KR102016238B1 - System and method for supervising doker container, computer readable medium for performing the method - Google Patents

Abstract

Docker container management method according to the invention using the log information divided into the error log and the scheduling information for determining the host for allocating the task using the log information including the normal log or error log for the docker-based container Generates execution priority information for determining the order of performing actions, which are tasks to be assigned to the host, and transmits a control signal to any one host so that a container for the task is created or regenerated from either host. Thus, the container can be efficiently managed using log information.

Description

Docker container management system and method, recording medium for performing the same {SYSTEM AND METHOD FOR SUPERVISING DOKER CONTAINER, COMPUTER READABLE MEDIUM FOR PERFORMING THE METHOD}

The present invention relates to a docker container management system and method, and a recording medium for performing the same. More particularly, the present invention relates to a docker container management system and method for managing a container using log information of containers created on a host, and to perform the same. It relates to a recording medium.

Linux supports LXC (Linux Container) as an operating system level virtualization method for running multiple Linux systems on a single control host. LXC can allocate and use isolated processes and resources without having to start a real virtual machine. Later, Docker based on LXC was released, and unlike the existing LXC, it was designed to be non-maintainable. Instead, it ensures the same operation without any modifications to any Linux environment, and it is configured more efficiently by minimizing included processes and libraries.

Docker has lowered the barriers to container technology, and Linux containers have begun to be used in a variety of places, the most common of which is cloud computing. The docker container technology does not affect the operation, the number of hosts, or the like, so that the desired operation can be expected only by executing a container prepared in advance.

This conventional container management technology can collect the log from the Docker container and store it in the central server. However, there is no logic that utilizes log information, so there is a limit that an operation beyond the visual confirmation of the log cannot be expected. Accordingly, there is a need for a technique for preparing a countermeasure according to job scheduling between each host or an operation error of a container by analyzing log information.

Korean Patent Registration No. 10-1810762 Korean Patent Publication No. 10-2017-0085072

An aspect of the present invention provides a docker container management system and method for providing an efficient scheduling and error response interface by collecting and analyzing log information generated in a docker-based container operating in multiple hosts, and a recording medium for performing the same. .

The technical problem of the present invention is not limited to the technical problem mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description.

Docker container management method according to an embodiment of the present invention, collecting log information including the normal log or error log for the docker-based container created and operated in the host from each host constituting a computer cluster, Generating scheduling information for determining a host for allocating a task using the log information divided into the normal log, and determining an order of performing operations, which are tasks to be allocated to the host using the log information divided into the error log. Generating execution operation priority information and determining one host to perform a task according to the execution operation priority based on the scheduling information and the priority information, and for the task at any one host, Control by any one of the hosts so that containers are created or regenerated Transmitting a signal.

The generating of the execution priority information may include setting the weight of the execution operation for regenerating the container according to the error log to be higher than the weight of the execution operation for creating a new container, thereby performing the execution operation according to the error log. It may be characterized by setting the priority over the performing operation for creating the new container.

The generating of the operation priority information may be performed by classifying a plurality of collected error logs for each container, calculating an abnormal operation rate for each container, and comparing the abnormal operation rates for each container with each other to perform an order of containers to be regenerated. It may be characterized by determining the operation priority.

The generating of the operation priority information may include calculating a first abnormal operation rate for the first container according to the first error log and a second abnormal operation rate for the second container according to the second error log. When the first abnormal operation rate is higher than the second abnormal operation rate, the second performing operation of regenerating the second container according to the second error log is performed by the first performing operation of regenerating the first container according to the first error log. It may be characterized by setting the priority over the operation.

Generating the execution operation priority information, if an abnormal operation rate for a specific container is less than or equal to a predetermined first threshold, set an execution operation for regenerating the specific container to a lower priority than an execution operation for creating a new container. It can be characterized by.

