoring source computer makes an inquiry to the agent program unit at a constant time interval to acquire the status information of the monitoring target object, and if there is a status information that is not collected, Is the status information collected Fault monitoring target object no one is characterized by comprising a manager program unit determines that occurred.

The object monitoring method of the present invention is an object monitoring method for a monitoring source computer to monitor a monitoring target object in a monitoring target computer connected to the monitoring source computer by a communication path. There are as many monitoring target objects as possible, each of which includes a predetermined non-overlapping state of one of the monitoring target objects at predetermined time intervals, and includes the collected state and the collected time. The step of notifying the state information, and the monitored computer acquires and stores the state information from each of the object monitoring program units, and stores it when an inquiry is received from the monitoring source computer. The status information of the monitoring target objects is collectively notified to the monitoring source computer. When the monitoring source computer makes an inquiry to the agent program unit at a fixed time interval to acquire the status information of the monitoring target object, and there is a status information that has not been collected, And a step of determining that a failure has occurred in the monitored object for which status information has not been collected.

The object monitoring program of the present invention is an object monitoring process for a monitoring source computer to monitor a monitoring target object in a monitoring target computer connected to the monitoring source computer by a communication path. There are as many monitoring objects as possible, each of which collects the status of one predetermined, non-overlapping monitoring object of the monitoring objects at regular time intervals, and notifies the status information including the collected status and the collection time. The processing and the status information from each of the object monitoring program units are acquired and stored, and when an inquiry is received from the monitoring source computer, the status information of the stored monitoring target objects is collectively stored as described above. The process of notifying the monitoring source computer is executed by the monitoring target computer and If the status information of the monitored object is acquired by inquiring to the agent program section at intervals, and the status information is not collected for some objects, a failure occurs in the monitored object for which the status information has not been collected. It is characterized in that the monitoring source computer is caused to execute a process of determining that the occurrence has occurred.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described in detail with reference to the drawings.

