JP3248485B2 - Cluster system, monitoring method and method in cluster system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a cluster system, and more particularly, to a monitoring method and a monitoring method in a cluster system.

[0002]

2. Description of the Related Art Referring to FIG. 5, in a conventional duplex system having a cluster configuration having an active device 1, a standby device 2, and a shared disk device 4, a failure occurs in the active device 1 performing a processing operation. Alternatively, if a failure occurs and the client 3 determines that the active device 1 cannot be processed, the processing task performed in the active device 1 is taken over by the standby device 2 and continued in the standby device 2. It is supposed to be.

[0003] In Japanese Patent Application Laid-Open No. 4-342333, a loop interface device and a loop monitoring unit are provided in a local area network having a loop-type transmission line, and a standby system is diagnosed by an operator inputting a test command. Diagnostic methods have been proposed.

However, in such a conventional system, when a processing task is taken over due to the occurrence of a failure or a failure in the active device 1, the processing task executed in the active device 1 is successfully performed by the standby device 2. May not be handed over, causing a decrease in the reliability of the system.

[0005] The case where the processing work is not successfully taken over is, for example, when some kind of failure or failure has occurred in the standby system 2 to take over the work, the client 3
Cannot be connected to the standby system 2 from the server, or the work executed in the active system 1 is insufficient for the storage capacity of the local memory 27 or the local disk 28 of the standby system 2
This is a case where the standby system 2 cannot execute the processing task after the processing is taken over due to a problem such as the inability of the standby system 2 to access the shared disk 4.

[0006]

In the conventional diagnostic system as described above, the operator only instructs the execution of a test to use the bypass mode to check whether or not a fault has occurred in the standby system in the system. Yes, no determination has been made regarding the availability of handover of the processing task during the operation of the active device. For this reason, in the conventional method, the reliability as to whether resources sufficient to execute processing tasks are secured at the time of processing transition of the standby system device has not been guaranteed.

[0007] For example, in the case where a job is executed using two central processing units (hereinafter referred to as CPUs), 30 MB of memory, and 4 GB of disk in the active system device 1, the standby system device 2 3 CPUs, 10 memory
It is assumed that 0 MB and 10 GB of disk can be used, respectively. Here, if one CPU and one
Even if the 0 MB memory fails, the standby system 2 has two CPUs, 90 MB of memory and 10 disks of disk.
Since each of the GBs can be used, the processing operation can be transferred from the active device 1. However, assuming that two CPUs and an 8 GB disk have failed in the standby device 2, the standby device 2 has one CPU and 10 memories.
Only 0 MB and 2 GB of the disk can be used, and the processing operation cannot be transferred from the active device 1.

[0008] An object of the present invention is to solve the above problem by providing a monitoring unit in each processing device, and constantly monitoring the resource utilization information of the active system and the resource status of the standby system. The purpose of the present invention is to guarantee the availability of the standby device according to the requirements.

[0009]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, a standby monitoring device in a cluster system according to the present invention comprises an operating device and a standby device connected to a network.
In a cluster system configured by connecting
In the system, the operation system device and the standby system device
Each has a monitoring unit that monitors faults and resources,
The active monitoring unit detects a failure in the active device and
Detects resource usage and increases the detected resource usage
The standby monitoring unit is notified, and the standby monitoring unit
Detect failures in equipment and detect available resources
The operation system device notified by the operation system monitoring unit.
Resource usage of the standby system detected from the resource usage of the
When the available amount is small, it is determined that an abnormality has occurred.
And features.

[0010]

Further, each of the active system device and the standby system device includes a monitoring line for monitoring, and monitors faults and resources and notifies when an abnormality is detected via the monitoring line.

[0012] Further, the operating device and the standby device are connected to each other by a communication line.

The monitoring section of the active system notifies the monitoring section of the standby system via the communication line of the detected use amount of the resource in its own device. If it is determined that the available resource amount of the standby device is smaller than the resource usage amount of the active device notified via the line, it is determined that a failure has occurred.

Further, the monitoring units of the active device and the standby device monitor each other via the communication line.

The resources are a CPU, a local memory and a local disk.

