Resource scheduling method, device, equipment and storage medium for container engine Docker

Abstract

The present invention discloses a resource scheduling method, apparatus, device, and storage medium for a container engine Docker. The method comprises: obtaining image layer data of a target image in a cluster image repository; performing hierarchical structure processing on the obtained image layer data; performing image matching between a node and the target image layer data according to the hierarchical structure to determine a matching result; and determining the node of the target image based on the matching result. Whenever the node of the target image is determined, the image layer data of the target image in the image repository is obtained, and the image layer data is matched with each node according to the hierarchical structure. The node of the target image is selected based on the matching result. Since the image matching processing is performed before the node of the target image is determined, the scheduler will assign the target image to the best node, solving the problem that the image within the node cannot match the best node during scheduling due to changes.

Description

Technical Field

[0001] The present invention relates to the field of Docker technology, and in particular to a resource scheduling method, device, equipment and storage medium for a container engine Docker. Background Art

[0002] In existing cluster image repositories, the scheduler will find the optimal node for allocation based on the description of the resource request and the set conditions. Generally, the scheduler will try to select the node where the corresponding image already exists. However, since the image within the node will be changed by the node, it is difficult to find the best matching node when scheduling the new image. Summary of the Invention

[0003] On the one hand, the present invention provides a resource scheduling method for a container engine Docker, the method comprising: obtaining image layer data of a target image in a cluster image repository; performing hierarchical structure processing on the obtained image layer data; performing image matching between a node and the target image layer data according to the hierarchical structure to determine a matching result; and determining a node of the target image based on the matching result.

[0004] In one possible implementation, obtaining the image layer data of the target image in the cluster image repository includes: synchronizing the image layer data of the target image in the cluster image repository to a preset information library; and obtaining the image layer data from the preset information library.

[0005] In one possible implementation, the method further includes: when a new image is pulled from the outside, synchronizing the image layer data of the new image to the preset information library.

[0006] In one possible implementation, after determining the node of the target image, the method further includes: obtaining the matched image layer data from the preset information library; and obtaining the corresponding image from an external image repository based on the image layer data.

[0007] In one possible implementation, mirror matching the node with the target image layer data according to the hierarchical structure to determine the matching result includes: using a multi-threaded mechanism to simultaneously mirror match the image layer data of the target image with multiple different nodes according to the hierarchical structure to respectively determine multiple intermediate results; sorting the multiple intermediate results according to a preset sorting method to determine the matching result.

[0008] In one possible implementation, the matching process between the node and the mirror layer data is as follows: if the mirror layer data and the original mirror data in the node do not have a common root node, the node and the mirror layer data are not matched; if the mirror layer data and the original mirror data in the node have a common root node, the path matching degree between the original mirror data and the mirror layer data is calculated to obtain a matching result.

[0009] In one possible implementation, the hierarchical structure processing specifically includes: processing the image layer data in the form of an acyclic graph.

[0010] On the other hand, the present invention provides a resource scheduling device for a container engine Docker, which includes: a layer data acquisition module for acquiring image layer data of a target image in a cluster image repository; a hierarchical processing module for performing hierarchical structure processing on the acquired image layer data; a layer data matching module for performing image matching between a node and the target image layer data according to the hierarchical structure to determine a matching result; and a layer data selection module for determining a node of the target image based on the matching result.

[0011] On the other hand, the present invention provides a resource scheduling device for a container engine Docker, the device comprising a memory and a processor; the memory is used to store instructions, and the instructions are used to control the processor to operate so as to implement any of the resource scheduling methods for the container engine Docker described above when executed.

[0012] Another aspect of the present invention provides a computer-readable storage medium, which includes a set of computer-executable instructions. When the instructions are executed, they are used to execute any of the resource scheduling methods for the container engine Docker described above.

