JP2025090892A - Container System - Google Patents

Abstract

To provide a method for automatically and appropriately changing a setting of a container when an error occurs on the container.SOLUTION: A container system includes: a container capable of executing an application; a container management unit; a container execution unit; a log storage unit; and a container setting change unit. The container management unit manages setting information related to setting of the container. The container execution unit operates the container according to the setting information. The log storage unit stores log information representing an event occurring in the container. The container setting change unit changes the setting information by referring to the log information when an error occurs on the container. The container execution unit operates the container based on the setting information changed by the container setting change unit.SELECTED DRAWING: Figure 1

Description

The present invention relates to a technology for dealing with errors or abnormalities that occur on a container.

Recently, application management technology that uses virtualized containers has been actively developed and utilized. A container can place one or multiple execution environments on a single OS (Operating System). Furthermore, many containers have a function for automatic recovery when an application running on the container is stopped or abnormally terminated. A method has been proposed for estimating the optimal placement of virtual machines/containers for a server selected based on priority (for example, Patent Document 1).

International Publication No. 2022/172385

As mentioned above, when an application running on a container stops or abnormally terminates, recovery processing is performed automatically. However, if recovery processing is performed with incorrect container settings, there is a risk that an error will occur and the restart procedure will be repeated. For this reason, currently, when an application running on a container stops or abnormally terminates, in many cases the user or administrator manually changes the container settings.

An object of one aspect of the present invention is to provide a method for automatically and appropriately changing the settings of a container when an error occurs on the container.

A container system according to one aspect of the present invention includes a container capable of executing an application, a container management unit that manages configuration information related to the configuration of the container, a container execution unit that operates the container according to the configuration information, a log storage unit that stores log information indicating events that occur in the container, and a container setting change unit that changes the configuration information by referring to the log information when an error occurs in the container. The container execution unit operates the container based on the configuration information changed by the container setting change unit.

According to the above-mentioned aspect, when an error occurs in a container, the settings of the container are automatically changed appropriately.

FIG. 2 is a diagram illustrating an example of a functional configuration of a container system according to an embodiment of the present invention. FIG. 1 is a diagram illustrating an example of a hierarchical model of a container system. FIG. 11 is a diagram illustrating an example of an operation sequence of the container system. 11 is a diagram illustrating an example of an operation sequence of the container system when an error occurs on the container. FIG. 4 illustrates an example of log information. FIG. 11 illustrates an example of setting change management information. FIG. 13 is a diagram illustrating an example of setting information before change. FIG. 11 is a diagram illustrating an example of changed setting information. 13 is a flowchart illustrating an example of a process of a container management unit. 13 is a flowchart illustrating an example of a process of a container setting change unit. FIG. 1 illustrates an example of a hardware configuration of a container system.

Figure 1 shows an example of the functional configuration of a container system according to an embodiment of the present invention. The container system 1 according to an embodiment of the present invention includes an application execution platform 10, a log management platform 20, and an application management platform 30. Note that the container system 1 may further include other functions not shown in Figure 1.

The application execution platform 10 includes multiple containers 11 (11a to 11n) and a container execution unit 12. Each container 11 provides a virtual execution environment on a common OS for the multiple containers 11. Hardware resources are allocated to each container 11. Specifically, at least processor resources and memory resources are allocated to each container 11. Then, the container 11 executes one or more applications using the allocated hardware resources. At this time, each container 11 creates log information that indicates events that occur in its own container. The log information is created periodically, for example. In addition, the log information may be created when an unexpected event occurs, such as an error.

The container execution unit 12 deploys the container 11 according to the configuration information provided by the application management infrastructure 30. For example, the container execution unit 12 allocates hardware resources specified by the configuration information to the corresponding container 11. In addition, the container execution unit 12 causes the corresponding container 11 to execute the application specified by the configuration information.

The log management platform 20 includes a log storage unit 21. The log storage unit 21 stores log information created by each container 11 in chronological order. It is also preferable that the log information be stored for each container 11.

