Method for expanding distributed database

Abstract

The application provides a distributed database expansion method, which comprises limiting network IO and disk IO resources in a distributed database expansion process by using a resource management group tool or a flow limiting algorithm. In the distributed database expansion method, the greatest influence is the network IO and disk IO used in the expansion. By means of resource allocation on the processes of the network IO and disk IO used in the expansion through the resource management group tool or flow rate limitation on the processes of the network IO and disk IO used in the expansion through the resource management group tool, the network IO and disk IO resources in the distributed database expansion process can be effectively limited, and thus the normal business of the users of the distributed database can be avoided from being affected in the distributed database expansion process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of database, in particular to a distributed database expansion method. BACKGROUND

[0002] With the development of the information age, more and more data needs to be stored, and the non-shared distributed database is used more and more in storing a large amount of data. With the growth of data size, the previous node planning cannot keep up with the changes, so node expansion is also more and more common. For some businesses that must be online all the time, in the process of database expansion, users require that the database expansion should have little impact on existing business, and should not occupy a large amount of resources, otherwise it may cause normal business to be unable to respond in time, affecting customer experience. Therefore, it is very important to control the expansion resource usage. The current mainstream way of expansion resource control is to use resource group to control, but only supports instance level and library level control, and there is no resource limitation function for controlling the entire process level. In the process of expanding the non-shared distributed database cluster, the network IO and disk IO used in the expansion process have the greatest impact on the normal business of the distributed database. How to limit the resources occupied by the network IO and disk IO used in the expansion is a problem that needs to be solved at present. SUMMARY

[0003] In view of the above problems, the present application is proposed in order to provide a distributed database expansion method which overcomes the above problems or at least partially solves the above problems, and can solve the problem of excessive resource occupation of network IO and / or disk IO used in the process of expanding the distributed database, thereby avoiding the impact on the normal business of the users of the distributed database.

[0004] Specifically, the present application provides a distributed database expansion method, comprising:

[0005] Limiting the resources of network IO and disk IO in the process of expanding the distributed database by using a resource management group tool and / or a flow limiting algorithm.

[0006] Optionally, the distributed database expansion method further comprises:

[0007] Executing an expansion tool to perform expansion;

[0008] The step of limiting the resources of network IO and disk IO in the process of expanding the distributed database by using the flow limiting algorithm comprises:

[0009] In the process of executing the expansion tool, the flow limiting rate is obtained and the algorithm flow limiting mode is started;

[0010] The algorithm flow limiting mode has the flow limiting algorithm, and the algorithm flow limiting mode is configured to limit the flow rate of the data read / write and / or the data sending required by the distributed database expansion process according to the flow limiting algorithm and the flow limiting rate, so as to limit the resources of the network IO and the disk IO in the distributed database expansion process.

[0011] Optionally, the flow limiting rate is obtained and the algorithm flow limiting mode is started in the process of executing the expansion tool, and the algorithm flow limiting mode comprises:

[0012] New nodes are added, and the expansion tool is controlled to initialize the new nodes;

[0013] The expansion tool is connected with a database server master node;

[0014] The expansion tool issues an expansion data redistribution command and a command for configuring a flow limiting rate parameter to the database server master node;

[0015] After the database server master node receives the expansion data redistribution command and the command for configuring the flow limiting rate parameter, the flow limiting rate is configured,

[0016] The database server master node performs data redistribution for expansion while starting the algorithm flow limiting mode.

[0017] Optionally, the database server master node performs data redistribution for expansion while starting the algorithm flow limiting mode, and the algorithm flow limiting mode comprises:

[0018] The database server master node sends the expansion data redistribution command and the flow limiting rate to lower data nodes of the database server;

[0019] The lower data nodes of the database server perform expansion;

[0020] In the expansion process of the lower data nodes of the database server, the flow limiting algorithm of the flow limiting rate is used through function calls of the data read / write or the data sending;

[0021] The flow limiting algorithm is run by using the flow limiting rate to limit the flow rate of the data read / write and / or the data sending required by the distributed database expansion process.

