CN103369054B - A kind of acquisition tasks management method and system - Google Patents

Abstract

Embodiments of the present invention provide a collection task management method and system. The method includes the following steps: the scheduling module loads the collection scheme, and reads the target node list from the collection scheme; the scheduling module generates a management task list according to the target node list; the scheduling module According to the preset balance mechanism, the collection tasks in the management task list are allocated to the collection nodes; the system includes a scheduling module, which is used to load the collection scheme, and read the target node list from the collection scheme; generate management tasks according to the target node list list; according to the preset balance mechanism, the collection tasks in the management task list are allocated to the collection nodes.

Description

A collection task management method and system

technical field

The invention relates to the field of communication technology, in particular to a collection task management method and system.

Background technique

A server cluster refers to a large-scale server that gathers many sub-servers together to perform the same service. From the perspective of the client, a server cluster is equivalent to a single server. Obviously, this kind of server cluster integrating multiple sub-servers has a higher computing speed than a single server.

The server cluster can be used as a large-scale data acquisition system. The sub-servers in the server cluster operate cooperatively. Each sub-server acts as a collection node and executes the corresponding collection tasks, which can realize parallel processing of multiple collection tasks, which can significantly improve The efficiency of data collection. Under this premise, reasonable task management and deployment for each collection node becomes the key to the balanced operation of the parent system.

The existing load balancer BIG-IP is a device that manages and distributes task traffic in the cluster data collection system. BIG-IP provides flexible algorithms to efficiently forward data streams to the clustered data acquisition systems to which it is connected. Facing the front-end operator, BIG-IP and the data acquisition system it manages can be equivalent to a virtual server, so that the operator does not need to operate on each acquisition node.

There are two management and inspection tools for collection nodes in BIG-IP, namely ECV (ExtendedContentVerification, extended content verification) and EAV (ExtendedApplicationVerification, extended application verification). Both can check the collection nodes and feed back information according to the request, judge the operation status of the collection nodes and the overall data collection system through the feedback information, and then manage and allocate collection tasks. The difference is that ECV responds to client's inspection requests, while EAV responds to application-specific inspection requests.

However, the defect of the two inspection tools in BIG-IP is that although they reflect the basic idea of managing and distributing collection tasks to a certain extent, in fact, the topology of collection nodes is complicated and dynamic in real time. Changing; the existing technology is still not able to achieve the task flow and load balance of the data collection system as a whole as the collection progresses, nor can it effectively perform task management and collection node maintenance of the data collection system.

Contents of the invention

In view of this, the purpose of the present invention is to provide a collection task management method and system, the method realizes the management and resource allocation of collection tasks and collection nodes during the process of data collection system executing collection tasks.

To achieve the above object, the present invention has the following technical solutions:

A collection task management method, said method comprising the following steps:

The scheduling module loads the collection plan, and reads the list of target nodes from the collection plan;

The scheduling module generates a management task list according to the target node list;

The scheduling module assigns the collection tasks in the management task list to the collection nodes according to the preset balance mechanism.

The scheduling module also reads the collection node list from the collection scheme; then the generation of the management task list according to the target node list is specifically:

Use the collection node list to extract the task list;

Compare the target node list with the ongoing task list, and use the part of the target node list that has changed compared with the ongoing task list as the management task list.

The task list extraction by using the collection node list is specifically as follows:

The scheduling module monitors the communication of the collection nodes in the collection node list, obtains the collection tasks being executed by each collection node, and gathers the collection tasks being executed into a task list.

The part of the target node list that is compared with the task list that changes is used as the management task list is specifically:

When the collection task exists in the target node list and does not exist in the task list, the collection task will be added to the management task list;

When the collection task does not exist in the target node list and exists in the task list, the collection task will be deleted from the management task list;

When the collection task exists in both the target node list and the progress task list, but any parameter changes, the collection task is regarded as a modification item in the management task list, and the changed parameters are subject to the parameters in the target node list.