The application management infrastructure 30 includes a container management unit 31 and a container setting change unit 32. The container management unit 31 manages setting information related to the setting of the container 11. The setting information is created for each container by, for example, a user of the container system 1. The setting information includes, for example, the following information:
(1) Image (2) Port (3) CPU
(4) Memory (5) Number of replicas (6) User permissions

The image represents the image of an application executed on container 11. The port represents the port used by the application. The CPU represents the upper limit of CPU usage used by container 11 (or allocated to container 11). The memory represents the upper limit of memory usage used by container 11 (or allocated to container 11). The number of replicas represents the number of replicas of an application. The user permissions represent the commands, etc. that the user can use.

The container management unit 31 provides the configuration information created for each container 11 to the container execution unit 12. As a result, the containers 11 operate according to the configuration information. For example, when the configuration information of container 11a indicates "Memory: 200 Mi" and the configuration information of container 11b indicates "Memory: 500 Mi", the container execution unit 12 allocates memory resources of 200 Mi to container 11a and memory resources of 500 Mi to container 11b.

The container setting change unit 32 can change the setting information managed by the container management unit 31. In this embodiment, when an application running on the container 11 terminates abnormally, the container setting change unit 32 refers to the log information stored in the log storage unit 21 and changes the corresponding setting information. For example, when an application running in the container 11a terminates abnormally, the container setting change unit 32 refers to the log information related to the container 11a and changes the setting information of the container 11a.

Figure 2 shows an example of a hierarchical model of the container system 1. The container system 1 is composed of hardware resources 41 for providing multiple execution environments, an OS 42, a container engine 43, and containers 11 (11a to 11n). The hardware resources 41 correspond to a CPU and memory prepared to realize the container system 1. Here, the hardware resources 41 include multiple CPU cores (or multiple processor cores). The hardware resources 41 may also include multiple memory devices. The OS 42 is not particularly limited, and any OS can be adopted. The container engine 43 is provided between the OS 42 and the container 11, and corresponds to the container management unit 31, the container setting change unit 32, and the container execution unit 12. The container engine 43 also includes an interface for accessing the log storage unit 21.

The container 11 is provided with one or more applications to be executed. It is also preferable that the container 11 be provided with various libraries.

The container 11 has a monitoring function for monitoring the status (stopped, started, running, abnormal, etc.) of applications running on the container itself. The container 11 also has a logging function. That is, the container 11 uses the monitoring function and logging function to create log information that indicates events that occur in the container itself. The log information is created periodically, for example. Alternatively, the log information may be created when an unexpected event occurs, such as an abnormal termination of an application. The container 11 also has a communication function. That is, the container 11 accepts instructions or settings from the container engine 43. The container 11 also transmits the created log information to the log storage unit 21.

Figure 3 shows an example of an operation sequence of the container system 1. Note that in the embodiment shown in Figure 3, one container 11 is depicted, but in reality, multiple containers 11a to 11n operate individually. Also, it is assumed that setting information related to the settings of each container 11 is created in advance and managed by the container management unit 31.

First, the container management unit 31 transmits the configuration information of the container 11 to the container execution unit 12. As described above, the configuration information includes information representing, for example, the image, port, CPU, memory, number of replicas, user authority, etc. The container execution unit 12 deploys the container 11 according to the configuration information provided by the container management unit 31. For example, the container execution unit 12 allocates the hardware resources specified by the configuration information to the container 11.

The container 11 uses a monitoring function to monitor the status of an application implemented in its own container. At this time, the monitoring function may periodically send a liveness confirmation signal to the application running on the container, and determine whether the application is operating normally or not based on the presence or absence of a reply. The container 11 then periodically creates status information representing the status of the application and sends it to the container management unit 31. Here, the status information includes, for example, time information and information representing the application status (stopped, started, running, abnormal, etc.). Thus, the container management unit 31 can detect the status of the application implemented in the container 11. For example, when an error occurs in an application implemented in the container 11, the container management unit 31 can detect the error from the status information.

The container 11 also periodically creates log information and sends it to the log storage unit 21. In addition, when an unexpected event (for example, an error such as an abnormal termination of an application) occurs, the container 11 sends log information indicating the event to the log storage unit 21. The log storage unit 21 then stores the log information received from the container 11 in chronological order.

It is preferable that the transmission of status information from the container 11 to the container management unit 31 and the transmission of log information from the container 11 to the log storage unit 21 are synchronized with each other. However, these operations do not necessarily need to be synchronized with each other.