[0022] Optionally, the flow limiting algorithm is run by using the flow limiting rate to limit the flow rate of the data read / write and / or the data sending required by the distributed database expansion process, and the flow limiting algorithm comprises:

[0023] The flow of the data read / write and / or the data sending is obtained;

[0024] comparing the flow of the data read / write and / or the data transmission with the flow corresponding to the throttling rate;

[0025] if the flow of the data read / write and / or the data transmission is less than the flow corresponding to the throttling rate, the data read / write and / or the data transmission is not throttled;

[0026] if the flow of the data read / write and / or the data transmission is greater than or equal to the flow corresponding to the throttling rate, process waiting is performed to throttle the data read / write and / or the data transmission, so that the maximum flow of the data read / write and / or the data transmission is the flow corresponding to the throttling rate.

[0027] Optionally, the method for expanding the distributed database further comprises:

[0028] statistically counting the process waiting time, wherein the process waiting time comprises:

[0029] recording the time of entering the throttling of the data read / write and / or the data transmission;

[0030] calculating the running time of the flow of the data read / write and / or the data transmission running according to the flow corresponding to the throttling rate;

[0031] determining the process waiting time according to the current time, the last time of entering the throttling of the data read / write and / or the data transmission, and the running time.

[0032] Optionally, it is judged whether the distributed database can support the use of the resource management group tool to throttle the network IO and the disk IO in the expansion process of the distributed database;

[0033] if the distributed database cannot support the use of the resource management group tool to control the network IO and the disk IO in the expansion process of the distributed database, the throttling algorithm is used to throttle the network IO and the disk IO in the expansion process of the distributed database;

[0034] if the distributed database can support the use of the resource management group tool to throttle the network IO and the disk IO in the expansion process of the distributed database, the resource management group tool is used to throttle the network IO and the disk IO in the expansion process of the distributed database.

[0035] Optionally, the use of the resource management group tool to control the network IO and the disk IO in the expansion process of the distributed database comprises:

[0036] creating an expansion control group by the resource management group tool and setting the limit of the expansion control group;

[0037] The expansion control group limits the resources of network IO and disk IO in the expansion process according to the limit;

[0038] Optionally, the creating of the expansion control group and the setting of the limit of the expansion control group by the resource management group tool comprises:

[0039] The resource manager is invoked on the database server master node by the resource management group tool, and a command of creating the expansion control group and setting the limit of the expansion control group is issued to the lower data nodes of the database server;

[0040] The resource manager is invoked on the database server master node by the resource management group tool, and a command of creating the expansion control group and setting the limit of the expansion control group is issued to the lower data nodes of the database server;

[0041] Optionally, the expansion method of the distributed database further comprises:

[0042] The expansion tool is executed to perform the expansion;

[0043] The expansion control group limits the resources of network IO and disk IO in the expansion process according to the limit;

[0044] The new node is added, and the expansion tool is controlled to initialize the new node;

[0045] The expansion tool is connected with the database server master node;

[0046] The expansion data redistribution command is issued to the database server master node by the expansion tool;

[0047] The database server master node receives and executes the expansion data redistribution command to perform the expansion, obtains the process for expansion, adds the process for expansion to the expansion control group, and the expansion control group limits the resources of network IO and disk IO in the process for expansion according to the limit.

[0048] Optionally, the expansion method of the distributed database further comprises:

[0049] The limit of the expansion control group is obtained at any time, and the limit of the expansion control group is updated, so that the expansion control group limits the resources of network IO and disk IO in the expansion process according to the updated limit.

[0050] In the distributed database expansion method, the network IO and the disk IO used in the distributed database expansion process have the greatest impact. Through the resource management group tool, the process of the network IO and the disk IO used in the expansion is allocated resources or the process of the network IO and the disk IO used in the expansion is flow rate limited through the resource management group tool, so that the resources of the network IO and the disk IO in the distributed database expansion process can be effectively limited, and then in the distributed database expansion process, the normal business of the user of the distributed database is avoided from being affected.