Further, when the monitoring units of the active system device and the standby system device detect a failure in the LAN control circuit of the own system, the monitoring unit reports the failure to the client through the other system.

Further, in a cluster system in which the active device and the standby device each include a monitoring unit for monitoring a fault and a resource, and are connected by a network, in a cluster system which is operating while the active device is operating, Of the standby device and the resource usage of the standby device, and the fault detection and the available resource of the standby device are detected. Is small, it is determined that a failure has occurred in the standby system device.

[0018]

Embodiments of the present invention will now be described in detail with reference to the drawings.

Referring to FIG. 1, an embodiment of a cluster system according to the present invention is an operation system 1 for executing a processing task.
And a standby device 2 that takes over the processing task when the active device 1 fails, an active device 1 and a standby device 2
, A client 3 for managing the cluster system, and a local area network (hereinafter, referred to as “LA”) connecting the active device 1, the standby device 2, and the client 3.
N ". ) 5 and a communication path 6 for connecting the active system device 1 and the standby system device 2 to each other.

The operation system device 1 includes a first LAN control circuit 11 connected to the LAN 5, a path control circuit 12 connected to the communication path 6, and a disk control device connected to the shared disk device 4 via a cable. And a circuit 13. Further, the operation system device 1 periodically checks the status of the LAN control circuit 11, the path control circuit 12, and the disk control circuit 13,
The usage and status of resources such as the usage of the PU 16, the local memory 17 and the local disk 18 are monitored, and during normal operation, the resource usage is sent to the standby system 2, and the failure in the operation system 1 is monitored. A monitoring unit 14 is provided for notifying the client 3 or the standby apparatus 2 of the detection at the time of detection. The monitoring unit 14 is connected to the LAN control circuit 11, the path control circuit 12, the disk control circuit 13, the CPU 16, the local memory 17, and the local disk 18 via a monitoring path 15 for monitoring.

The standby system 2 has the same configuration as the active system 1 and has a LAN control circuit 21, a path control circuit 22, and a disk control circuit 23.
The state of the N control circuit 11, the path control circuit 12, and the disk control circuit 13 and the available amount and state of resources such as the CPU 26, the local memory 27, and the local disk 28 are monitored to detect a failure or resource shortage. The system device 1 includes a monitoring unit 24 for notifying the fact and a monitoring path 25 for monitoring.

Normally, the client 3 issues a processing request to the active device 1 via the LAN 5 and causes the active device 1 to execute a processing task.

The operating system device 1 performs a processing operation in response to a processing request from the client 3, and at this time, the monitoring unit 14 transmits the LAN control circuit 11 via the monitoring path 15 to the LAN control circuit 11.
The statuses of the path control circuit 12 and the disk control circuit 13 and the usage of the CPU 16, the local memory 17, and the local disk 18 are monitored, respectively, and the active CPU is monitored.
16, local memory 17 and local disk 18
Monitoring unit 2 of the standby system 2 via the communication path 6
Notify 4. Here, if the monitoring unit 14 detects a failure, the monitoring unit 14 notifies the client 3 to that effect.

The monitoring unit 24 of the standby system device 2 monitors each unit in the standby system device 2 while a job is being executed in the active system device 1. That is, it is determined whether the connection between the client 3 and the standby device 2, the transfer of the processing task from the active device 1 to the standby device 2, and the access from the standby device 2 to the shared disk 4 are possible. To do
The monitoring of the LAN control circuit 21, the path control circuit 22, the disk control circuit 23, the CPU 26, the local memory 27, and the local disk 28 is performed via the monitoring path 25. At the same time, the monitoring unit 24 operates the CPU 16, the memory 17, and the disk 18 of the active device 1 sent from the monitoring unit 14 of the active device 1 via the communication path 6.
Is compared with the available capacity of the CPU 26, the memory 27, and the disk 28 of the standby device 2, and the processing operation being executed in the active device 1 can be executed even if it is taken over by the standby device 2. It is determined whether or not an appropriate resource is secured in the standby system device 2. Then, upon detecting a failure or resource shortage, the monitoring unit 24 notifies the client 3 to that effect. As described above, when the usable amount of the resources in the standby device 2 is larger than the used amount of the resources in the active device 1, the work being processed by the active device 1 can be taken over by the standby device 2. The standby device 2 is monitored according to the operating status of the active device 1 such that the standby device 2 is determined.

