A CEPH parameter recommendation method, device, equipment and storage medium

Abstract

The present invention provides a CEPH parameter recommendation method, device, equipment and storage medium, and the method includes: docking CEPH to ensure that CEPH can be executed command operation; establishing an indicator collector, collecting and storing CEPH configuration parameters based on the indicator collector; performing benchmark modulation on the configuration parameters to determine the parameter boundary of CEPH; obtaining the resource configuration of the host machine and user requirements, and determining the target parameter recommendation model from the general configuration model based on the sequence model and the deep deterministic policy gradient model based on the resource configuration and user requirements; determining the recommended parameters of CEPH based on the target parameter recommendation model and the parameter boundary, and configuring the recommended parameters to CEPH. The present invention improves the accuracy and determination efficiency of the recommended parameters.

Description

Technical Field

[0001] The present invention relates to the technical field of CEPH parameter tuning, and in particular to a CEPH parameter recommendation method, device, equipment and storage medium. Background Art

[0002] With the continuous development of distributed storage technology, the scale of storage is getting larger and larger, and the amount of data stored is also getting larger and larger. With the development of information technology, the operation and progress of the entire society are greatly dependent on computer information systems. Due to the characteristics of CEPH's 3 copies of data, it can greatly ensure that data is not lost and is widely used in various fields. CEPH is a distributed storage software. While ensuring data security and stability, its performance issues are very important. For example, under special circumstances, when the computer system needs to obtain a large amount of basic information, health information, and trajectory information of personnel, and needs to return to the user in seconds or even milliseconds, it can ensure that the nucleic acid site is normal and not congested in a large area. This requires CEPH to have strong writing and retrieval performance, and the above performance of CEPH is closely related to the parameter configuration of CEPH.

[0003] CEPH itself has hundreds of configuration parameters to control factors such as the cache, memory, CPU, and the frequency of writing data to storage. In the prior art, when tuning CEPH operating parameters, a common practice is to hire dedicated storage experts or storage operation and maintenance personnel to help adjust the parameters, but the parameter adjustment cycle is long, the efficiency is low, and the labor cost is too high.

[0004] Therefore, there is an urgent need to provide a CEPH parameter recommendation method, device, equipment and storage medium to solve the above technical problems. Summary of the invention

[0005] In view of this, it is necessary to provide a CEPH parameter recommendation method, device, equipment and storage medium to solve the technical problems of low efficiency and high cost of CEPHE parameter recommendation in the prior art.

[0006] In one aspect, the present invention provides a CEPH parameter recommendation method, comprising:

[0007] Connect to CEPH to ensure that the CEPH can execute command operations;

[0008] Establishing an indicator collector, and collecting and storing the configuration parameters of the CEPH based on the indicator collector;

[0009] Performing a reference modulation on the configuration parameters to determine the parameter boundaries of the CEPH;

[0010] Obtain resource configuration of the host machine and user requirements, and determine a target parameter recommendation model from a general configuration model based on a sequence model and a deep deterministic policy gradient model based on the resource configuration and the user requirements;

[0011] Determine the recommended parameters of the CEPH based on the target parameter recommendation model and the parameter boundary, and configure the recommended parameters to the CEPH.

[0012] In some possible implementations, the resource configuration includes GPU idle resources and / or CPU idle resources, and the user demand includes an accuracy requirement value; and determining a target parameter recommendation model from a general configuration model based on a sequence model and a deep deterministic policy gradient model based on the resource configuration and the user demand includes:

[0013] When both the GPU idle resources and the CPU idle resources are greater than the preset idle resources, or the accuracy requirement value is greater than the preset requirement value, determining that the deep deterministic policy gradient model is the target parameter recommendation model;

[0014] When the GPU idle resources or the CPU idle resources are less than or equal to the preset idle resources, and the accuracy requirement value is less than or equal to the preset requirement value, the general configuration model based on the sequence model is determined as the target parameter recommendation model.

[0015] In some possible implementations, performing a reference modulation on the configuration parameter to determine a parameter boundary of the CEPH includes:

[0016] Determine an independent parameter in the configuration parameters and an associated parameter associated with the independent parameter;

[0017] Determining an association relationship between the independent parameter and the associated parameter;

[0018] An independent parameter boundary of the independent parameter is determined, and an associated parameter boundary of the associated parameter is determined based on the independent parameter boundary and the associated relationship.

[0019] In some possible implementations, the CEPH parameter recommendation method further includes:

[0020] defining a benchmark sample, and inputting the benchmark sample into the CEPH configured with the recommended parameters;

[0021] A verification index after the CEPH executes the benchmark example is obtained, and the recommended parameter is evaluated based on the verification index.