[0013] In an embodiment of the present invention, whenever the node of the target image is determined, the image layer data of the target image in the image warehouse is obtained, the image layer data is matched with each node according to the hierarchical structure, and the node of the target image is selected based on the matching result. Since the image matching processing is performed before the node of the target image is determined, the scheduler will assign the target image to the best node, which solves the problem that the image in the node cannot be matched with the best node during scheduling due to changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily apparent by reading the following detailed description with reference to the accompanying drawings, in which several embodiments of the present invention are shown by way of example and not limitation, in which:

[0015] In the drawings, the same or corresponding reference numerals denote the same or corresponding parts.

[0016] Figure 1 This is a schematic diagram of the implementation process of a resource scheduling method for a container engine Docker according to an embodiment of the present invention;

[0017] Figure 2 This is a structural diagram of a master node and a slave node in a resource scheduling method for a container engine Docker according to an embodiment of the present invention;

[0018] Figure 3 This is a schematic diagram of the data structure of an image in a resource scheduling method for a container engine Docker according to an embodiment of the present invention;

[0019] Figure 4 This is a schematic diagram of the hierarchical structure of image data in a preset information repository in a resource scheduling method for a container engine Docker according to an embodiment of the present invention;

[0020] Figure 5 This is a schematic diagram of the implementation process of nodes and target images in a resource scheduling method for a container engine Docker according to an embodiment of the present invention;

[0021] Figure 6 This is a schematic diagram of matching the original image and image layer data in a node in a resource scheduling method for a container engine Docker according to an embodiment of the present invention;

[0022] Figure 7 This is a schematic diagram of the entire scheduling process in a resource scheduling method for a container engine Docker according to an embodiment of the present invention;

[0023] Figure 8 This is a schematic diagram of the structural composition of a resource scheduling device for a container engine Docker according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to make the purposes, features, and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0025] Combine Figure 1 As shown, the present invention provides a resource scheduling method for a container engine Docker, the method comprising:

[0026] Step 101: Obtain the image layer data of the target image in the cluster image repository;

[0027] Step 102, performing hierarchical structure processing on the acquired image layer data;

[0028] Step 103: Mirror the node and the target image layer data according to the hierarchical structure to determine the matching result;

[0029] Step 104: Determine the node of the target image based on the matching result.

[0030] In this embodiment, in step 101, the cluster specifically refers to a Docker cluster, which is specifically composed of a group of machines running Docker. The Docker cluster can generally be built using a cluster management tool such as Docker swarm or kubernetes. Figure 2 As shown in the figure, in a Docker cluster, each machine in the cluster is called a node. One of the machines is the manager of the cluster, called the master node, and the rest are called slave nodes. All commands are executed by the master node.

[0031] The cluster image registry is mainly used to store the images required by each machine to run Docker. It can upload (push) and download (extract) images. The registry can be a public repository such as Docker Hub or Alibaba Cloud Registry, or a private container image registry created for the Docker project.

[0032] The target image is the new image applied to the current Docker project. It can be one or more. Each image is composed of multiple layers of data. Figure 3 As shown, for example, it includes the boot file system bootfs of the kernel at the bottom layer, the second layer is the file system rootfs of the Centos image, and above it can be other images such as mysql (image 1), Tomcat (image 2) and read-write layer (container), etc. Of course, the image structure is not limited to Figure 3 The structure shown can also be other forms of structures.

[0033] Combine Figure 2 As shown, when executing step 101, the image layer data of the target image can be obtained from the cluster image repository through the master node Master. Of course, in addition to obtaining the image layer data, other image information can also be obtained, such as code, runtime library, environment variables and configuration files.

[0034] In step 102, after obtaining the image layer data of the target image, the obtained image layer data is subjected to hierarchical structure processing. The hierarchical structure processing is specifically based on hierarchical sorting of the image layer data in the target image according to the original layer structure to generate an acyclic graph form or a tree form, that is, the base layer in the target image (reference Figure 3 The bootfs layer in the file is listed in the first layer, and then the layer data of the next layer is obtained step by step according to the layer structure and arranged in layers in sequence. One of the acyclic graph forms after arrangement can be referred to Figure 4 As shown, Figure 4 It contains six layers, of which the fourth layer contains two branches and the sixth layer contains three branches. Each layer includes layer data and mirror data.