[0051] The above and other objects, advantages and features of the present application will become more apparent from the following detailed description of specific embodiments thereof, taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0052] Some specific embodiments of the present application will be described in detail in the following with reference to the drawings, which are given by way of illustration only and thus are not restrictive of the present application. Like reference numerals refer to like elements or parts throughout the description of the drawings. It should be understood that the drawings are not necessarily to scale. In the drawings:

[0053] Figure 1 is a schematic flow chart of a distributed database expansion method according to an embodiment of the present application;

[0054] Figure 2 is a schematic flow chart of a distributed database expansion method according to another embodiment of the present application;

[0055] Figure 3 is a schematic flow chart of a distributed database expansion method according to another embodiment of the present application. DETAILED DESCRIPTION

[0056] A distributed database system usually uses small computer systems, each of which can be placed in a place alone, and each of which can have a complete copy or a partial copy of a DBMS and has its own local database. Many computers located in different places are connected to each other through a network to form a complete, global, logically centralized and physically distributed large database.

[0057] With the development of information age, more and more data need to be stored, and the non-shared distributed database is used more and more in storing a large amount of data, and as the data scale grows, the previous node planning cannot keep up with the changes, so the node expansion is also more and more common. For some businesses that must be online all the time, in the process of database expansion, users require that the database expansion should have little impact on the existing business, and cannot occupy a large amount of resources, otherwise it may cause normal business to be unable to respond in time, affecting customer experience. Therefore, it is very important to control the resource usage of expansion. The mainstream way of expansion resource control is to use resource groups to control, but only instance level and library level are supported, and there is no resource limit function for the whole process level in the expansion process. In the process of expanding the non-shared distributed database cluster, the network IO and disk IO used in the expansion process have the greatest impact on the normal business of the distributed database, and how to limit the resource occupation of the network IO and disk IO used in the expansion is a problem that needs to be solved urgently. In view of the above problems, the present application is proposed to provide a distributed database expansion method which can overcome the above problems or at least partially solve the above problems, and can solve the problem of excessive resource occupation of network IO and disk IO used in the distributed database expansion process, thereby avoiding the impact of normal business of users of the distributed database.

[0058] Figure 1 is a schematic flow chart of the distributed database expansion method according to an embodiment of the present application. As shown in Figure 1 , and referring to Figures 2 to 3 , the embodiment of the present application provides a distributed database expansion method, which comprises limiting the network IO and disk IO resources in the distributed database expansion process by using a resource management group tool or a flow limiting algorithm.

[0059] In the distributed database expansion process, the network IO and disk IO used in the expansion have the greatest impact, and by using the resource management group tool to separately allocate resources to the processes of network IO and disk IO used in the expansion or by using the flow limiting algorithm to limit the flow rate of the processes of network IO and disk IO used in the expansion, the resources of network IO and disk IO in the distributed database expansion process can be effectively limited, and then in the distributed database expansion process, the normal business of users of the distributed database can be avoided from being affected, and the user satisfaction can be improved.

[0060] In some embodiments of the present application, the method for expanding the distributed database further comprises executing the expansion tool to perform the expansion. The throttling algorithm is used to limit the network IO and disk IO resources in the expansion process of the distributed database. In the process of executing the expansion tool, the throttling rate is obtained and the algorithm throttling limit mode is started. The algorithm throttling limit mode has a throttling algorithm, and the algorithm throttling limit mode is configured to limit the flow rate of data reading and writing and / or data sending required in the expansion process of the distributed database according to the throttling algorithm and the throttling rate, so as to limit the network IO and disk IO resources in the expansion process of the distributed database.

[0061] By expanding the tool to expand and obtaining the throttling rate in the expansion process, the throttling rate parameter can be used to limit the flow rate of data reading and writing and / or data sending in the throttling algorithm. The data reading and writing and / or data sending with large flow rate will be limited, so as to reduce the influence of these data reading and writing and / or data sending with large flow rate on the normal business of the user using the distributed data.

[0062] In some embodiments of the present application, in the process of executing the expansion tool, the throttling rate is obtained and the algorithm throttling limit mode is started, which includes adding a new node and controlling the expansion tool to initialize the new node. The initialized new node only has metadata. The connection between the expansion tool and the database server master node is established. The expansion data redistribution command and the command for configuring the algorithm throttling rate parameter are sent to the database server master node through the expansion tool. After the database server master node receives the expansion data redistribution command and the command for configuring the algorithm throttling rate parameter, the throttling rate is configured. The data is redistributed through the database server master node to perform the expansion, and the algorithm throttling limit mode is started.

