CN103326897A - Distributed computing environment general monitoring device and failure detection method - Google Patents

Abstract

A general monitoring system for a distributed computing environment, including a connectivity monitoring system, which includes a connectivity monitoring module set on a client, a connectivity response module set on a server, and a connection between the connectivity monitoring module and the connectivity response The connectivity monitoring channel of the module, and the connectivity monitoring system is used to detect whether the network interconnection environment or the server is connected; the service availability monitoring system, the service availability monitoring system includes a heartbeat monitoring module arranged on the client, The heartbeat response module set on the server and the heartbeat monitoring channel connecting the heartbeat monitoring module and the heartbeat response module, the service effectiveness monitoring system is used to detect whether the server fails. Avoid the inefficiency and hysteresis of manual intervention and fault judgment, give full play to the ability of large-scale equipment in the centralized central office, improve its availability, and protect the investment benefits of operators.

Description

A general monitoring device and failure detection method for a distributed computing environment

【Technical field】

The invention relates to a general monitoring software device and a failure detection method in a distributed computing environment, which are applicable to unattended command response processing and batch command automatic processing.

【Background technique】

For the client-server software system based on network communication, including the distributed client-service system based on the Internet, and the plug-in type main control-line card device based on the main board of the chassis, all have the network distributed characteristics of client and service interaction, For example, in the field of passive optical network (PON) broadband access, a large number of frame plug-in devices are used, especially under the trend of flattening data communication networks and access networks, the In order to carry more user access, the configuration and distribution of services need to be realized by issuing commands and batch commands from the client to the server. During the process of device activation and restart, batch command execution needs to be completed under unattended conditions. Automatically restore the previously saved configuration. These processes are closely related to user experience and directly affect the quality of service. On the other hand, the network management needs to obtain the running status of the device and collect device performance data to achieve fault management. It also sends commands to the device. way achieved. Frame plug-in equipment needs to carry more users, and its availability is required to be as high as 99.99999%. Therefore, the key components of the equipment need to have active and standby redundancy, and the slots supporting service boards also need to be increased to support more Access users. The frame plug-in device is usually composed of a main control board, multiple line cards and auxiliary boards. The former requires active and standby redundancy, and the relationship between the main control board and the service board (line card) is the relationship between the client and the server. , the chassis backplane usually provides a large number of slots, and the boards are interconnected through the network, which is a typical distributed computing command execution environment.

For a client and server system in a distributed computing environment, the client sends a command to the server to obtain the server's response. When the interconnection environment, network failure or server failure will affect the process of command request and response, especially when automatic configuration During the batch execution of commands, the entire processing transaction will be affected due to the obstruction of intermediate commands. Therefore, a system framework and processing method are needed to achieve fault tolerance and ensure a smooth processing flow.

Figure 1 shows the command request-response process in the prior art. For various command interaction situations, including batch command processing, there will be a large difference in the delay of the command response processing process, and the command response may be caused by the failure of the service module. The processing process is infinitely extended without a response result, or the response result cannot be delivered due to environmental failures, resulting in unlimited waiting for the command result reception process, which has a serious impact on the client module command interaction, especially the automatic batch command processing. If the timeout threshold is artificially set during the waiting process of command result reception, and the longest delay in the command response processing process shall prevail, it is difficult to determine due to the variability of the computing environment of the service module, and it may also seriously affect the execution of batch commands due to the failure of the service module . From another point of view, due to the large difference in the delay in the command processing process and the command response processing process, the waiting time of the command result receiving waiting process needs to be preset according to the longest time. In addition, the network communication channel may fail, or the service module may fail. All will cause the client module not to receive the response. If you wait for the longest time, you may enter blind waiting because it is difficult to judge the fault: if the waiting time is set too short, the client module will ignore the effective response of the service module; if the waiting time is set too long, the validity and availability of the service module Hysteresis, which can result in reduced system command processing efficiency.

【Content of invention】

The purpose of the present invention is to provide a general-purpose monitoring software device and failure detection method for distributed computing environments, which can avoid blind waiting, avoid hysteresis in judging validity and availability, and improve system command processing efficiency.

The present invention provides a general monitoring software device for distributed computing environment, which is characterized in that it includes:

A connectivity monitoring system, the connectivity monitoring system comprising a connectivity monitoring module set on the client, a connectivity response module set on the server, and a connectivity monitoring channel connecting the connectivity monitoring module and the connectivity response module, And the connectivity monitoring system is used to detect whether the network interconnection environment or the server is connected;

A service effectiveness monitoring system, the service effectiveness monitoring system includes a heartbeat monitoring module set on the client, a heartbeat response module set on the server, and a heartbeat monitoring channel connecting the heartbeat monitoring module and the heartbeat response module, the service The availability monitoring system is used to detect whether the server is invalid.