[0035] In step 103, the node and the target image layer data are mirrored according to the hierarchical structure, specifically:

[0036] After the image registry proxy performs hierarchical structure processing, the master node Master obtains the image layer data of each layer in sequence starting from the first layer according to the hierarchical structure, and matches the image layer data with multiple different nodes (i.e., different machines). The data matching method can be to match the image layer data with the original image in each node for data similarity to obtain a matching result. Similarly, according to the above logic, after the matching of each layer of image layer data is completed, the matching results of the target image and multiple nodes are obtained.

[0037] In step 104, after obtaining the matching results of each layer, the node with the highest data matching degree can be selected as the best node for the target image, or one of multiple nodes that simultaneously meet the preset similarity value can be selected as the best node for the target image.

[0038] Therefore, whenever the node of the target image is determined, the image layer data of the target image in the image warehouse is obtained, and the image layer data is matched with each node according to the hierarchical structure. The node of the target image is selected based on the matching result. Since the image matching processing is performed before the node of the target image is determined, the scheduler will assign the target image to the best node, which solves the problem that the image in the node cannot be matched with the best node during scheduling due to changes.

[0039] In one possible implementation, obtaining image layer data of a target image in a cluster image repository includes:

[0040] Synchronize the image layer data of the target image in the cluster image warehouse to the preset information library;

[0041] Get image layer data from the preset repository.

[0042] In this embodiment, the specific process of step 101 is:

[0043] First, the image layer data of the target image is downloaded from the public cluster image repository through the Docker command, and the image layer database is synchronously stored in the preset information library, where the preset information library can be a local database or a cloud database, preferably a private database, which is convenient for storing image data that is not disclosed to the public and has privacy.

[0044] When image layer data needs to be obtained, the image layer data of the target image is obtained from the preset information library.

[0045] Furthermore, the Docker cluster architecture in this embodiment is as follows Figure 2 As shown, Figure 2 It includes a master node Master and three slave nodes Worker. Of course, the number of slave nodes is not limited to three, it can be any number. The master node Master is used to manage multiple slave nodes Worker. The slave node Worker also contains the container generated by the original image in the slave node.

[0046] In addition to the regular cluster image registry and scheduler, the master node Master also has an image registry proxy and a node selector plugin.

[0047] The image registry proxy interacts with the cluster image registry, each slave node Worker, and the node selector plugin respectively, and is used to execute step 101 to obtain the image layer data of the target image from the cluster image registry, and step 102 to perform hierarchical structure processing on the image layer data and store it in the preset information library Meta store. It is also used to obtain node Worker information and send it to the node selector plugin.

[0048] The node selector plugin interacts with the scheduler to receive the worker information of the image registry proxy. It then performs step 103 to match the node with the target image layer data according to the hierarchical structure, calculates and generates a matching result, and sends the matching result to the scheduler. The scheduler then performs step 104 to determine the best node for assigning the target image. The assignment result is sent to the image registry proxy, and the image registry proxy assigns the target image to the corresponding best worker node.

[0049] In one possible implementation, when a new image is pulled from the outside, the image layer data of the new image is synchronized to a preset information repository.

[0050] In this embodiment, during the acquisition of the target image's image layer data, if the target image does not exist in the pre-set repository, the target image is acquired from the public cluster image repository using the Docker pull command. After acquiring the new target image, the corresponding image layer data is synchronized to the pre-set repository to facilitate matching of the image layer data in subsequent steps.

[0051] In one embodiment, after determining the node of the target image, the method further includes:

[0052] Obtain the matching image layer data from the preset information library;

[0053] Based on the image layer data, obtain the corresponding image from the external image repository.

[0054] In this embodiment, after the target image determines the corresponding node, a request to pull the image can be sent to the image registry proxy through the master node. The image registry proxy responds to the request and obtains the image layer data corresponding to the node from the preset information library. Based on the obtained image layer data, it obtains the image file that meets the image layer data from the external image repository and downloads it.