[0063] In some embodiments, establishing the connection between the expansion tool and the database server master node includes connecting the database server master node by the expansion tool, and sending the connection information to the database server master node. The database server master node receives the connection information and establishes the connection with the expansion tool.

[0064] Further, in some embodiments of the present application, the data redistribution is performed by the database server master node for scaling out, while the algorithmic throttling restriction mode is started, including the database server master node sending the scaling out data redistribution command and the throttling rate to the lower level data nodes of the database server. The lower level data nodes of the database server perform scaling out. In the scaling out process of the lower level data nodes of the database server, the throttling algorithm is used by function calls of data reading and writing or data sending with the throttling rate. The throttling algorithm is run with the throttling rate to limit the flow rate of data reading and writing and / or data sending required in the distributed database scaling out process. That is, the algorithmic throttling restriction mode finally throttles the flow rate of data reading and writing and data sending required in the scaling out process to limit the network IO and disk IO resources in the distributed database scaling out process.

[0065] Further, in some embodiments of the present application, as shown in Figure 2 the throttling algorithm is run with the throttling rate to limit the flow rate of data reading and writing and / or data sending required in the distributed database scaling out process, including: obtaining the flow of data reading and writing and / or data sending. Comparing the flow of data reading and writing and / or data sending with the flow corresponding to the throttling rate. If the flow of data reading and writing and / or data sending is less than the flow corresponding to the throttling rate, the data reading and writing and / or data sending is not limited. If the flow of data reading and writing and / or data sending is greater than or equal to the flow corresponding to the throttling rate, the process is waiting to limit the data reading and writing and / or data sending, so that the maximum flow of data reading and writing and / or data sending is the flow corresponding to the throttling rate.

[0066] The obtaining of the flow of data reading and writing and / or data sending includes counting the flow of data reading and writing and / or data sending by a counter. In some embodiments, the comparison of the flow of data reading and writing and / or data sending with the flow corresponding to the throttling rate can be the comparison of the flow of data reading and writing and / or data sending with the product of the throttling rate and the process time of data reading and writing and / or data sending. That is, if the flow of data reading and writing and / or data sending is greater than the product of the throttling rate and the process time of data reading and writing and / or data sending, it is considered that the flow of data reading and writing and / or data sending is too large, and if it continues, it will occupy too many resources and affect the normal business of users of the distributed database, so the process of data reading and writing and / or data sending can only wait, thereby playing a limiting role.

[0067] In some embodiments of the present application, as Figure 2As shown, the distributed database expansion method further comprises counting the process waiting time. The process waiting time comprises recording the time of entering the limited data read-write and / or data sending. The running time of the flow of data read-write and / or data sending running according to the flow corresponding to the flow control rate is calculated. The process waiting time is determined according to the current time, the time of the last entering the limited data read-write and / or data sending, and the running time. That is, the process waiting time is equal to the data sending time minus the time of the last entering the limited data read-write and / or data sending. By using this method to control the process waiting time, the process of data read-write and / or data sending can be more accurately controlled, the process waiting time process can be avoided, and the expansion progress of the distributed database can be delayed.

[0068] In some embodiments of the present application, the resource management group tool is used to control the network IO and disk IO resources in the distributed database expansion process, which comprises creating an expansion control group by the resource management group tool and setting the limit of the expansion control group. The expansion control group limits the network IO and disk IO resources in the expansion process according to the limit.

[0069] Further, in some embodiments of the present application, as shown, Figure 3 The creation of the expansion control group and the setting of the limit of the expansion control group by the resource management group tool comprises invoking the resource manager on the database server master node by the resource management group tool, and issuing the command of creating the expansion control group and setting the limit of the expansion control group to the lower data nodes of the database server. The lower data nodes of the database server execute local operations by invoking the resource manager on the lower data nodes through the resource management group tool to create the expansion control group and set the limit of the expansion control group.

[0070] In some embodiments of the present application, before invoking the resource manager on the database server master node by the resource management group tool, the distributed database expansion method further comprises adding the resource management group tool. The resource management group tool is a tool for managing the control group, which is used to create, set, view, and destroy the control group.