In this way, in the container system 1, the specified application is executed on the container 11 according to the configuration information. Then, the log information of each container 11 is stored in chronological order in the log storage unit 21.

Figure 4 shows an example of the operation sequence of the container system 1 when an error occurs on the container. In this example, it is assumed that an error occurs while an application is being executed on the container 11, causing the application to terminate abnormally.

When an abnormal termination of an application is detected, the container 11 notifies the container management unit 31 of this using the status information. At this time, the status information includes information indicating that the application terminated abnormally and information indicating the time when the application terminated abnormally. The container 11 also creates log information related to the abnormal termination of the application and sends it to the log management platform 20. The log management platform 20 then stores the log information received from the container 11 in the log storage unit 21. The log information includes at least an error log related to the cause of the abnormal termination of the application, in addition to information indicating the time when the application terminated abnormally.

Figure 5 shows an example of log information. In this embodiment, the log information includes information for identifying the container 11 in which the error occurred, information for identifying the application that terminated abnormally, a port number, an error log related to the cause of the abnormal termination (Reason: CrashLoopBackOff), and the like.

When the container management unit 31 receives status information indicating an abnormal termination of an application, it transmits error information to the container setting change unit 32. The error information is generated based on the status information received from the container 11, and includes information identifying the container in which the error occurred and information indicating the time when the error occurred. Note that the information identifying the container in which the error occurred is identified by the container management unit 31 detecting the source of the status information.

When error information is received from the container management unit 31, the container setting change unit 32 uses the error information to refer to and search for log information stored in the log storage unit 21. Specifically, the container setting change unit 32 searches for log information related to the "container in which the error occurred" in a time period close to the "time the error occurred." In this way, the container setting change unit 32 obtains log information related to the abnormal termination of the application.

Then, the container setting change unit 32 determines countermeasures against abnormal termination of the application based on the acquired log information. Here, the container setting change unit 32 holds setting change management information in which countermeasures for resolving errors related to the error log are set, in association with an error log that may be described in the log information. Then, the container setting change unit 32 uses this setting change management information to determine countermeasures against abnormal termination of the application.

Figure 6 shows an example of configuration change management information. In this embodiment, the configuration change management information describes the cause of an error and countermeasures for the error log. The countermeasures indicate a method for resolving the error that occurred, and more specifically, how to change the container configuration information to resolve the error that occurred. For example, it is described that the cause of "OOM (Out of Memory) Killed" is "insufficient memory required to run the container," and the countermeasure is "increase the memory capacity allocated to the container by 1 GB." Note that the configuration change management information is assumed to have been created in advance by, for example, a user or administrator of the container system 1.

The container setting change unit 32 can therefore determine a measure to take against the abnormal termination of an application by searching the acquired log information using the error information. For example, when the error log "OOMKilled" is found from the acquired log information, the container setting change unit 32 determines that the measure to resolve the error that has occurred is to "increase the memory capacity allocated to the container by 1 GB."

The container management unit 31 also transmits the setting information of the container 11 in which the error occurred to the container setting change unit 32. Note that in the example shown in FIG. 4, the error information and the setting information are transmitted separately, but the embodiment of the present invention is not limited to this procedure. In other words, the container management unit 31 may transmit the error information related to the error that occurred and the setting information of the container 11 in which the error occurred together to the container setting change unit 32.

The container setting change unit 32 changes the contents of the setting information received from the container management unit 31 according to the measures determined using the log information. The container setting change unit 32 then transmits the changed setting information to the container management unit 31.

Figures 7 and 8 show an example of change information. Specifically, Figure 7 shows an example of setting information before the change that is sent from the container management unit 31 to the container setting change unit 32. Also, Figure 8 shows an example of setting information after the change that is sent from the container setting change unit 32 to the container management unit 31.

As shown in FIG. 7, the change information includes information related to user authority, information related to image, and information related to memory/CPU, which can be changed by the container setting change unit 32. Information related to user authority indicates commands that can be used by a user of the container system 1. Information related to image indicates an application image to be executed on the container 11. Information related to memory/CPU indicates the memory capacity allocated to the container 11 and the number of CPU cores allocated to the container 11.