Next, the operation of the embodiment of the present invention will be described.

Referring to FIGS. 1 and 2, the monitoring unit 14 in the active device 1 first checks the status of resources such as the CPU 16, the local memory 17 and the local disk 18 in the active device 1 through the monitoring path 15. In addition to monitoring the usage status and the status of the LAN control circuit 11, the path control circuit 12, and the disk control circuit 13, the monitoring unit 24 of the standby system 2 is also monitored (step S1 in FIG. 2). Then, as a result of this monitoring, when the occurrence of an abnormality inside the active device 1 or the abnormality notification from the monitoring unit 24 of the standby device 2 is detected, the monitoring unit 14 notifies the client 3 of the occurrence of the abnormality (step S21). on the other hand,
If no abnormality is detected, the resource utilization information of the active device 1 is transmitted to the monitoring unit 24 of the standby device 2 through the communication path 6 (Step S3).

On the other hand, the monitoring unit 24 of the standby device 2 transmits the status and available amount of resources such as the CPU 26, the local memory 27 and the local disk 28 in the standby device 2 through the monitoring path 25 in the standby device 2. , LAN control circuit 21, path control circuit 22, and disk control circuit 2
3 and the state of the monitoring unit 14 of the active device 2 is monitored (step S11). Then, as a result of this monitoring, when the occurrence of an abnormality inside the standby system device 2 or the notification of the abnormality from the monitoring unit 14 of the operation system device 1 is detected, the client 3 is notified (step S2).
1) If no notification of an abnormality is detected, a comparison is made between the resource usage of the active device 1 transmitted via the communication path 6 and the available capacity of the resources of the standby device 2 (step S13). . Here, even one of the resources such as the CPU, the local memory, and the local disk is stored in the active system device 1.
When it is determined that the resource usage of the corresponding resource is insufficient, the client 3 is notified of the fact (step S14). On the other hand, if the available resources of the resources in the standby device 2 are all larger than the available resources of the active device 1, the process ends without any problem. The client 3 is the operating system device 1
Alternatively, when the occurrence of a problem such as a failure is notified from the standby device 2, a measure such as notifying the operator is taken.

Next, the case where an abnormality is detected in the above will be described in detail.

First, consider the case where the monitoring unit 24 of the standby apparatus 2 detects an abnormality. The monitoring unit 24 first determines the location where the abnormality has occurred (step S21), and determines the location of the abnormality with the path control circuit 22 and the disk control circuit 2 indicated by hatching in FIG.
3. If any one of the CPU 26, the local memory 27, the local disk 28, and the monitoring unit 14 of the operation apparatus 1 or the LAN 5, or notifies the client 3 via the LAN control circuit 21 and the LAN 5 of the occurrence of an abnormality. Is performed (step S22). On the other hand, as illustrated in FIG. 4, when an abnormality of the LAN control circuit 21 is detected, the monitoring unit 24 transmits the operation system device via the path control circuit 22, the communication path 6, and the path control circuit 12. The first monitoring unit 14 is notified of the occurrence of an abnormality in the LAN control circuit 21 (step S23). Then, the monitoring unit 14 having received the notification of the occurrence of the abnormality notifies the client 3 via the LAN control circuit 11 and the LAN 5 of the LAN control circuit 2 of the standby system 2.
A notification of the occurrence of abnormality 1 is made (step S24).

Similarly, when an abnormality other than the LAN control circuit 11 is detected by the monitoring unit 14 of the active device 1,
Client 3 via AN control circuit 11 and LAN 5
Is notified of the occurrence of an error. On the other hand, when an abnormality is detected in the LAN control circuit 11, the standby system device 2 is transmitted via the path control circuit 12, the communication path 6, and the LAN control circuit 22.
Is notified of the occurrence of an abnormality in the LAN control circuit 11. Then, the monitoring unit 24 of the standby apparatus 2 receiving this notification notifies the client 3 of the occurrence of the abnormality of the LAN control circuit 11 via the LAN control circuit 21 and the LAN 5.