[0055] In one possible implementation, performing mirror matching between the node and the target mirror layer data according to the hierarchical structure and determining the matching result includes:

[0056] Using a multi-threaded mechanism, the image layer data of the target image is simultaneously image matched with multiple different nodes according to the hierarchical structure, and multiple intermediate results are determined respectively;

[0057] Sort the multiple intermediate results according to the preset sorting method to determine the matching results.

[0058] In this embodiment, the multi-thread mechanism can realize parallel processing, thus preventing a task from occupying CPU time for a long time. Figure 5 As shown, the specific matching process in step 103 is:

[0059] First, the image layer data of the target image is obtained in sequence, and the multi-threaded mechanism of the machine is started to match all the original images in all nodes in the current Docker project with the same image layer data at the same time. After each node is matched, all the original images in each node are sorted, and the matching degree of the original image with the highest matching degree is selected as the intermediate result of the node.

[0060] All nodes are sorted according to the matching degree, and the node with the highest matching degree is selected as the best node for the target mirror.

[0061] In one embodiment, the process of matching nodes with image layer data is as follows:

[0062] If the image layer data and the original image data in the node do not have a common root node, the node and the image layer data do not match;

[0063] In the case where the mirror layer data and the original mirror data in the node have a common root node, the path matching degree between the original mirror data and the mirror layer data is calculated to obtain a matching result.

[0064] In this embodiment, combined with Figure 5 As shown, first determine whether the original image data and the image layer data in the node have a common root node. For details, please refer to Figure 6 As shown, the root node is specifically the mirror layer data and the first layer of the original mirror data, that is, it is determined whether the first layer of the original mirror data is consistent with the first layer of the mirror layer data. If the first layers of the two are inconsistent, the matching degree between the two is directly determined to be zero. On the contrary, if the first layers of the two are consistent, the subsequent steps are continued.

[0065] The subsequent steps are to iterate the layer data of the next layer backward in sequence according to the path of the hierarchical structure of the image. During the iteration process, if the layer data of the two are still consistent, the matching degree is increased (for example, the matching value is increased by 1) until data inconsistency occurs, and the iteration is terminated. The total path matching degree is generated and returned.

[0066] refer to Figure 6 As shown, Figure 6First, the first layer of data matching is performed. If the data of the first layer is consistent, the second layer of data matching is performed backward. If the data of the second layer is consistent, and so on, each successful match increases the matching degree. When matching the fourth layer Image3 and Image4, the data of the two are inconsistent, the iteration ends. If the increased matching degree value is 1, then Figure 6 The matching degree is 4, completing the matching process.

[0067] Combine Figure 2 and Figure 7 As shown, the entire scheduling method is:

[0068] First, the image data of the target image is synchronously obtained from the cluster image registry through the image registry proxy in the master node Master, including image layer data, operating environment, code, and other data. The image layer data is hierarchically structured and stored in the preset information library Meta store.

[0069] Then, the scheduler in the master node Master sends a node selection request to the node selector plugin. The node selector plugin responds to the request by sending an instruction to the image registry proxy to obtain the target image adapter node. The image registry proxy executes the instruction to obtain the image information from the preset information library Meta store. The preset information library Meta store executes the instruction to feedback the image data to the image registry proxy. The image registry proxy receives the image data and feeds it back to the node selector plugin.

[0070] The node selector plugin calculates the matching degree between the acquired image layer data and multiple different nodes, and feeds back multiple matching scores to the scheduler.

[0071] The scheduler assigns containers to each node based on the matching score. Each node sends a request to the image registry proxy to pull the corresponding image layer data based on its own container. The image registry proxy responds to the request to obtain the image layer data from the preset information library Meta store, and pulls the corresponding image from the cluster image registry based on the image layer data, so as to finally generate the corresponding container based on the image and assign it to each node.

[0072] like Figure 8 As shown, based on the resource scheduling method of the container engine Docker provided above, the present invention further provides a resource scheduling device of the container engine Docker, the device comprising:

[0073] The layer data acquisition module 201 is used to obtain the image layer data of the target image in the cluster image repository;

[0074] The hierarchical processing module 202 is used to process the acquired image layer data into a hierarchical structure;

[0075] The layer data matching module 203 is used to perform mirror matching between the node and the target mirror layer data according to the hierarchical structure and determine the matching result;

[0076] The layer data selection module 204 is used to determine the node of the target image based on the matching result.