First, the first embodiment of the present invention will be described in detail. Referring to FIG. 1, in the first exemplary embodiment of the present invention, a monitoring source computer 100 and a monitoring target computer 2
00 and 00. These computers operate under program control and are connected to each other via a communication path 300. The monitoring source computer 100 includes a manager program unit 110. The monitoring target computer 200 includes an agent program unit 210, object monitoring program units 220-1, 220-2, 220-3, and monitoring target objects 230-1, 230-2, 230-.
3 and 3. The manager program unit 110 communicates with the agent program unit 210 via the communication path 300, and acquires the information collected by the agent program unit 210. The agent program section 210 is
Object monitoring program units 220-1, 220-
2, 220-3 through the computer internal communication path 400 to communicate with the object monitoring program units 220-1, 2
Information regarding the states of the monitoring target objects 230-1, 230-2, and 230-3 collected by the 20-2 and 220-3 is acquired. Here, the agent program unit 21
0 is composed of the minimum necessary computer resources, and uses the memory lock function and the task priority control function provided by the operating system of the monitoring target computer 200 to execute other programs (object monitoring program) on the monitoring target computer 200. The operation is performed so as not to be affected by the operation of the unit 220 and the failure of other resources (including the monitoring target object). The memory lock function is one of the functions provided by the operating system.It has a function that secures the memory required by the target task in advance and locks it to prevent the paging mechanism from taking the memory from the task. Say. As a result, the monitored computer 200
Even when the memory load is high, the agent program unit 210 can continue to operate in the memory secured in advance. Generally, since the amount of memory that can be locked is limited, the memory cannot be locked for all modules. Therefore, the agent program unit 210 limits its function and minimizes the required resource amount. The task priority control function refers to a function of controlling the execution priority of each task operating on the system. By setting the task priority of the agent program unit 210 higher than that of other tasks, the agent priority control function The program unit 210 can operate even when the load on the monitored computer 200 is high. The object monitoring program units 220-1, 220-2, 220-3 monitor the monitoring target objects 230-1, 230-2, 230-3 operating on the monitoring target computer, respectively, and their states, information collection times, etc. To the agent program unit 210 as status information. It should be noted that each of the object monitoring program units 220-1, 220-2, 220-3 operates without depending on a monitoring target object other than the monitoring target object that it monitors. next,
The overall operation of this embodiment will be described in detail with reference to FIGS. Here, as an example, the monitoring target object 230-1 is a file on a disk, the monitoring target object 230-2 is a free memory, and the monitoring target object 230-3 is an execution waiting task. FIG. 2 shows the object monitoring program units 220-1, 220-2,
220-3 shows a processing flow of 220-3. The object monitoring program units 220-1, 220-2, 220-3 have monitoring target objects 230-1, 230-2, 230-.
The state of No. 3 is collected (step A1), and it is used as the state information together with the information collection time in the agent program unit 2
When notified to 10 (step A2), the agent program unit 210 acquires and stores the state information.
Then, the object monitoring program units 220-1, 2
20-2 and 220-3 perform a waiting process (step A3) at a predetermined monitoring interval, and return to step A1 again. FIG. 3 shows the agent program section 21.
0 represents a processing flow of 0. Agent program section 2
10 waits for an inquiry from the manager program unit 110 (step B1), and when the inquiry is received (step B2), collectively stores the stored state information (step B3) and notifies the manager program unit 110. (Step B4). FIG. 4 shows a processing flow of the manager program unit 110. The manager program section 110 instructs the agent program section 210 to monitor objects 230-1, 230-2, 2
Status information is acquired by inquiring about the status of 30-3 (step C1), and each monitoring target object 230 is acquired.
The information collection times of -1, 230-2 and 230-3 are checked (step C2). When it is detected that the status information cannot be collected at a predetermined monitoring interval, the corresponding object monitoring program 23
It is detected that an abnormality has occurred in 0-1, 230-2, and 230-3 (step C3). If no particular abnormality is detected, a waiting process (step C4) according to the monitoring interval is performed, and the process returns to step C1. Here, it is assumed that a failure has occurred in the monitored object 230-1, that is, the file on the disk. In this case, the object monitoring program unit 220-1 cannot collect the state in step A1 and does not notify the state information in step A2. On the other hand, the other object monitoring program units 220-2 and 220-3 have monitoring target objects 220-2 and 220 to be monitored, respectively.
-3 is different, the status information can be notified to the agent program unit 210 in step A2 without being affected even if the monitored object 220-1, that is, the file on the disk, becomes a failure. When the agent program section 210 receives a request from the manager program section 110 in this state (step B2), the agent program section 210 notifies the manager program section 110 of the state information of the monitoring target object 230 (steps B3 and B4). When the manager program unit 110 acquires the state information from the agent program unit 210 (step C1), it detects that the state information of the monitoring target object 230-1 has not been collected (steps C2 and C3), and the monitoring target object is detected. It is detected that a failure has occurred in 230-1 (step C5). As described above, ideally, as in the present embodiment, it is preferable to monitor one object to be monitored 230 in association with one object monitoring program unit 220, but to monitor one object to be monitored 230. One object monitoring program unit 220 may not be associated with each other. For example, to monitor two files on the same disk and the controller (interface card etc.) connected to that disk,
The object monitoring program 220 is associated with each of the two files and monitored. In this case, each of the two object monitoring programs 220 needs to access and monitor the file via the controller, and it is assumed that the controller (interface card or the like) connected to the disk is operating normally. However, if the two object monitoring programs 220 cannot collect the status information of the monitored object at the same time, the two object monitoring programs 220 need to store in the controller or disk which is a common part for monitoring the files. It can be judged that a failure has occurred. In this way, more information may be acquired in some cases by operating a plurality of object monitoring programs 220 that use the common part. Next, a second embodiment of the present invention will be described in detail with reference to the drawings. In FIG. 1, the monitoring target object 230 is replaced by the monitoring target computer 20.
0 is defined as various functions provided by the system, and each object monitoring program unit 220 is configured to use each monitoring target object 230. In addition, the object monitoring program unit 220 has a plurality of functions.
May be configured to be used by. Each object monitoring program unit 220 periodically confirms that the function operates normally, and the result is confirmed by the agent program unit 21.
Notify 0. Subsequent operations are the same as those in the first embodiment, so description thereof will be omitted. As described above, it is possible to confirm that the various functions provided by the system operate normally, and conversely, it is possible to detect a failure in the monitored computer by detecting that the function does not operate normally. Next, a third embodiment of the present invention will be described in detail with reference to the drawings. In the configuration of the first embodiment, in particular, the agent program unit 210 is connected to the monitored computer 2
00 system core (eg UNIX (registered trademark)
If so, include it in the kernel part). With such a configuration, the agent program section can be separated from other general tasks, and the fault tolerance can be further increased.