The preset equalization mechanism is specifically:

According to the number of collection tasks in the target node list and the number of collection nodes, calculate the value interval of the load task of each collection node;

Then the allocation of the collection tasks in the management task list to the collection nodes according to the preset balancing mechanism is specifically:

Assign the collection tasks in the management task list to the collection nodes, so that the number of collection tasks assigned to each collection node is within the value range of the load task.

The method also includes:

Set a flag bit for the management task list, if a changed collection task is recorded in the management task list, define the flag bit of the management task list as true;

After the collection tasks in the management task list are allocated to the collection nodes according to the preset balancing mechanism, the flag bit of the management task list is changed to false.

The method also includes:

The scheduling module sends test instructions to the collection nodes in the collection node list, causing the collection nodes to test themselves according to the test instructions; the scheduling module receives feedback information from the collection nodes, and modifies the collection node list according to the feedback information.

The modification of the collection node list according to the feedback information is specifically:

If the feedback information of the collection node shows that the test fails, the collection node is deleted from the collection node list, and the collection node is added to the dead node list.

The method also includes:

The dispatching module sends at least one test instruction to the collection nodes in the dead node list, and when the feedback information of the collection node shows that the test is successful at least once, the collection node is added to the collection node list.

The method also includes:

When adding a collection node, add the newly added collection node to the list of collection nodes.

A collection task management system, the system specifically includes:

The scheduling module is used to load the collection plan, read the target node list from the collection plan; generate a management task list according to the target node list; assign the collection tasks in the management task list to the collection nodes according to the preset balancing mechanism.

The scheduling module includes:

The reading unit is used to load the collection plan, and read the target node list and the collection node list from the collection plan;

The processing unit is used to extract the task list by using the collection node list;

The management unit is used to compare the target node list with the task list, and use the part of the target node list that has changed compared with the task list as the management task list;

The allocation unit is configured to allocate the collection tasks in the management task list to the collection nodes according to a preset balancing mechanism.

The scheduling module also includes:

The test unit is used to send test instructions to the collection nodes in the collection node list, causing the collection nodes to test themselves according to the test instructions; the scheduling module receives feedback information from the collection nodes, and modifies the collection node list according to the feedback information.

It can be seen from the above technical solutions that the beneficial effects of the present invention are: in the case of dynamic changes in the collection strategy, a corresponding management task list can be generated, and the distribution of collection tasks in the collection nodes can be further modified according to the balancing mechanism, so as to realize the load on the collection nodes balance; by comparing the target node list with the task list, only the changed acquisition tasks are written into the management task list, saving communication resources and system resources; by setting the flag bit for the management task list, it is also possible to avoid processing Duplication of the management task list has an impact; send test instructions to the collection node for testing, understand the status of the collection node in real time, modify the list of collection nodes when the collection node changes, and realize dynamic management of the collection node for easy collection More accurate management and balanced distribution of tasks.

Description of drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are For some embodiments of the present invention, those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is the flow chart of the method described in the embodiment of the present invention;

Fig. 2 is a flow chart of the method described in another embodiment of the present invention;

Fig. 3 is a schematic structural diagram of the system according to the embodiment of the present invention.

detailed description

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

In the present invention, the method is co-implemented while the cluster data collection system (which may be referred to as the parent system of the collection task management method and system in the present invention) executes the collection task. The existing cluster data acquisition system includes multiple acquisition nodes that collect data from the target node, and the system first needs to formulate acquisition strategies for all acquisition nodes and target nodes, and the acquisition nodes perform data acquisition according to the acquisition strategy formulated by the parent system .

The collection strategy stipulates that the collection node collects target data for a certain target node at a certain time, that is, specifies the collection tasks for the target node; the collection strategy contains one or more collection tasks to be assigned to different collection nodes Above; the acquisition strategy is the basis for the acquisition node to perform tasks. It should be noted that, in the parent system, collection tasks are generally performed periodically, that is, according to the collection tasks, the collection nodes collect data from the target nodes in a specific cycle in a continuous cycle. In the overall solution of the present invention, the acquisition policy content is not changed.