The generating of the operation priority information may include determining that the error log for the specific container is a temporary error when the abnormal operation rate for the specific container is less than or equal to a second threshold smaller than the first threshold value. The execution operation for creating a container may be excluded from the execution operation priority information.

The host collects the normal log for a predetermined time and periodically transmits the error log, and immediately transmits the error log and generates the execution priority information. The host logs are sent to the error log each time the error log is received. The abnormal operation rate of the container may be calculated, and the execution operation priority information may be updated in real time in comparison with the previously generated execution operation priority information.

The generating of the scheduling information may be characterized in that the scheduling information is periodically updated based on the normal log collected at predetermined time intervals.

In addition, it may be a computer-readable recording medium having a computer program recorded thereon for performing the docker container management method according to the present invention.

In addition, the docker container management system according to an embodiment of the present invention, the host receiving the control signal to create a docker-based container, collects log information for the container in operation and classifies it as a normal log or error log and transmits; Receiving the log information classified as a normal log or an error log from each host constituting the computer cluster, and generating scheduling information for determining a host for allocating a job using the log information classified as the normal log, Generate execution operation priority information for determining an order of performing operations, which are tasks to be allocated to a host using log information divided into the error logs, and based on the scheduling information and the priority information, Determine which host to perform the operation according to, In the host and a management server that transmits a control signal to the one of the host so that the container is generated or regenerated on the job.

According to one aspect of the present invention, the state of the docker container can collectively manage log information about the operation information, and can improve the job scheduling or container module between each host by analyzing the log information, error log By dynamically determining and providing a countermeasure according to an operation error of the container, the load on the container can be reduced.

1 is a conceptual diagram illustrating a schematic configuration of a docker management system according to the present invention.
FIG. 2 is a diagram illustrating a specific configuration of the host of FIG. 1.
3 is a diagram illustrating a specific configuration of the management server of FIG.
4 is a diagram illustrating a data flow in the docker management system of FIG. 1.
5 is a flowchart illustrating a schematic flow of a docker container management method according to an embodiment of the present invention.

DETAILED DESCRIPTION The following detailed description of the invention refers to the accompanying drawings that show, by way of illustration, specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that the various embodiments of the present invention are different but need not be mutually exclusive. For example, certain shapes, structures, and characteristics described herein may be embodied in other embodiments without departing from the spirit and scope of the invention in connection with one embodiment. In addition, it is to be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the invention. The following detailed description, therefore, is not to be taken in a limiting sense, and the scope of the present invention, if properly described, is defined only by the appended claims, along with the full range of equivalents to which such claims are entitled. Like reference numerals in the drawings refer to the same or similar functions throughout the several aspects.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the drawings.

1 is a conceptual diagram illustrating a schematic configuration of a docker container management system according to an embodiment of the present invention.

Specifically, the docker container management system 1000 according to an embodiment of the present invention includes at least one host (100_1, 100_2, 100_3) and the management server 200. Hereinafter, for convenience of description, any one host among the plurality of hosts 100_1, 100_2, and 100_3 constituting the cluster will be described as a host 100.

The host 100 may communicate with other devices by accessing a network, and the host 100 processing an assigned task may be a computer device. However, the host 100 is not limited thereto, and may include other electronic devices capable of performing the functions described below, such as a smartphone, a tablet PC, a notebook, a wearable device, and the like.

The host 100 may execute an application (application) in an independent environment according to the Linux container (LinuX Container) technology. To this end, the host 100 may create at least one docker-based container 110_1 to 110_6. Hereinafter, for convenience of description, any one of the containers generated in one host 100 will be described by representing the container 110 as a representative.

The container 110 may include a Linux module (hereinafter, referred to as a module) that bundles an application and all of its dependencies, libraries, other binaries and configuration files in one package. Accordingly, the host 100 may independently execute the application for each generated container 110. In this process, the host 100 may collect log information about the respective containers 110. For example, the first host 100_1 collects log information about the first container 110_1 and the second container 110_2, and the second host 100_2 collects the third container 110_3 generated by itself and Log information on the fourth container 110_4 may be collected. The log information may be information representing status information and operation information of the Linux module isolated in the container. Hereinafter, for convenience of description, the log information will be described as being defined as information on the state and operation of the container.