In this embodiment, for example, the error log "OOMKilled" is searched for from the log information corresponding to the container 11 in which the error occurred. Here, the countermeasure for dealing with this error log is, as described above, "increase the memory capacity allocated to the container by 1 GB." Therefore, in this case, the container setting change unit 32 changes the description related to memory in the setting information. Specifically, as shown in FIG. 8, the upper limit value of the memory is rewritten from "200 Mi" to "1200 Mi." Then, the container setting change unit 32 transmits the changed setting information to the container management unit 31. At this time, the container setting change unit 32 may transmit a container execution command to the container management unit 31 together with the changed setting information.

The container management unit 31 provides the container execution unit 12 with the configuration information whose contents have been changed by the container configuration change unit 32, and instructs it to execute the container 11. The container execution unit 12 then redeploys (or restarts) the container 11 in accordance with the changed configuration information. The container 11 then operates in accordance with the changed configuration information.

In the example shown in Figures 7 and 8, the memory upper limit has been increased from 200 Mi to 1200 Mi in the configuration information for the container 11. Therefore, the container execution unit 12 increases the memory capacity allocated to the container 11 in accordance with this configuration information. The container 11 then executes the application with the increased memory capacity. This prevents errors from occurring due to a lack of memory.

In this way, in the container system 1 according to an embodiment of the present invention, when an error occurs in the container 11 and causes an application to abnormally terminate, the cause of the error is identified and measures to resolve the error are determined. The configuration information of the container 11 in which the error occurred is then changed in accordance with the determined measures, and the container is redeployed according to the changed configuration information. This prevents the same error from occurring repeatedly in the container 11.

Figure 9 is a flowchart showing an example of the processing of the container management unit 31. In this embodiment, it is assumed that configuration information related to the settings of the container 11 has been created.

In S1, the container management unit 31 transmits the configuration information of the container 11 to the container execution unit 12. As a result, the container execution unit 12 deploys the container 11, and the container 11 executes the implemented application.

In S2, the container management unit 31 monitors the status information sent from the container 11. Then, when status information is received indicating that an error such as an abnormal termination of an application has occurred, the processing of the container management unit 31 proceeds to S3.

In S3, the container management unit 31 sends error information indicating that an error has occurred in the container 11 to the container setting change unit 32. The error information includes information identifying the container 11 in which the error has occurred. In addition, in S4, the container management unit 31 sends setting information of the container 11 in which the error has occurred to the container setting change unit 32. Note that the container management unit 31 may execute S3 and S4 simultaneously. Thereafter, the container management unit 31 waits for a reply from the container setting change unit 32.

In S5, the container management unit 31 receives the configuration information of the container 11 whose contents have been changed by the container setting change unit 32. Then, in S6, the container management unit 31 transmits the changed configuration information to the container execution unit 12. As a result, the container execution unit 12 redeploys the container 11 based on the changed configuration information.

Figure 10 is a flowchart showing an example of the processing of the container setting change unit 32. In S11, the container setting change unit 32 monitors whether error information is sent from the container management unit 31. Then, when error information is received from the container management unit 31, the processing of the container setting change unit 32 proceeds to S12. Note that the container setting change unit 32 receives the setting information of the container 11 in which the error occurred together with the error information. Alternatively, the container setting change unit 32 receives the setting information of the container 11 in which the error occurred after receiving the error information.

In S12 to S13, the container setting change unit 32 refers to the log information stored in the log storage unit 21 based on the error information. At this time, the container setting change unit 32 searches for an error log related to the container 11 specified by the error information (i.e., the container 11 in which the error occurred).

In S14, the container setting change unit 32 identifies the cause of the error corresponding to the error log searched in S13, and determines a measure to resolve the error. At this time, the container setting change unit 32 refers to setting change management information that has been created in advance (for example, the setting change management information shown in FIG. 6).

In S15, the container setting change unit 32 changes the setting information of the container 11 in which the error occurred in accordance with the countermeasure determined in S14. At this time, the container setting change unit 32 changes the setting information, for example, to increase the upper limit of memory/CPU resources to be allocated to the container 11. Alternatively, the container setting change unit 32 changes the setting information to add authority to the user of the container system 1.