[0022] In some possible implementations, after performing reference modulation on the configuration parameter to determine the parameter boundary of the CEPH, the method further includes:

[0023] The parameter boundaries are persistently stored.

[0024] In some possible implementations, the determining the recommended parameters of the CEPH based on the target parameter recommendation model and the parameter boundary further includes:

[0025] generating initial parameters based on the parameter boundaries, and generating parameters to be confirmed based on the initial parameters and the target parameter recommendation model;

[0026] Configuring the parameters to be confirmed to the CEPH, and collecting and storing the indicator information of the CEPH based on the indicator collector;

[0027] It is determined whether the indicator information meets the requirements. If the indicator information meets the requirements, the parameters to be confirmed are the recommended parameters. If the indicator information does not meet the requirements, the parameters to be confirmed are updated.

[0028] In some possible implementations, the indicator information includes the number of mon connection sessions, read and write efficiency, OSD write delay count, OSD write total delay, OSD read delay count, OSD read total delay, RBD write delay count, RBD write total delay, RBD read delay count and RBD read total delay.

[0029] On the other hand, the present invention also provides a CEPH parameter recommendation device, comprising:

[0030] A CEPH docking unit, used for docking with CEPH to ensure that the CEPH can execute command operations;

[0031] An indicator collector establishing unit, used to establish an indicator collector, and to collect and store configuration parameters of the CEPH based on the indicator collector;

[0032] a parameter boundary determination unit, configured to perform reference modulation on the configuration parameter to determine a parameter boundary of the CEPH;

[0033] A parameter recommendation model determination unit, used to obtain the resource configuration of the host machine and the user demand, and determine the target parameter recommendation model from a general configuration model based on a sequence model and a deep deterministic policy gradient model based on the resource configuration and the user demand;

[0034] A recommended parameter configuration unit is used to determine the recommended parameters of the CEPH based on the target parameter recommendation model and the parameter boundary, and configure the recommended parameters to the CEPH.

[0035] On the other hand, the present invention also provides a CEPH parameter recommendation device, including a memory and a processor, wherein:

[0036] The memory is used to store programs;

[0037] The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the CEPH parameter recommendation method described in any of the above implementations.

[0038] On the other hand, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instruction, which, when executed by a processor, can implement the steps of the CEPH parameter recommendation method described in any of the above implementations.

[0039] The beneficial effect of adopting the above embodiment is: the CEPH parameter recommendation method provided by the present invention, based on the target parameter recommendation model determined from the general configuration model based on the sequence model and the deep deterministic policy gradient model, realizes automatic recommendation of parameters without human intervention, reduces labor costs, and improves the efficiency of parameter recommendation.

[0040] Furthermore, before generating recommended parameters through the target parameter recommendation model, the present invention first performs benchmark modulation on the configuration parameters to determine the parameter boundaries of CEPH, which can make the recommended parameters more reasonable and accurate, avoid invalid recommendations, and further improve the accuracy and determination efficiency of the recommended parameters.

[0041] Furthermore, the present invention can improve the compatibility of the determined target parameter recommendation model with the host machine and user needs by determining the target parameter recommendation model from the general configuration model based on the sequence model and the deep deterministic policy gradient model based on resource configuration and user needs, thereby improving the user satisfaction of the CEPH parameter recommendation method. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0043] Figure 1 A schematic diagram of an embodiment of the CEPH parameter recommendation method provided by the present invention;

[0044] Figure 2 For the present invention Figure 1 A schematic flow chart of an embodiment of S103;

[0045] Figure 3 A schematic diagram of a flow chart of an embodiment of evaluating recommended parameters provided by the present invention;

[0046] Figure 4 For the present invention Figure 1 A schematic diagram of an embodiment of S105;

[0047] Figure 5 A schematic diagram of the structure of an embodiment of a CEPH parameter recommendation device provided by the present invention;

[0048] Figure 6 A schematic diagram of the structure of an embodiment of a CEPH parameter recommendation device provided by the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments 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 creative work are within the scope of protection of the present invention.

[0050] It should be understood that the schematic drawings are not drawn to scale. The flowchart used in the present invention shows the operations implemented according to some embodiments of the present invention. It should be understood that the operations of the flowchart can be implemented out of order, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art, under the guidance of the content of the present invention, can add one or more other operations to the flowchart, and can also remove one or more operations from the flowchart. Some of the block diagrams shown in the accompanying drawings are functional entities, which do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor systems and / or microcontroller systems.