The management server 200 may manage a plurality of hosts 100 by clustering. That is, the management server 200 may process a job using a clustering technique using the plurality of hosts 100 as one. The management server 200 may automatically distribute the received traffic according to the load balancing technique and allocate the traffic to each host 100.

The management server 200 may collect log information from the hosts 100 in the cluster. The management server 200 may perform scheduling to determine the order of allocating a task to the host 100 by using the collected log information, and cope with a failure of the container for a specific host 100. Can provide.

The management server 200 may be in the form of a virtualized cloud server, but may also be provided in the form of a physical server.

Hereinafter, each component shown in FIG. 1 will be described in detail with reference to FIGS. 2 to 4.

FIG. 2 is a block diagram illustrating specific functions of the host 100 of FIG. 1.

The host 100 may generate the container 110 at the request of the management server 200. The host 100 may create a docker-based container 110, collect log information about the container 110, and transmit the collected log information to the management server 200. In this case, the container 110 created in the host 100 is any one of the general containers 111_1 and 111_2 and the log container 112, and the host 100 includes a plurality of general containers 111_1 and 111_2 and one log. The container 112 may be created and managed. Hereinafter, for convenience of description, any one of the plurality of general containers 111_1 and 111_2 generated in the host 100 will be described as a representative container 111.

The general container 111 may generate log information about an operation state of an application and transmit log information to the host 100 through a remote system log (syslog). In detail, the general container 111 may transmit an application execution result through a standard stlog or a standard error stderr through a remote syslog. Here, the standard output may be an output path through which log information generated while the application is normally operated is transmitted, and the standard error may be an output path through which log information generated when an error occurs during the execution of the application.

The log container 112 may collect log information stored by mounting the log information generated by the plurality of general containers 111 existing in the host 100 in a database or a stored path. The log container 112 may be divided into a normal log or an error log. The normal log is log information transmitted through standard output, and may be information generated while the container 110 is normally executed as described above. On the other hand, the error log is log information transmitted through a standard error, and may be information generated when the container 110 causes abnormal operation or stops due to an invalid request or a defect thereof. The log container 112 may identify which path the log information is transmitted to and divide the log information into a normal log or an error log. The log container 112 may transmit the divided log information to the management server 200.

Meanwhile, in the illustrated embodiment, each host 100 is illustrated as creating two general containers 111_1 and 111_2 and one log container 112, but the number of containers 110 generated is limited. Do not put

3 is a diagram illustrating a specific function of the management server 200 of FIG.

The management server 200 includes a communication unit 210, a control unit 220 and a database unit 230. The components of the communication unit 210, the control unit 220, and the database unit 230 are arbitrarily divided for convenience of description, and at least two of the above-described elements are integrated into one component. The components of may perform complex functions. Hereinafter, the above-described components will be described in detail.

The communication unit 210 may receive log information received from the log container 112. The communication unit 210 transmits the received normal log and error log to the database unit 230 to be stored, and transmits log information to the control unit 220 to provide information for reference in managing the host 100. can do. In addition, the communication unit 210 may transmit a control signal for managing a job or the container 110 assigned by the management server 200 to the host 100 set by the control unit 200.

The control unit 220 may analyze the log information received from the hosts 100 constituting the same cluster through the communication unit 210 and generate a control signal to be transmitted to the specific host 100 constituting the cluster according to the result. have. To this end, the controller 220 includes an analyzer 221 and a container manager 222. In this regard, it will be described with reference to FIG.

4 is a conceptual diagram illustrating movement of data generated in the docker container management system 1000 of FIG. 1.

The analyzing unit 221 may analyze the log information and the management server 200 may generate various analysis information for managing the host 100 and the container 110. Here, the analysis information includes information on execution time and occupancy rate of each container 110, information on abnormal operation rate, and information on execution operation priority.