[0071] Further, in some embodiments of the present application, as shown, Figure 3As shown, the distributed database expansion method further comprises executing an expansion tool to perform expansion. The expansion control group limits the resources of network IO and disk IO in the expansion process according to the quota, including adding new nodes and controlling the expansion tool to initialize the new nodes. The expansion tool is connected with the database server master node. The expansion tool sends a data redistribution command to the database server master node. The database server master node receives the data redistribution command and performs expansion to obtain expansion processes, and adds the expansion processes to the expansion control group. The expansion control group limits the resources of network IO and disk IO in the expansion processes according to the quota.

[0072] That is, the database server master node sends the data redistribution command to the lower data nodes of the database server, and the lower data nodes of the database server perform expansion. The expansion processes are obtained and added to the expansion control group. The expansion control group limits the resources of network IO and disk IO in the expansion processes according to the quota. Adding the expansion processes to the expansion control group includes adding the expansion processes, corresponding child processes and processes used during expansion of the lower data nodes to the expansion control group, and the expansion control group automatically limits the resources of the expansion processes, corresponding child processes and processes used during expansion of the lower data nodes.

[0073] In some embodiments of the present application, the distributed database expansion method further comprises acquiring the quota of the expansion control group at any time, and updating the quota of the expansion control group, so that the expansion control group limits the resources of network IO and disk IO in the expansion process according to the updated quota. That is, the quota of the expansion control group is updated according to the business volume of the user. When the business volume of the user using the distributed database is large, the quota of the expansion control group is reduced. When the business volume of the user using the distributed database is small, the quota of the expansion control group is increased. For example, during working hours on weekdays, the business volume of the user using the distributed database is relatively large, so the quota of the expansion control group can be reduced. During the remaining time, as the business volume of the user using the distributed database decreases, the quota of the expansion control group can be increased. Thus, the expansion of the distributed database is completed as soon as possible while minimizing the impact on the business volume of the user of the database.

[0074] In some embodiments of the present application, the method for expanding the distributed database further comprises judging whether the distributed database can support the use of the resource management group tool to limit the network IO and disk IO resources in the distributed database expansion process. If the distributed database cannot support the use of the resource management group tool to control the network IO and disk IO resources in the distributed database expansion process, a flow limiting algorithm is used to limit the network IO and disk IO resources in the distributed database expansion process. If the distributed database can support the use of the resource management group tool to limit the network IO and disk IO resources in the distributed database expansion process, the resource management group tool is used to limit the network IO and disk IO resources in the distributed database expansion process.

[0075] The embodiment of the present application further provides a machine readable storage medium, which stores a machine executable program, and the machine executable program is executed by a processor to implement the method in any of the above embodiments.

[0076] The flowchart provided by the embodiment is not intended to indicate that the operations of the method should be performed in any specific order, or that all of the operations included in the method should be included in every case. In addition, the method can include additional operations. Additional changes can be made to the above method within the scope of the technical ideas provided by the embodiment method.

[0077] At this point, those skilled in the art should recognize that, although the present application has been fully illustrated and described herein with a plurality of exemplary embodiments, many other variations or modifications in accordance with the principles of the present application can be directly determined or deduced from the disclosure of the present application without departing from the spirit and scope of the present application. Therefore, the scope of the present application should be understood and recognized as covering all such other variations or modifications.

Claims

1. A method for scaling a distributed database, the method comprising: The method comprises: controlling network IO and disk IO resources in the distributed database expansion process by using a resource management group tool and / or a flow control algorithm; The method further comprises: judging whether the distributed database can support the use of the resource management group tool to control network IO and disk IO resources in the distributed database expansion process; if the distributed database cannot support the use of the resource management group tool to control network IO and disk IO resources in the distributed database expansion process, using the flow control algorithm to control network IO and disk IO resources in the distributed database expansion process; if the distributed database can support the use of the resource management group tool to control network IO and disk IO resources in the distributed database expansion process, using the resource management group tool to control network IO and disk IO resources in the distributed database expansion process; The method further comprises: executing an expansion tool to perform expansion; The method of using the flow control algorithm to control network IO and disk IO resources in the distributed database expansion process comprises: acquiring a flow control rate and starting an algorithm flow control mode during execution of the expansion tool; The algorithm flow control mode has the flow control algorithm, and is configured to control the flow rate of data reading and writing and / or data sending required by the distributed database expansion process according to the flow control algorithm and the flow control rate, so as to control network IO and disk IO resources in the distributed database expansion process. The method of using the resource management group tool to control network IO and disk IO resources in the distributed database expansion process comprises: creating an expansion control group by using the resource management group tool and setting a limit for the expansion control group; The expansion control group limits network IO and disk IO resources in the expansion process according to the limit.