[0051] In the description of the embodiments of the present invention, "and / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent three situations: A exists alone, A and B exist at the same time, and B exists alone.

[0052] Reference to an "embodiment" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0053] The present invention provides a CEPH parameter recommendation method, device, equipment and storage medium, which are described below respectively.

[0054] Figure 1 A flow chart of an embodiment of the CEPH parameter recommendation method provided by the present invention is shown in FIG. Figure 1 As shown, the CEPH parameter recommendation method includes:

[0055] S101, connect to CEPH to ensure that CEPH can execute command operations;

[0056] S102, establish an indicator collector, and collect and store CEPH configuration parameters based on the indicator collector;

[0057] S103, performing benchmark modulation on the configuration parameters to determine the parameter boundaries of CEPH;

[0058] S104, obtaining the resource configuration of the host machine and the user demand, and determining the target parameter recommendation model from a Sequential Model-based Algorithm Configuration (SMAC) model and a Deep Deterministic Policy Gradient (DDPG) model based on the resource configuration and the user demand;

[0059] S105 : Determine the recommended parameters of CEPH based on the target parameter recommendation model and the parameter boundary, and configure the recommended parameters to CEPH.

[0060] Compared with the prior art, the CEPH parameter recommendation method provided in the embodiment of the present invention realizes automatic parameter recommendation based on the target parameter recommendation model determined from the general configuration model based on the sequence model and the deep deterministic policy gradient model, without the need for human intervention, thereby reducing labor costs and improving the efficiency of parameter recommendation.

[0061] Furthermore, before generating recommended parameters through the target parameter recommendation model, the embodiment of the present invention first performs benchmark modulation on the configuration parameters to determine the parameter boundaries of CEPH, which can make the recommended parameters more reasonable and accurate, avoid invalid recommendations, and further improve the accuracy and determination efficiency of the recommended parameters.

[0062] Furthermore, the embodiments of the present invention can improve the compatibility of the determined target parameter recommendation model with the host machine and user needs by determining the target parameter recommendation model from the general configuration model based on the sequence model and the deep deterministic policy gradient model based on resource configuration and user needs, thereby improving the user satisfaction of the CEPH parameter recommendation method.

[0063] Among them, the connection with CEPH in step S101 is specifically as follows: on the one hand, some common commands of CEPH are encapsulated by using shell, including the ceph tell injectargs command for dynamically modifying parameters, ceph daemonmgr.x dump_osd_network and perf dump for obtaining osd status, and mon dump for obtaining mon status, etc. On the other hand, Python is used to encapsulate the start and stop of CEPH, configuration file modification and distribution, host machine indicator collection, rbd block device creation processing, etc., so as to achieve normal start and stop, and copy ceph.conf files to containers or system-side operations.

[0064] The implementation principle is to encapsulate different operations and call the built-in command line of CEPH. In subsequent operations, the corresponding interface methods are called respectively to be compatible with different CEPH architectures.

[0065] In a specific embodiment of the present invention, CEPH includes container CEPH and physical deployment CEPH, and step S101 is specifically as follows: for container CEPH, firstly, for static parameters or modification of configuration files, the kubernetes configmap will be encapsulated and mounted as the ceph.conf configuration file in the container; secondly, CEPH is run in the form of resources such as daemonset, statefulset, deployment, etc., and exists on the kubernetes platform in the form of different namespaces and pods; furthermore, the process of configuring the recommended parameters to access the API, services, and pod resources of the Kubernetes api-server is encapsulated into an API, and the process of pod->recommended device, recommended device->pod is encapsulated into pod_to_app and app_to_pod functions based on the TCP protocol; finally, when the recommended parameters need to be configured to the ceph pod, the relevant functions can be called to implement the logic of copying the file to the container, thereby realizing access and operation. For the physically deployed CEPH, the operations such as start and stop, copy, dynamic / static parameter modification are also encapsulated into python and Shell commands. When the recommended parameters need to be pushed to the physically deployed CEPH, the relevant shell commands are called to realize access and operation.

[0066] Since the SMAC model is characterized by fast optimization and small resource occupation, and the DDPG model is characterized by more accurate control of the details of parameter values, based on the difference between the two models, in some embodiments of the present invention, resource configuration includes GPU idle resources and / or CPU idle resources, and user requirements include accuracy requirement values; then step S104 includes:

[0067] When both the GPU idle resources and the CPU idle resources are greater than the preset idle resources, or the accuracy requirement value is greater than the preset requirement value, the deep deterministic policy gradient model is determined as the target parameter recommendation model;

[0068] When the GPU idle resources or the CPU idle resources are less than or equal to the preset idle resources, and the accuracy requirement value is less than or equal to the preset requirement value, the general configuration model based on the sequence model is determined as the target parameter recommendation model.