The acquisition strategy can be reflected through the list of target nodes. The target node list includes information such as the name and address of each target node; the target node list may also include information on the manner in which each target node is collected, such as a collection period. This means that the target nodes included in the target node list are all the nodes that need to collect data in the collection strategy, that is, the target of the collection task execution; so each target node actually corresponds to a collection task in the collection strategy , the target node list also reflects the sum of collection tasks in the collection strategy, which is equivalent to the collection strategy.

In order to allocate collection tasks during the collection process and control the collection nodes to perform collection tasks in a targeted manner, it is also necessary to configure the collection node list. The collection node list and the target node list together constitute a complete collection scheme in the data collection process.

The collection node list includes information such as the name, address, operation status, and distribution status of each collection node in the parent system. That is, the collection node list not only describes which collection nodes in the parent system participate in the execution of the collection scheme, but also shows whether the collection nodes are assigned and are executing collection tasks.

The parent system operates according to the above-mentioned collection scheme, and realizes controlling the collection node to collect data from the target node. In the present invention, the management of collection tasks is realized on the premise of operating on the above two lists. The above application background is applicable to the following embodiments.

Referring to FIG. 1 , it is a specific embodiment of the acquisition task management method described in the embodiment of the present invention. In this embodiment, the above collection node list and target node list will be combined to generate a management task list, and then according to the management task list, effective management of collection tasks in a dynamically changing environment will be realized, so that each collection node can run in a balanced and stable manner. The method described in this embodiment specifically includes the following steps:

Step 101, the scheduling module loads the collection scheme, and reads the list of target nodes from the collection scheme.

It can be further known from the implementation background of the method that the target node list means a collection strategy, and the collection strategy includes at least one collection task. In this embodiment, the scheduling module can load the acquisition scheme from the database provided in the parent system.

The above steps are the essential preparatory steps for the implementation of the method. By loading the collection scheme, the scheduling module obtains the list of target nodes, and can further obtain the list of collection nodes, so that the prerequisite for subsequent steps is met. It should also be noted that the scheduling module is the part in charge of control in the method described in this embodiment.

In addition, the scheduling module can also be connected to the monitoring port, and use the monitoring port to directly receive relevant instructions from the mother system.

Step 102, the scheduling module generates a management task list according to the target node list.

In this embodiment, the management of the collection node load is generally based on a collection task as a unit, and the management of the operation and load status of the collection nodes is realized by distributing collection tasks in a balanced manner. The data presentation form of a collection task (task) in the system can refer to the following table:

parameter name meaning server_id target node ID4 --> server_type target node type raid_type Target node disk type server_ip target node ip address bmc_ip The ip address of the target node bmc bmc_user The username of the target node bmc bmc_password The password of the target node bmc ipmi_period Target node ipmi collection cycle snmp_period target node snmp collection cycle

Table 1

In the collection scheme, tasks are assigned to the collection nodes in advance. However, with the progress of data collection, the overall collection scheme will change dynamically, and the collection tasks and collection nodes of each collection node may change; for example, the parent system reconfigures the collection strategy, and the collection tasks in the original collection strategy It may be modified or deleted, and new collection tasks may be added to the collection strategy; or some collection nodes may crash, or new collection nodes may be added to the system. Therefore, during the collection process, the task load of each collection node will be unbalanced to varying degrees. It is difficult for existing technologies to adapt to similar dynamic changes. The purpose of the task management method described in this embodiment is to improve the unbalanced task load of collection nodes by dynamically changing the original distribution of collection tasks.

In this embodiment, the collection task management method does not change the collection strategy, that is, it does not modify the content of the target node list. However, the parent system can re-edit the collection strategy through instructions, and then modify the target node list according to the new collection strategy.

As mentioned above, it has been clarified that the content in the target node list is essentially the same as the content in the collection strategy. Each target node corresponds to a collection task. If the parent system changes the collection strategy, this change will also be reflected in the target node list. Therefore, the target node list can reflect dynamic changes. In this embodiment, the target node list includes 6 collection tasks, and the following table shows the data form of the node list:

Table 2

Table 2 shows the six tasks in the target node Table Task1~Task6. The ID and other information of the target node corresponding to each task are also correspondingly displayed. For details, please refer to Table 1, which will not be repeated here.