The monitoring unit 14 of the active device 1 and the standby device 2
Monitors the status of each other via the communication path 6, so that a failure of the monitoring unit on the other side can be detected, and normal monitoring can be performed due to the failure of the monitoring unit. It can be prevented that the notification is not performed or a failure notification cannot be issued.

[0032]

As described above, according to the present invention, in a cluster system having an active device and a standby device, a CPU, a local memory, and a local disk which can execute processing tasks being executed in the active device are used. By providing a monitoring unit that monitors whether the capacity and the like can be secured by the standby system device and whether there is any abnormality in each control device for accessing the LAN and the shared disk device, the standby unit according to the operation status of the active system is provided. It is possible to determine the availability of the system, and it is possible to achieve high reliability as a cluster system that can take over and execute processing tasks without any problem to the standby system device when an abnormality occurs in the active system device. .

[Brief description of the drawings]

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

FIG. 2 is a flowchart illustrating an operation of the exemplary embodiment of the present invention.

FIG. 3 is a diagram showing a first state of the embodiment of the present invention.

FIG. 4 is a diagram showing a second state of the embodiment of the present invention.

FIG. 5 is a diagram showing a configuration of a conventional technique.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 active system device 2 standby system device 3 client 4 shared disk 5 LAN 6 communication path 7, 8 cable 11, 21 LAN control circuit 12, 22, path control circuit 13, 23 disk control circuit 14, 24 monitoring unit 15, 25 for monitoring Path 16, 26 CPU 17, 27 Local memory 18, 28 Local disk

Claims (8)

(57) [Claims] In a monitoring system in a cluster system in which an active device and a standby device are connected by a network , each of the active device and the standby device has a fault.
A monitoring unit that monitors harm and resources, wherein the active monitoring unit detects a failure in the active device and
Detects resource usage and increases the detected resource usage
Notifying the standby monitoring unit, the standby monitoring unit detects a failure in the standby device and
Detects the amount of resources that can be used.
Detected from the resource usage of the operating device notified from the
When the available resources of the standby system
A monitoring method in a cluster system, wherein it is determined that normal has occurred . 2. The system according to claim 1, wherein each of the active device and the standby device includes a monitoring line for monitoring, and performs monitoring of a failure and a resource and notification of abnormality detection via the monitoring line. The monitoring method in the cluster system according to claim 1 . 3. The cluster system includes a communication line that connects the active device and the standby device, and a usage amount of the resource detected by the monitoring unit of the active system is set via the communication line. The monitoring method according to claim 2, wherein the monitoring method is notified to a monitoring unit of the system. 4. The monitoring method according to claim 3 , wherein the monitoring units of the active device and the standby device monitor each other via the communication line. Wherein said resource, CPU, claim 1, characterized in that it comprises a local memory and local disk
Or a monitoring method in the cluster system described in 2 or 4 . 6. The monitoring unit of the active device and the standby device detects a failure in its own LAN control circuit.
Monitoring method in a cluster system according to claim 1, wherein the reporting a failure to the client through the other system. 7. A monitoring method in a cluster system in which an active device and a standby device each include a monitoring unit for monitoring a failure and a resource, and wherein the active device and the standby device are connected by a network, Detecting a failure in the active device and detecting the amount of resources used; detecting a failure in the standby device and detecting the amount of available resources; and ,
Comparing the available amount of resources of the standby device; and, if the available amount of the standby device is smaller than the available amount of resources of the active device, the standby device fails. Determining that an error has occurred. 8. An active device and a standby device, wherein each of the active device and the standby device includes a monitoring unit that monitors a failure and a resource, and the monitoring unit of the active device includes: The failure detection in the active device and the detection of the resource usage are performed. The monitoring unit of the standby device detects the failure in the standby device and detects the available amount of the resource. A cluster system which determines that an abnormality has occurred when the available amount of resources of the standby system device is smaller than the used amount of the standby system device.