2. The distributed database expansion method of claim 1, wherein The method of acquiring a flow control rate and starting an algorithm flow control mode during execution of the expansion tool comprises: adding a new node and controlling the expansion tool to initialize the new node; establishing a connection between the expansion tool and a database server master node; sending an expansion data redistribution command and a command for configuring an algorithm flow control rate parameter to the database server master node by using the expansion tool; configuring the flow control rate after the database server master node receives the expansion data redistribution command and the command for configuring an algorithm flow control rate parameter; performing data redistribution by using the database server master node to expand the database, and starting an algorithm flow control mode.

3. The distributed database expansion method of claim 2, wherein The method of performing data redistribution by using the database server master node to expand the database, and starting an algorithm flow control mode comprises: sending the expansion data redistribution command and the flow control rate to lower-level data nodes of the database server by using the database server master node; expanding the database by using the lower-level data nodes of the database server. In the expansion process of the lower data nodes of the database server, the throttling algorithm is run by using the throttling rate through the function call of data reading and writing or data sending; The throttling algorithm is run by using the throttling rate to limit the flow rate of data reading and writing and / or data sending required in the expansion process of the distributed database.

4. The expansion method of the distributed database according to claim 3, characterized in that, the throttling algorithm is run by using the throttling rate to limit the flow rate of data reading and writing and / or data sending required in the expansion process of the distributed database, comprising: obtaining the flow of data reading and writing and / or data sending; comparing the flow of data reading and writing and / or data sending with the flow corresponding to the throttling rate; if the flow of data reading and writing and / or data sending is less than the flow corresponding to the throttling rate, not limiting the data reading and writing and / or data sending; if the flow of data reading and writing and / or data sending is greater than or equal to the flow corresponding to the throttling rate, process waiting is performed to limit the data reading and writing and / or data sending, so that the maximum flow of the data reading and writing and / or data sending is the flow corresponding to the throttling rate.

5. The method of claim 4, wherein, Further comprising: statistically obtaining the process waiting time, which comprises: recording the time of entering the limitation of data reading and writing and / or data sending; calculating the running time of the flow of data reading and writing and / or data sending running according to the flow corresponding to the throttling rate; determining the process waiting time according to the current time, the last time of entering the limitation of data reading and writing and / or data sending, and the running time.

6. The expansion method of the distributed database according to claim 1, characterized in that, the creation of the expansion control group and the setting of the limit of the expansion control group by using the resource management group tool comprises: calling the resource manager on the database server master node by using the resource management group tool, and issuing the command of creating the expansion control group and setting the limit of the expansion control group to the lower data nodes of the database server; the lower data nodes of the database server call the resource manager on the lower data nodes by using the resource management group tool to perform local operations, create the expansion control group, and set the limit of the expansion control group.

7. The method of claim 1, wherein, Further comprising: performing the expansion tool to perform expansion; the resource limit of the network IO and the disk IO in the expansion process by the expansion control group according to the limit comprises: adding new nodes and controlling the expansion tool to initialize the new nodes; establishing the connection between the expansion tool and the database server master node; issuing the expansion data redistribution command to the database server master node by the expansion tool; The database server master node receives the expansion data redistribution command and executes to expand, obtains an expansion process, and adds the expansion process to the expansion control group, and the expansion control group limits the resources of network IO and disk IO in the expansion process according to the limit; The expansion method of the distributed database further comprises: At any time, the limit of the expansion control group is obtained, and the limit of the expansion control group is updated, so that the expansion control group limits the resources of network IO and disk IO in the expansion process according to the updated limit.

Images