Because the target node list represents a collection of collection tasks, in theory, the target node list can be directly used as a management task list for subsequent balanced distribution of tasks. But in fact, it is impossible to see the change of the collection task only based on the target node list itself. If the target node list is directly used as the management task list to assign tasks, it is equivalent to re-executing the task for all collection nodes. Allocation; but in fact, the change of the target node list is generally not particularly large, so a complete re-allocation wastes a lot of communication resources and system resources. In this embodiment, the following preferred solutions can be adopted to obtain the management task list:

If the load of the collection nodes can be adjusted only for the part of the target node list that changes, resources will be greatly saved. In fact, in the system, changes in the collection strategy cannot be directly communicated to the collection nodes, and the collection nodes only perform tasks according to the original assignment. Therefore, by comparing the list of target nodes and the collection tasks currently executed by the collection nodes, changes in the list of target points can be found.

In this embodiment, the task list is extracted according to the collection node list, that is, the scheduling module monitors the communication of the collection nodes in the collection node list, directly obtains the collection tasks currently being executed by each collection node by monitoring the communication content, and sends all the collection tasks that will be executed The task set is combined into a task list. In terms of implementation, the scheduling module can use the listening port to communicate and control the collection nodes.

Assume that in this embodiment, there are two collection nodes, namely the first collection node and the second collection node. By monitoring the communication of the two collection nodes, two task orders can be obtained correspondingly, as follows:

First task list:

table 3

The first in-progress task list indicates that the first collection node is executing two tasks, Task1 and Task7.

The second task list:

Table 4

Similarly, the second execution task list indicates that the second collection node is executing Task3, Task4 and Task5, and the information of the corresponding target nodes is also displayed in a prescribed format.

Combining the above-mentioned first task list and the second task list can obtain a task list that can display the task execution and distribution of all collection nodes in the collection node list, that is, know which specific tasks each node is performing.

However, the task list shows which collection tasks each collection node needs to perform after the previous task assignment to the collection nodes. If the list of target nodes has changed, but the task distribution of the collection nodes has not changed accordingly, it is very likely that the load of the collection nodes will be unbalanced.

In this embodiment, the target node list is compared with the task list, and the part of the target node list that has changed from the task list is used as the management task list. The management task list can be generated according to the following mechanism:

When a collection task exists in the target node list and does not exist in the task list, the collection task will be added to the management task list;

When a collection task does not exist in the target node list and exists in the task list, the collection task will be deleted from the management task list;

When a collection task exists in the target node list and the task list at the same time, but any parameter is changed, the collection task is regarded as a modification item in the management task list, and the changed parameters are subject to the parameters in the target node list .

In this embodiment, comparing the list of target nodes and the task list, it can be found that Task2 and Task6 belong to added items; Task7 belongs to deleted items; Task1, Task3 and Task4 all belong to modified items (the parameter snmp_period has changed). Task5 has not changed. Thus, the management task list obtained through comparison avoids repeated communication and processing of Task5 and saves system resources. It should be noted that the total number of collection tasks and the number of unchanged collection tasks in this embodiment are relatively small, and it is often quite considerable to save this part of resources in a real large-scale parent system.

According to the above method, the specific data form of the management task list is as follows:

add Task2, Task6 mod Task1, Task3, Task4 delete Task7

table 5

Step 103, the scheduling module assigns the collection tasks in the management task list to the collection nodes according to the preset balancing mechanism.

In this embodiment, the preset equalization mechanism can be expressed as follows:

According to the number of collection tasks and the number of collection nodes in the target node list, the value interval of the load task of each collection node is calculated; and the number of collection tasks assigned to each collection node is within the value interval of the load task.