[0077] In this embodiment, in the layer data acquisition module 201, the cluster specifically refers to a Docker cluster, which is specifically composed of a group of machines running Docker. The Docker cluster can generally be built through a cluster management tool such as Docker swarm or kubernetes. Figure 2 As shown in the figure, in a Docker cluster, each machine in the cluster is called a node. One of the machines is the manager of the cluster, called the master node, and the rest are called slave nodes. All commands are executed by the master node.

[0078] The cluster image registry is mainly used to store the images required by each machine to run Docker. It can upload (push) and download (extract) images. The registry can be a public repository such as Docker Hub or Alibaba Cloud Registry, or a private container image registry created for the Docker project.

[0079] The target image is the new image applied to the current Docker project. It can be one or more. Each image is composed of multiple layers of data. Figure 3 As shown in the figure, for example, the boot file system bootfs of the kernel is included in the bottom layer, the file system rootfs of the Centos image is included in the second layer, and other images such as MySQL (image 1), Tomcat (image 2) and read-write layer (container) can be located above it.

[0080] Combine Figure 2 As shown, when executing step 101, the image layer data of the target image can be obtained from the cluster image repository through the image registry proxy in the master node Master. Of course, in addition to obtaining the image layer data, other image information can also be obtained, such as code, runtime library, environment variables and configuration files, etc., where the image registry proxy is preset in advance and is used to record all image information.

[0081] In the hierarchical processing module 202, after obtaining the image layer data of the target image, the obtained image layer data is subjected to hierarchical structure processing. The hierarchical structure processing is specifically based on hierarchical sorting of the image layer data in the target image according to the original layer structure to generate an acyclic graph form or a tree form, that is, the base layer (such as Figure 3 The bootfs layer in the file is listed in the first layer, and then the layer data of the next layer is obtained step by step according to the layer structure and arranged in layers in sequence. One of the acyclic graph forms after arrangement can be referred to Figure 4 As shown, Figure 4 There are six layers in it, the fourth layer contains two branches, and the sixth layer contains three branches.

[0082] In the layer data matching module 203, the node and the target image layer data are mirrored according to the hierarchical structure, specifically:

[0083] After the image registry proxy performs hierarchical structure processing, the node selector plugin in the master node Master obtains the image layer data of each layer in sequence starting from the first layer according to the hierarchical structure, and matches the image layer data with multiple different nodes (i.e., different machines). The data matching method can be to match the image layer data with the original image in each node for data similarity to obtain a matching result. Similarly, according to the above logic, after the matching of each layer of image layer data is completed, the matching result of the target image and multiple nodes is obtained.

[0084] In the layer data selection module 204, after obtaining the matching results of each layer, the node with the highest data matching degree can be selected as the best node for the target image, or one of the nodes that simultaneously meet the preset similarity value can be selected as the best node for the target image.

[0085] Therefore, whenever the node of the target image is determined, the image layer data of the target image in the image warehouse is obtained, and the image layer data is matched with each node according to the hierarchical structure. The node of the target image is selected based on the matching result. Since the image matching processing is performed before the node of the target image is determined, the scheduler will assign the target image to the best node, which solves the problem that the image in the node cannot be matched with the best node during scheduling due to changes.

[0086] Based on the resource scheduling method of the container engine Docker provided above, the present invention further provides a resource scheduling device of the container engine Docker, the device including a memory and a processor;

[0087] The memory is used to store instructions, and the instructions are used to control the processor to operate so as to implement any resource scheduling method of the container engine Docker during execution.

[0088] In an embodiment of the present invention, a computer-readable storage medium includes a set of computer-executable instructions, which, when executed, are used to obtain image layer data of a target image in a cluster image repository; perform hierarchical structure processing on the obtained image layer data; perform image matching between the node and the target image layer data according to the hierarchical structure to determine the matching result; and determine the node of the target image based on the matching result.