The analyzing unit 221 may generate information on the execution time and the occupancy rate of each container 110 by using the log information classified as a normal log among the collected log information. In this case, since the docker-based container is generated in the form of an image, the analysis unit 221 may generate information about execution time and occupancy rate of each container image. That is, the analysis unit 221 may analyze the log information in which the container 110 operates normally to determine which applications are running for which hosts each of the clusters 100 constituting the cluster. Information on the share of the host 100 and the share of the entire cluster can be checked. If it is determined that a large amount of resources for performing a specific task is required, the management server 200 may control to create a container 110 for the same application to the plurality of hosts 100 in the cluster.

In addition, the analyzer 221 may generate information on an abnormal operation rate by using log information classified as an error log among the collected log information. In detail, the analysis unit 221 may classify the collected error logs for each container and calculate information on abnormal operation rates for each container image, that is, for each application. The information on the abnormal operation rate may be information about a probability that the container 110 malfunctions or stops due to a bad request or a defect thereof. When the error log is received, the analysis unit 221 checks which container 110 the error log corresponds to, and checks the normal log for the container 110 among the error logs previously stored in the database unit 230. By referring to the error log, information about an abnormal operation rate, which is an error recurrence probability for the container in which the error log is generated, may be calculated.

The analysis unit 221 may generate information on the execution operation priority using the information on the abnormal operation rate. The performing operation priority may be information referred to in the process of determining which task to process in the process of allocating a task to the host 100 by the management server 200. For example, the management server 200 is required to process a first job to regenerate the first container, a second job to create a new second container, and a third job to regenerate the third container According to the information on the priority of the execution operation generated by the analysis unit 221, it is possible to determine a job to be processed first.

To this end, the analysis unit 221 may determine the execution operation priority according to the type of work and the abnormal operation rate. In detail, the analyzer 221 may give a weight to a job that needs to be reproduced as compared to a job of newly creating a container. According to the above-described embodiment, the analyzer 221 may place the first task and the third task in a higher priority than the second task in the information on the priority of the execution operations. In addition, the analyzer 221 may compare the abnormal operation rate of the first job and the abnormal operation rate of the third job and set the priority of the job having the high abnormal operation rate as a priority. That is, when the error recurrence probability of the first task is higher than the error recurrence probability of the third task, it may be determined to process the first task before the third task. Finally, the analysis unit 221 may generate information on the execution operation priority for allocating a task in the order of the first task, the third task, and the second task. In addition, when there are a plurality of jobs for newly creating a container, the analyzer 221 may set priorities according to the importance of the containers.

In this process, when the error log is received, the analysis unit 221 may update the execution operation priority in real time. As described above, the log container 112 may transmit the divided log information to the management server 200. In this case, the log container 112 may periodically transmit log information divided into normal logs to the management server 200 after collecting for a predetermined time. On the other hand, when the log container 112 receives log information divided into error logs, the log container 112 may immediately transmit the received log information to the management server 200. When the error log is received, the analyzer 221 may calculate an abnormal operation rate of the container with respect to the corresponding error log and compare the information with the previously generated execution priority. For example, when the analysis unit 221 receives the error log for the fourth container, after calculating the abnormal operation rate of the fourth container, the analysis operation priority that reflects the processing order between the first to third jobs generated earlier You can compare the information with ranking. In this case, when the abnormal operation rate of the fourth job, which is a regeneration job for the fourth container, is higher than the abnormal operation rate of the first job, information about the execution operation priority may be updated to process the fourth job as the highest priority. . As such, each time the error log is received, the analysis unit 221 may update the information on the priority of the execution operation in real time.