In this embodiment, the target node list includes 6 acquisition tasks, and the number of acquisition nodes is 2, so the average number of acquisition tasks that each acquisition node should load is 3; is less, so you can directly set the value interval of the load task to be equal to 3; in other environments, there may be a large number of collection tasks and collection nodes, and the average value may appear in decimals, so the value interval of the load task can be appropriately relaxed accordingly, so that It is no longer a single numerical value; if the value range of the load task is relaxed in this embodiment, the value range of the load task can be set as 2-4. Of course, according to different collection tasks and the number levels of collection nodes, the range of the load task value range can also be appropriately changed as needed. When the number of tasks executed by all collection nodes falls within the range of load task values, it is considered that the load of all collection nodes is balanced.

In this embodiment, when the collection tasks are assigned according to the management task list, the collection tasks that have not changed will generally not be reassigned, and for the collection tasks that have changed, the modified items are often given priority; that is, for example, this implementation In the example, the acquisition tasks such as Task1, Task3, or Task4 that only change parameters will not change the original acquisition nodes that execute the acquisition tasks as much as possible, thereby saving the workload of the system.

For example, for the 6 collection tasks and 2 collection nodes in this embodiment, they can be allocated in the following manner, and the value interval of the load task is equal to 3.

For the first collection node, Task1 and Task7 are originally executed; Task1 is a modified item, which is replaced based on the parameters in the target node list; Task7 is a deleted item, which is directly deleted. At this time, only one collection task is assigned to the first collection node, which is smaller than the value interval of the load task, so the two added items Task2 and Task6 are assigned to the first collection node for execution. After redistribution, the first collection node will execute three collection tasks Task1, Task2 and Task6 to meet the value interval of the load task.

The second collection node originally executes three collection tasks of Task3, Task4 and Task5, among which Task3 and Task4 are modification items; Task3 and Task4 can be replaced according to the parameters in the target node list, and Task5 remains unchanged. The second collection node executes three collection tasks, and the number meets the value range of the load task.

Therefore, in this embodiment, the load tasks of the two collection nodes are three, achieving load balance.

In this embodiment, a flag can also be set for the management task list, and the flag can have two positions of true and false. Whether it is necessary to redistribute the acquisition tasks according to the balance mechanism needs to be determined according to the flag of the management task list. judge. Specifically, when the flag bit is true, it is determined that reallocation is required, otherwise, reallocation is not required.

If in step 102, the target node list is not equal to the task list, that is, when the management task list records a collection task with a change between the two, define the flag bit of the management task list as true. When the flag bit is true, the process of reallocating collection tasks according to the balance mechanism in this step is started. After the allocation process is over, the flag bit of the management task list needs to be changed to false. If the flag bit is false, this step is not triggered, so that the processed management task list can be prevented from affecting task allocation again.

From the above technical solutions, it can be seen that the beneficial effect of the collection task management method described in this embodiment is: the method can generate a management task list correspondingly when the collection strategy changes dynamically, and further modify the collection task according to the balance mechanism. The distribution of collection tasks in the nodes realizes the balance of the load on the collection nodes; by comparing the list of target nodes with the task list, only the changed collection tasks are written into the management task list, which saves communication resources and system resources; Setting the flag bit can also prevent the repeated influence of the processed management task list.

The acquisition task management method described in the foregoing embodiments realizes the load balancing of the acquisition nodes during the acquisition process, that is to say, the impact of the dynamic change of the acquisition strategy is avoided in a targeted manner. In order to further address the dynamic change of the collection node state, the following optimization scheme can also be combined into the overall technical scheme of the method, as shown in Figure 2 for details:

Step 201, the scheduling module sends test instructions to the collection nodes in the collection node list, causing the collection nodes to test themselves according to the test instructions; the scheduling module receives feedback information from the collection nodes, and modifies the collection node list according to the feedback information.

As mentioned above, in the process of executing the collection strategy of the parent system, not only the collection strategy may change, but also the state of the collection node may change, and the collection node executing the collection task may fail or crash; so if the normal execution of the parent system is to be maintained It is very important to test the collection nodes in real time, understand the status of each collection node and make management actions.

In this embodiment, the scheduling module can periodically send test instructions to the collection nodes in the collection node list, and the collection nodes can perform tests according to the test instructions and feed back information to the scheduling module. The scheduling module thus learns the status of each collection node. Generally speaking, the collection nodes that are running normally can complete the test according to the test instructions, and the feedback information will show that the test is successful; but if a collection node is in a dead state, the test will not be completed, and a test failure will be returned. Feedback.