On the basis of the above technical solution, the connectivity monitoring module periodically sends a connectivity detection command, and if the connectivity response module does not respond for several cycles, it is judged that the network interconnection environment or the server is not connected.

On the basis of the above technical solution, the heartbeat monitoring module initiates a validity detection command, and if the heartbeat response module does not respond within a certain period of time, it is judged that the server is invalid.

On the basis of the above-mentioned technical solution, the general-purpose monitoring software device for the distributed computing environment is also provided with a chassis interconnection mechanism.

On the basis of the above technical solution, the chassis interconnection mechanism is a chassis backplane for inserting server boards, and the chassis backplane is provided with detection elements for detecting whether the service boards are connected.

The present invention also provides a failure detection method using the general monitoring software device of the distributed computing environment, and the failure detection method includes the following two tasks:

Task 1: The connectivity monitoring system periodically monitors whether the network interconnection environment or the server is connected, and the service effectiveness monitoring system periodically monitors whether the server is invalid;

Task two:

A: The client issues a command, waits for a time slice for receiving the command response, checks whether it receives the server response, and continues to execute the command if it receives the server response, and enters step B if it does not receive the server response;

B: Detect whether the connectivity response module responds within the corresponding cycle, if the connectivity response module responds, then enter step C, if the connectivity monitoring system does not respond, the command times out;

C: Detect whether the heartbeat response module responds. If the heartbeat response module responds, return to step A. If the heartbeat response module does not respond, the command times out.

On the basis of the above technical solution, the task 1 and task 2 are executed concurrently.

On the basis of the above technical solution, when the command times out, the client suspends or transfers to other redundant services.

Compared with the prior art, the present invention detects whether the network interconnection environment or the server is connected through the connectivity monitoring system, and detects whether the server is invalid through the service effectiveness monitoring system, so that unattended network interconnection and chassis equipment can be Large batches of configuration commands and device status reading commands can be executed better, and rapid fault location and supporting processing greatly improve the efficiency of command execution. Avoid the inefficiency and hysteresis of manual intervention and fault judgment, give full play to the ability of large-scale equipment in the centralized central office, improve its availability, and protect the investment benefits of operators.

【Description of drawings】

FIG. 1 is a schematic diagram of a command request-response process in the prior art;

Fig. 2 is a system architecture and a functional block diagram of the present invention;

Fig. 3 is a flowchart of the failure detection method of the present invention.

【Detailed ways】

Please refer to FIG. 2 , which is a diagram of the system architecture and functional modules of the present invention. The client 103 and the server 104 are connected through a command channel 111 . The general monitoring software device in the distributed computing environment includes a connectivity monitoring system and a service availability monitoring system, and the connectivity monitoring system includes a connectivity monitoring module 105 arranged on the client, a connectivity response module 106 arranged on the server, and Connecting the connectivity monitoring module 105 and the connectivity monitoring channel 112 of the connectivity answering module 106, the connectivity monitoring system is used to detect whether the network interconnection environment or the server is connected. The service availability monitoring system includes a heartbeat monitoring module 101 arranged on the client, a heartbeat response module 102 arranged on the server, and a heartbeat monitoring channel 110 connecting the heartbeat monitoring module 101 and the heartbeat response module 102. The service availability monitoring system is used for Check if the server is down.

The connectivity monitoring module 105 periodically sends a connectivity detection command, and if the connectivity response module 106 does not respond for several cycles, it is determined that the network interconnection environment or the server 104 is not connected. At this time, a fault is reported and the command is stopped, or the client 103 turns to other redundant services.

The heartbeat monitoring module 101 initiates a validity detection command. If the heartbeat response module 102 does not respond within a certain period of time, it is judged that the server 104 is invalid and the command is stopped, or the client 103 switches to other redundant services.

When the chassis interconnection mechanism 108 exists, the detection element is provided on the chassis interconnection mechanism 108 . The detecting element detects whether the frame interconnection mechanism 108 is connected correctly. In this embodiment, the chassis interconnection mechanism 108 is a chassis backplane, which provides slots for the server boards 104 . The detecting element detects whether the board card of the server 104 is installed. If the board of the server 104 is not installed, it is judged that the server 104 is invalid.