In some other embodiments, the analyzer 221 may additionally perform a process of comparing the abnormal operation rate with a predetermined threshold value and then generate information on the execution operation priority. In detail, the analyzer 221 may compare the abnormal operation rate of the first task and the abnormal operation rate of the third task with a preset threshold. At this time, if it is determined that the abnormal operation rate of the first job is greater than or equal to the threshold value, the execution operation priority determined in the previous step may be maintained. On the other hand, if it is determined that the abnormal operation rate of the third task is less than the threshold value, the analyzer 221 may set the rank of the third task to be lower than the second task. In other words, the analysis unit 221 sets the job to be regenerated to the priority of the job to be regenerated from the container 110, the abnormal operation rate for the job to be regenerated the container 110, In comparison with the threshold value, a specific task that needs to be regenerated according to the result of the container 110 may be set as a lower priority than a task that needs to be newly created.

In some other exemplary embodiments, the analyzer 221 may exclude the abnormal operation rate from the processing work object based on a result of comparing the abnormal operation rate with another threshold value. If the threshold value according to the above-described embodiment is assumed as the first threshold value, the analyzer 221 may compare the abnormal operation rate with the second threshold value smaller than the first threshold value in the process of comparing the first threshold value. Can be. If it is confirmed that the abnormal operation rate for the third job of the above-described embodiment is less than the second threshold value, the analyzer 221 may exclude the third job from the information on the priority of the execution operation. In other words, if a container malfunctions due to a temporary bad request, the abnormal operation rate is minimal. Therefore, if a one-time error log occurs, it is very likely that the container will operate normally without recreating the container. The cost savings can be expected to improve the overall system performance.

The container manager 222 may be responsible for scheduling the container 110 to distribute the load of the work of the hosts 100 in the cluster by using a load balancing technique. The container manager 222 may generate scheduling information for determining a host for allocating a task using information on execution time and occupancy generated by the analyzer 221 and information on execution priority.

In detail, the container manager 222 determines the type of a task to be currently processed (whether to regenerate an existing container or create a new container) by referring to the information on the execution priority, and provides information on the execution time and the occupancy rate. With reference to this, scheduling information for determining which host 100 in the cluster to allocate the task may be generated. The container manager 222 may transmit a control signal for commanding creation or regeneration of the container to the specific host 100 according to the generated scheduling information.

The database unit 230 may store log information received from the host 100 through the communication unit 210 and analysis information generated by the control unit 220.

As described above, when the docker container management system 1000 according to the present invention is used, the log information containing the state or operation information of the docker-based container may be integrated and managed by the management server 200 which is a concept of a central host. It is easy to extend new services based on. In addition, the management server 200 may improve the load balancing based scheduling or container module by analyzing the log information, and the error and failure of the container 110 using the log information classified as the error log as described above. It can provide intelligent container management service by applying the algorithm to tolerate.

5 is a flowchart illustrating a schematic flow of a docker container management method according to an embodiment of the present invention.

The management server 200 may collect log information from the hosts 100 configuring the cluster (510). In the host 100, a plurality of containers 110 including a general container 111 and a log container 112 are generated, and the host 100 normalizes log information collected from each general container 111. The log may be classified into an abnormal log and transmitted to the management server 200. In this case, the host 100 immediately transmits the error log to the management server 200, but the normal log is kept and may be periodically transmitted whenever a certain time passes.

The management server 200 may generate analysis information using the collected log information (520). In detail, the management server 200 may generate information on execution time and occupancy rate of each container 110 by using the normal log, and generate scheduling information based on the information. The management server 200 may generate information on an abnormal operation rate by using an error log, and generate priority operation priority information according to the abnormal operation rate.

The management server 200 may transmit a control signal to any one host 100 according to scheduling information and processing target priority information (530). The specific host 100 receiving the control signal may generate or regenerate the specific container according to the type of the control signal.

Such a technology for providing a docker container management method may be implemented in the form of program instructions that may be implemented as an application or executed through various computer components, and recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination.

The program instructions recorded on the computer-readable recording medium are those specially designed and configured for the present invention, and may be known and available to those skilled in the computer software arts.

Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical recording media such as CD-ROMs, DVDs, and magneto-optical media such as floptical disks. media), and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like.

Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules to perform the process according to the invention, and vice versa.

Although the above has been described with reference to the embodiments, those skilled in the art will understand that various modifications and changes can be made without departing from the spirit and scope of the invention as set forth in the claims below. Could be.

By applying the docker container management system and method according to the present invention to the docker container integrated management system, it is possible to secure the source technology of the convergence technology between cloud and AI to advance the cloud server and improve the national SW industry competitiveness of the 4th industry. have.

100: host
200: management server

Claims (10)

Collecting log information including a normal log or an error log of a docker-based container generated and operated on the host from each host constituting the computer cluster;
Generating scheduling information for determining a host for allocating a task using the log information divided into the normal log, and determining an order of performing operations, which are tasks to be allocated to the host using the log information divided into the error log. Generating performing operation priority information; And
Determine one host to perform a task according to the execution priority based on the scheduling information and the priority information, and the one of the ones to generate or regenerate a container for the task on the one host. Transmitting a control signal to a host,
Generating the performing operation priority information,
Characterizing the plurality of error logs collected by each container to calculate the abnormal operation rate for each container, and comparing the abnormal operation rate for each container with each other to determine the priority of the operation operation which is the order of the container to be regenerated, How to manage Docker containers.
The method of claim 1,
Generating the performing operation priority information,
By setting the weight of the performing operation for regenerating the container according to the error log higher than the weighting of the performing operation for creating a new container, the performing operation according to the error log takes precedence over the performing operation for creating the new container. A docker container management method, characterized in that the ranking is set.
delete The method of claim 1,
Generating the performing operation priority information,
Calculate a first abnormal operation rate for the first container according to the first error log and a second abnormal operation rate for the second container according to the second error log, wherein the first abnormal operation rate is the second abnormal operation rate; If the ratio is higher than the ratio, the first performing operation to regenerate the first container according to the first error log is set to have a higher priority than the second performing operation to regenerate the second container according to the second error log. How to manage Docker containers.
The method of claim 1,
Generating the performing operation priority information,
And when an abnormal operation rate for a specific container is equal to or less than a predetermined first threshold value, setting an execution operation for regenerating the specific container to a lower priority than an operation operation for creating a new container.
The method of claim 5,
Generating the performing operation priority information,
If the abnormal operation rate for the specific container is less than or equal to the second threshold value smaller than the first threshold value, the execution operation for generating the specific container is determined by determining that the error log for the specific container is a temporary error. Docker container management method, characterized in that excluded from the priority information.
The method of claim 1,
The host collects the normal log for a predetermined time and periodically transmits the error log, and immediately transmits the error log.
The generating of the execution operation priority information may include calculating an abnormal operation rate of the container for the error log each time the error log is received, and comparing the execution operation priority information with previously generated execution operation priority information. Docker container management method, characterized in that for updating in real time.
The method of claim 7, wherein
Generating the scheduling information,
And periodically updating the scheduling information based on the normal log collected at predetermined time intervals.
A computer-readable recording medium having a computer program recorded thereon for performing the docker container management method of claim 1.
A host configured to receive a control signal, generate a docker-based container, collect log information on the operated container, and classify it into a normal log or an error log; And
Receiving the log information classified as a normal log or an error log from each host constituting the computer cluster, and generating scheduling information for determining a host for allocating a job using the log information classified as the normal log, Generate execution operation priority information for determining an order of performing operations, which are tasks to be allocated to a host using log information divided into the error logs, and based on the scheduling information and the priority information, And a management server which determines one host to perform a task according to the above, and transmits a control signal to the one host so that a container for the task is created or regenerated from the one host.
The management server,
Docker container management to classify the plurality of error logs collected for each container to calculate the abnormal operation rate for each container, and to determine the priority of the execution operation which is the order of containers to be regenerated by comparing the abnormal operation rate for each container with each other system.

Images