[0009]

As described above, the present invention has the following effects. The first effect is that the fault tolerance of the monitoring system is improved. The reason is that the resources required for the operation of the agent program part on the monitored computer, which is the core of the monitoring system, are minimized and the operation of this part is guaranteed. Also, by dividing each object monitoring program for each monitoring target, it is configured that a failure of a certain object monitoring program does not affect the operations of the agent program unit and other object monitoring programs. The second effect is that the existence of the object monitoring program for each monitoring target object makes it possible to detect where the failure on the monitoring target computer has occurred. The reason is that the object monitoring program is subdivided for each monitoring target.

[Brief description of drawings]

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

FIG. 2 is a flowchart of the operation of the first embodiment of the present invention.

FIG. 3 is a flowchart of the operation of the first exemplary embodiment of the present invention.

FIG. 4 is a flowchart of the operation of the first embodiment of the present invention.

[Explanation of symbols]

100 monitoring source computer 110 Manager Program Department 200 Computers to be monitored 210 Agent Program Department 220-1 Object Monitoring Program Unit 220-2 Object monitoring program section 220-3 Object Monitoring Program Unit 230-1 Monitoring target object 230-2 Monitoring target object 230-3 Monitoring target object 300 communication channels 400 Computer internal communication path

Claims (3)

[Claims] 1. An object monitoring method in which a monitoring source computer monitors a monitoring target object in a monitoring target computer connected to the monitoring source computer via a communication path, wherein the monitoring target computer is the number of the monitoring target objects. An object that collects the status of one predetermined, non-overlapping monitored object among the monitored objects at regular time intervals, and notifies the status information including the collected status and the collected time. The monitoring program section and the object monitoring program section acquire and store the state information, and when an inquiry is received from the monitoring source computer, the stored state information of the monitored object is summarized. And an agent program unit for notifying the monitoring source computer. Each of the object monitoring program units operates independently of monitoring target objects other than the monitoring target objects monitored by the object monitoring program unit, and the agent program unit operates and monitors the object monitoring program unit. The monitoring source computer performs an inquiry to the agent program unit at a fixed time interval to acquire the status information of the monitoring target object and collects the status information. An object monitoring method characterized by comprising a manager program section that judges that a failure has occurred in the monitored object for which the status information has not been collected, if there is any that has not been collected. 2. An object monitoring method, wherein a monitoring source computer monitors a monitoring target object in a monitoring target computer connected to the monitoring source computer by a communication path, wherein the monitoring target computer is the number of the monitoring target objects. Exists, each of which collects the state of one predetermined non-overlapping monitored object of the monitored objects at regular time intervals, and notifies the state information including the collected state and the collected time. And the monitoring target computer acquires and stores the state information from each of the object monitoring program units, and when an inquiry is received from the monitoring source computer, the status of the stored monitoring target object A step of collectively reporting information to the monitoring source computer, and the monitoring source computer If the status information of the monitoring target object is acquired by inquiring to the agent program unit at regular time intervals, and there is any status information that has not been collected, the monitoring target object for which the status information has not been collected And a step of determining that a failure has occurred in the object monitoring method. 3. An object monitoring process in which a monitoring source computer monitors a monitoring target object in a monitoring target computer connected to the monitoring source computer via a communication path, and there are as many monitoring target objects as there are monitoring target objects. A process of collecting the state of one predetermined non-overlapping monitored object of the monitored objects at regular time intervals and notifying the status information including the collected status and the collected time; and the object monitoring program A process of acquiring and storing the state information from each of the units, and collectively reporting the stored state information of the monitoring target objects to the monitoring source computer when an inquiry is received from the monitoring source computer. Is executed by the computer to be monitored, and the agent program section is executed at regular time intervals. If the status information of the above-mentioned monitored objects is acquired by performing a match and there is any status information that has not been collected, it is determined that a failure has occurred in the monitored object for which the status information has not been collected. An object monitoring program that causes the monitoring source computer to execute the processing to be performed.