For example, in the specific scenario of the previous embodiment, the scheduling module sends test instructions to the first collection node and the second collection node at the same time; assuming that the first collection node is working normally, its feedback information shows that the test is successful, and the second collection node crashes , the feedback message is test failure.

Further, after receiving the feedback information, the scheduling module will make corresponding management according to the feedback information, that is, modify the collection node list according to the feedback information. The collection nodes included in the collection node list are all collection nodes that are executing tasks, and once a dead collection node appears, it needs to be deleted from the collection node list. Of course, if all the collection nodes work normally, the list of collection nodes will not be modified here.

In this embodiment, the second collection node is deleted from the list of collection nodes and added to the list of dead nodes.

In addition, in order to improve the accuracy of the collection node test, this embodiment will further include:

Step 202 , the scheduling module sends at least one test instruction to the collection nodes in the dead node list, and when the feedback information from the collection nodes indicates that the test is successful at least once, the collection node is added to the collection node list.

The acquisition nodes added to the dead node list are all acquisition nodes whose feedback information is a test failure in a test. However, since the test has only been performed once, it cannot be ruled out that the test fails due to accidental factors; or the collection nodes that may enter the dead node list have returned to normal operation after maintenance, and the collection task can be performed again. Therefore, in this embodiment, repeated testing is performed on the collection nodes in the dead node list, that is, the test instruction is repeatedly sent at least once, preferably three times in this embodiment. When the feedback information of at least one of the three test instructions is that the test is successful, it is considered that the collection node can normally perform the collection task, and the collection node is re-added to the collection node list.

In this embodiment, for the case where the parent system adds a collection node during the execution of the collection strategy, the following steps can be followed:

Step 203, when adding a collection node, add the newly added collection node into the list of collection nodes.

In either case, after changing the collection node list, it is necessary to reset the task list according to the new collection node list as described in step 102 in the previous embodiment, and then further reset the management task list, and finally The collection tasks of the collection nodes are redistributed, so as to achieve load balance again under the condition of dynamic changes of the collection nodes.

The beneficial effects of the method described in this embodiment are: the dispatching module sends test instructions to the collection nodes for testing, the state of the collection nodes is understood in real time, and the list of collection nodes is modified when the collection nodes change, realizing dynamic pairing The collection nodes are managed to facilitate more accurate management and balanced distribution of collection tasks.

Corresponding to the system described in the embodiment shown in Figure 1, the present invention further discloses a collection task management system, the system is established based on the method, and the core technical solutions of the two are consistent, and the system specifically includes:

The scheduling module is used to load the collection plan, read the target node list from the collection plan; generate a management task list according to the target node list; assign the collection tasks in the management task list to the collection nodes according to the preset balancing mechanism.

As shown in Figure 2, the scheduling module includes:

The reading unit is used to load the collection plan, and read the target node list and the collection node list from the collection plan;

The processing unit is used to extract the task list by using the collection node list;

The management unit is used to compare the target node list with the task list, and use the part of the target node list that has changed compared with the task list as the management task list;

The allocation unit is configured to allocate the collection tasks in the management task list to the collection nodes according to a preset balancing mechanism.

From the above technical solutions, it can be seen that the beneficial effect of the system is: the system can generate a management task list correspondingly when the collection strategy changes dynamically, and further modify the distribution of collection tasks in the collection nodes according to the balance mechanism, so as to realize Load balancing on the collection nodes; by comparing the target node list with the task list, only the changed collection tasks are written into the management task list, saving communication resources and system resources.

In addition, the following preferred solutions can also be combined in the system: the scheduling module also includes:

The test unit is used to send test instructions to the collection nodes in the collection node list, causing the collection nodes to test themselves according to the test instructions; the scheduling module receives feedback information from the collection nodes, and modifies the collection node list according to the feedback information.