In S16, the container setting change unit 32 transmits the changed setting information to the container management unit 31. The container management unit 31 receives the changed setting information in S5 shown in FIG. 9. Therefore, the container 11 is redeployed based on the setting information whose contents have been changed by the container setting change unit 32.

<Variations>
In the above-described embodiment, the container 11 monitors the state of the application, and when an error such as an abnormal termination of the application occurs, the container 11 notifies the container management unit 31 of state information, but the embodiment of the present invention is not limited to this configuration. For example, the container management unit 31 may monitor the state of the container 11. In this case, the container management unit 31 may periodically transmit a survival confirmation signal to each container 11 and determine whether or not each container 11 is operating normally based on the presence or absence of a reply.

In the above-described embodiment, the error information notified from the container management unit 31 to the container setting change unit 32 includes information indicating the time when the error occurred, but the embodiment of the present invention is not limited to this configuration. For example, the container setting change unit 32 that has received the error information may extract an error log corresponding to a new error by searching for unsearched log information. Note that, if log information is recorded for each container, it is sufficient to search for unsearched log information among the log information related to the container 11 specified by the error information (i.e., the container 11 in which the error occurred).

<Hardware Configuration>
11 shows an example of a hardware configuration of the container system 1. The container system 1 is realized by a computer 200 including a processor 201, a memory 202, a storage device 203, an input/output device 204, a recording medium reader 205, and a communication interface 206.

The processor 201 includes multiple processor cores, and controls the operation of the container system 1 by executing a container management program stored in the storage device 203. The container management program includes program code that describes the procedures of the flowcharts shown in Figures 9 and 10. Thus, when the processor 201 executes this program, the functions of the container execution unit 12, container management unit 31, and container setting change unit 32 shown in Figure 1 are provided. In addition, some of the multiple processor cores that make up the processor 201 are allocated to realize the container 11.

The memory 202 is used as a working area for the processor 201. A part of the memory 202 is allocated to realize the container 11. The storage device 203 stores the container management program and other programs. The log storage unit 21 is realized using the storage device 203.

The input/output device 204 includes input devices such as a keyboard, a mouse, a touch panel, and a microphone. The input/output device 204 also includes output devices such as a display device and a speaker. The recording medium reader 205 can acquire data and information recorded on the recording medium 210. The recording medium 210 is a removable recording medium that can be attached to and detached from the computer 200. The recording medium 210 is realized by, for example, a semiconductor memory, a medium that records signals by optical action, or a medium that records signals by magnetic action. The container management program may be provided to the computer 200 from the recording medium 210. The communication interface 206 provides a function for connecting to a network. When the container management program is stored in the program server 220, the computer 200 may acquire the container management program from the program server 220.

1 Container system 10 Application execution platform 11 (11a to 11n) Container 12 Container execution unit 20 Log management platform 21 Log storage unit 30 Application management platform 31 Container management unit 32 Container setting change unit

Claims (5)

A container in which the application can run;
a container management unit that manages setting information related to the setting of the container;
a container execution unit that operates the container in accordance with the setting information;
a log storage unit for storing log information representing events occurring in the container;
a container setting change unit that changes the setting information by referring to the log information when an error occurs in the container,
The container system according to claim 1, wherein the container execution unit operates the container based on the setting information changed by the container setting change unit. The container setting change unit includes:
Identifying the cause of the error that occurred on the container by referring to the log information;
The container system according to claim 1 , further comprising: changing the setting information in response to the identified cause. the container setting change unit holds setting change management information in which a measure for eliminating an error related to the error log is set in association with the error log that may be described in the log information;
The container setting change unit includes:
searching the log information to identify an error log corresponding to an error that occurred on the container;
Referencing the setting change management information, and determining a measure to be taken in response to the identified error log;
The container system according to claim 1 , wherein the setting information is changed in accordance with the determined countermeasure. the container outputs status information indicating that an error has occurred in the container when an application running on the container has abnormally terminated, and stores log information including an error log indicating the content of the error that has occurred in the container in the log storage unit;
The container system according to claim 3, characterized in that the container setting change unit searches the log information in response to the output of the status information to identify an error log corresponding to an error that has occurred in the container. The setting information describes hardware resources to be allocated to the container,
The container system according to claim 1 , wherein the container setting change unit changes the setting information so that an amount of hardware resources allocated to the container is increased.

Images