[0089] Based on the resource scheduling method of the container engine Docker provided above, the present invention provides a computer-readable storage medium on the other hand, which includes a set of computer-executable instructions, which are used to execute any of the resource scheduling methods of the container engine Docker mentioned above when the instructions are executed.

[0090] In an embodiment of the present invention, a computer-readable storage medium includes a set of computer-executable instructions, which, when executed, are used to obtain image layer data of a target image in a cluster image repository; perform hierarchical structure processing on the obtained image layer data; perform image matching between the node and the target image layer data according to the hierarchical structure to determine the matching result; and determine the node of the target image based on the matching result.

[0091] Therefore, whenever the node of the target image is determined, the image layer data of the target image in the image warehouse is obtained, and the image layer data is matched with each node according to the hierarchical structure. The node of the target image is selected based on the matching result. Since the image matching processing is performed before the node of the target image is determined, the scheduler will assign the target image to the best node, which solves the problem that the image in the node cannot be matched with the best node during scheduling due to changes.

[0092] In the description of this specification, reference to the terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and integrate different embodiments or examples described in this specification, as well as features of different embodiments or examples, unless otherwise inconsistent.

[0093] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of such features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0094] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A resource scheduling method for a container engine Docker, the method comprising: Get the image layer data of the target image in the cluster image repository; Process the acquired image layer data into a hierarchical structure; Performing mirror matching between the node and the mirror layer data according to the hierarchical structure, and determining a matching result; The mirror matching of the node with the image layer data according to the hierarchical structure to determine the matching result includes: using a multi-threaded mechanism to simultaneously mirror match the image layer data of the target image with multiple different nodes according to the hierarchical structure to respectively determine multiple intermediate results; sorting the multiple intermediate results according to a preset sorting method to determine the matching result; the data matching method of the mirror matching is data similarity matching; Based on the matching result, the node of the target image is determined.

2. The method according to claim 1, wherein obtaining the image layer data of the target image in the cluster image repository comprises: Synchronize the image layer data of the target image in the cluster image warehouse to the preset information library; The image layer data is obtained from the preset information library.

3. The method according to claim 2, further comprising: When a new image is pulled from the outside, the image layer data of the new image is synchronized to the preset information library.

4. The method according to claim 2, after determining the node of the target mirror, further comprising: Obtaining the matched image layer data from the preset information library; Based on the image layer data, the corresponding image is obtained from the external image repository.

5. According to the method of claim 1, the matching process of the node and the image layer data is: In the case where the image layer data and the original image data in the node do not have a common root node, the node does not match the image layer data; In the case that the mirror layer data and the original mirror data in the node have a common root node, the path matching degree between the original mirror data and the mirror layer data is calculated to obtain a matching result.

6. The method according to claim 1, wherein the hierarchical structure processing is specifically: The image layer data is processed in the form of an acyclic graph.

7. A resource scheduling device for a container engine Docker, the device comprising: The layer data acquisition module is used to obtain the image layer data of the target image in the cluster image repository; A hierarchical processing module is used to process the acquired image layer data into a hierarchical structure; A layer data matching module, configured to perform mirror matching between the node and the mirror layer data according to the hierarchical structure and determine a matching result; The mirror matching of the node with the image layer data according to the hierarchical structure to determine the matching result includes: using a multi-threaded mechanism to simultaneously mirror match the image layer data of the target image with multiple different nodes according to the hierarchical structure to respectively determine multiple intermediate results; sorting the multiple intermediate results according to a preset sorting method to determine the matching result; the data matching method of the mirror matching is data similarity matching; The layer data selection module is used to determine the node of the target image based on the matching result.

8. A resource scheduling device for a container engine Docker, the device comprising a memory and a processor; The memory is used to store instructions, and the instructions are used to control the processor to operate so as to implement the resource scheduling method of the container engine Docker according to any one of claims 1 to 6 when executed.

9. A readable storage medium comprising a set of computer-executable instructions, which, when executed, are used to execute the resource scheduling method for the container engine Docker according to any one of claims 1 to 6.

Images