The beneficial effects of the above optimal solution are: send test instructions to the collection node for testing, understand the state of the collection node in real time, modify the list of collection nodes when the collection node changes, and realize dynamic management of the collection node. For more accurate management and balanced distribution of collection tasks.

The above is only a preferred embodiment of the present invention, it should be pointed out that, for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications can also be made. It should be regarded as the protection scope of the present invention.

Claims (11)

1. a collection task management method, is characterized in that, described method comprises the following steps: The scheduling module loads the collection plan, and reads the list of target nodes from the collection plan; The scheduling module generates a management task list according to the target node list; The scheduling module assigns the collection tasks in the management task list to the collection nodes according to the preset balance mechanism; Wherein, the scheduling module also reads the collection node list from the collection plan; then the generation of the management task list according to the target node list is specifically: Use the collection node list to extract the task list; Compare the target node list with the ongoing task list, and use the part of the target node list that has changed compared with the ongoing task list as the management task list. 2. The method according to claim 1, characterized in that, the extraction of the task list by using the collection node list is specifically: The scheduling module monitors the communication of the collection nodes in the collection node list, obtains the collection tasks being executed by each collection node, and gathers the collection tasks being executed into a task list. 3. The method according to claim 1, characterized in that, the part of the target node list that is compared with the changed task list as the management task list is specifically: When the collection task exists in the target node list and does not exist in the task list, the collection task will be added to the management task list; When the collection task does not exist in the target node list and exists in the task list, the collection task will be deleted from the management task list; When the collection task exists in both the target node list and the progress task list, but any parameter changes, the collection task is regarded as a modification item in the management task list, and the changed parameters are subject to the parameters in the target node list. 4. The method according to claim 1, wherein the preset equalization mechanism is specifically: According to the number of collection tasks in the target node list and the number of collection nodes, calculate the value interval of the load task of each collection node; Then the allocation of the collection tasks in the management task list to the collection nodes according to the preset balancing mechanism is as follows: Assign the collection tasks in the management task list to the collection nodes, so that the number of collection tasks assigned to each collection node is within the value range of the load task. 5. according to the method described in any one of claim 1-4, it is characterized in that, described method also comprises: Set a flag bit for the management task list, if a changed collection task is recorded in the management task list, define the flag bit of the management task list as true; After the collection tasks in the management task list are allocated to the collection nodes according to the preset balancing mechanism, the flag bit of the management task list is changed to false. 6. The method according to claim 1, wherein the method further comprises: The scheduling module sends test instructions to the collection nodes in the collection node list, causing the collection nodes to test themselves according to the test instructions; the scheduling module receives feedback information from the collection nodes, and modifies the collection node list according to the feedback information. 7. The method according to claim 6, wherein the modification of the collection node list according to the feedback information is specifically: If the feedback information of the collection node shows that the test fails, the collection node is deleted from the collection node list, and the collection node is added to the dead node list. 8. The method according to claim 7, wherein the method further comprises: The dispatching module sends at least one test instruction to the collection nodes in the dead node list, and when the feedback information of the collection node shows that the test is successful at least once, the collection node is added to the collection node list. 9. The method according to any one of claims 6-8, wherein the method further comprises: When adding a collection node, add the newly added collection node to the list of collection nodes. 10. A collection task management system, characterized in that the system specifically includes: The scheduling module is used to load the collection plan, read the target node list from the collection plan; generate a management task list according to the target node list; assign the collection tasks in the management task list to the collection nodes according to the preset balancing mechanism; The scheduling module includes: The reading unit is used to load the collection plan, and read the target node list and the collection node list from the collection plan; The processing unit is used to extract the task list by using the collection node list; The management unit is used to compare the target node list with the task list, and use the part of the target node list that has changed compared with the task list as the management task list; The allocation unit is configured to allocate the collection tasks in the management task list to the collection nodes according to a preset balancing mechanism. 11. The system according to claim 10, wherein the scheduling module further comprises: The test unit is used to send test instructions to the collection nodes in the collection node list, causing the collection nodes to test themselves according to the test instructions; the scheduling module receives feedback information from the collection nodes, and modifies the collection node list according to the feedback information.