The following introduces the method for failure detection by using the general monitoring software device detection of the distributed computing environment of the present invention. The detection method includes the following two concurrently executed tasks:

Task 1: The connectivity monitoring system periodically monitors whether the network interconnection environment 107 or the server 104 is connected, and the service availability monitoring system periodically monitors whether the server 104 is invalid. The connectivity monitoring module 105 in the connectivity monitoring system periodically issues a connectivity detection command, and if the connectivity response module 106 does not respond for several cycles, it is determined that the network interconnection environment or the server 104 is not connected.

Task two:

A: The client 103 accepts to wait for a time slice after issuing the command, and checks whether it receives a response from the server 104. If it receives a response from the server 104, it continues to execute the command. If it does not receive a response from the server 104, it proceeds to step B;

B: detect whether the connectivity response module 106 has a response in the corresponding cycle, if the connectivity response module 106 has a response, then enter step C, if the connectivity response module 106 has no response, then judge that the command is overtime;

C: Detect whether the heartbeat response module 102 responds. If the heartbeat response module 102 responds, return to step A and continue to wait for the next time slice. If the heartbeat response module 102 does not respond, the command times out. When the command times out, an error alarm signal can be issued to suspend the command processing, or transfer to other redundant services to enter the process of sending-receiving commands according to the preset.

The general monitoring software device and failure detection method of the distributed computing environment of the present invention can quickly and effectively detect relevant failures and failures in the process of command execution, and avoid resource waste caused by long time-out waiting when connectivity or services fail, And blind maintenance work that cannot distinguish the cause of the failure. Rapid fault location and supporting processing greatly improve the efficiency of command execution. Avoid the inefficiency and hysteresis of manual intervention and fault judgment, give full play to the ability of large-scale equipment in the centralized central office, improve its availability, and protect the investment benefits of operators.

Claims (8)

1. A general monitoring system for a distributed computing environment, characterized in that: comprising, A connectivity monitoring system, which includes a connectivity monitoring module set on the client, a connectivity response module set on the server, and a connectivity monitoring channel connecting the connectivity monitoring module and the connectivity response module, and the connectivity The monitoring system is used to detect whether the network interconnection environment or the server is connected; A service validity monitoring system, the service validity monitoring system includes a heartbeat monitoring module set on the client, a heartbeat response module set on the server, and a heartbeat monitoring channel connecting the heartbeat monitoring module and the heartbeat response module, the service validity The performance monitoring system is used to detect whether the server fails. the 2. a kind of general monitoring software device of distributed computing environment as claimed in claim 1, it is characterized in that: described connectivity monitoring module periodically sends connectivity detection order, if described connectivity response module does not respond in multiple cycles , it is judged that the network interconnection environment or the server is not connected. the 3. a kind of general monitoring software device of distributed computing environment as claimed in claim 1, it is characterized in that: described heartbeat monitoring module initiates validity detection command, if described heartbeat response module does not respond within preset time, then Judgment server failure. the 4. A general-purpose monitoring software device for a distributed computing environment as claimed in claim 1, characterized in that: said general-purpose monitoring software device for a distributed computing environment is also provided with a chassis interconnection mechanism. the 5. A kind of distributed computing environment universal monitoring software device as claimed in claim 4, it is characterized in that: the machine frame interconnection mechanism is the machine frame backplane for inserting the server board, and the machine frame The backplane is provided with a detection element for detecting whether the service board is connected. the 6. A failure detection method using the distributed computing environment general monitoring software device described in any one of claims 1-4, characterized in that: the failure detection method comprises the following two tasks: Task 1: The connectivity monitoring system periodically monitors whether the network interconnection environment or the server is connected, and the service effectiveness monitoring system periodically monitors whether the server is invalid; Task two: A. The client issues a command, and waits for a time slice to accept the command, and checks whether it receives a server response. If it receives it, it continues to execute the command; if it does not receive it, it enters B; B. Detect whether the connectivity response module has a response in the corresponding period, if the connectivity response module has a response, then enter step C, if the connectivity response module does not respond, the command times out; C. Detect whether the heartbeat response module responds. If the heartbeat response module responds, return to step A. If the heartbeat response module does not respond, the command times out. the 7. The failure detection method according to claim 6, wherein said task one and said task two are executed concurrently. the 8. The failure detection method according to claim 6, wherein when the command times out, the client transfers to other redundant services or suspends. the

Images