[0069] It should be noted that in addition to the accuracy requirement value, user requirements can also be time requirement values, that is, how long it takes to complete the parameter recommendation. If the time requirement value is less than the time threshold, the general configuration model based on the sequence model is determined as the target parameter recommendation model. Otherwise, the deep deterministic policy gradient model is determined as the target parameter recommendation model.

[0070] Among them, the specific process of the SMAC model to determine the recommended parameters is: initialize the Bayesian loop, randomly generate two hyperparameter combinations of the SMAC model, one of the two hyperparameter combinations is the best combination, and the other is the challenge combination; perform iterative operations, and the iterative operations include: compare the performance of the best combination and the challenge combination on the target algorithm, update the best combination according to the comparison result, and obtain the updated best combination; in response to determining that the performance of the updated best combination on the target algorithm meets the preset conditions, determine the updated best combination as the current optimal hyperparameter combination and output it, that is, generate recommended parameters.

[0071] The specific process of the DDPG model to determine the recommended parameters is as follows: set the state, action, action parameters and reward, use the parameters as the state, use the CEPH operating indicators corresponding to the parameters as the action, and establish a reward function to determine the reward. Iterate based on the state, action, action parameters and reward. When the training requirements are met, the parameters are used as recommended parameters.

[0072] Since CEPH needs to configure many parameters, in order to ensure the rationality of the balance and ensure the accuracy and efficiency of the determined recommended parameters, in some embodiments of the present invention, Figure 2 As shown, step S103 includes:

[0073] S201, determining an independent parameter and an associated parameter associated with the independent parameter in the configuration parameters;

[0074] S202, determining the correlation relationship between the independent parameter and the correlation parameter;

[0075] S203: Determine independent parameter boundaries of the independent parameters, and determine associated parameter boundaries of the associated parameters based on the independent parameter boundaries and the associated relationships.

[0076] The embodiment of the present invention can ensure the accuracy and speed of setting the boundary of the associated parameter based on the association relationship between the associated parameter and the independent parameter by determining the independent parameter and the associated parameter associated with the independent parameter in the configuration parameter.

[0077] In a specific embodiment of the present invention, the minimum value of osd_memory_target is greater than

[0078] osd_memory_base+osd_memory_cache_min

[0079] It is subject to the ratio limit of osd_memory_target_cgroup_limit_ratio, and needs to be converted according to 1024. The value divided by the two must be greater than the specified storage space size and must be a prime number.

[0080] Specifically, the minimum value of the RBD cache size is 1G and the maximum value is memory_num / 16.

[0081] The association relationship includes, but is not limited to: a power of 2, N times of a page size, a specified percentage threshold, a multiple based on a minimum value, or a modulo operation, etc.

[0082] To verify the accuracy and feasibility of the recommended parameters, in some embodiments of the present invention, Figure 3 As shown, the CEPH parameter recommendation method also includes:

[0083] S301, defining a benchmark sample, and inputting the benchmark sample into CEPH configured with recommended parameters;

[0084] S302: Obtain verification indicators after CEPH executes the benchmark sample, and evaluate the recommended parameters based on the verification indicators.

[0085] The embodiment of the present invention evaluates the recommended parameters through benchmark samples, and the accuracy and feasibility of the recommended parameters can be learned based on the evaluation results, thereby ensuring the rationality of the recommended parameters.

[0086] The benchmark examples can be random / sequential read / write of CEPH RBD, which meets TPC-C and TPC-H standards. The benchmark examples can also be more complex query, calculation, insertion and other examples that CEPH has encountered in production.

[0087] Specifically, the verification indicators include the number of mon connection sessions, read and write efficiency (TPS and QPS), OSD write delay count, OSD write total delay, OSD read delay count, OSD read total delay, RBD write delay count, RBD write total delay, RBD read delay count, and RBD read total delay.

[0088] Among them, TPS is the number of transactions per second (Transcantion Per Second), and QPS is the number of requests and queries per second (Query Per Second).

[0089] To facilitate subsequent calling, querying and tracing of the recommended parameters, in some embodiments of the present invention, after step S103, the process further includes: persistently storing the parameter boundaries.

[0090] The specific process of persistently storing the boundary parameters is as follows: storing the boundary parameters in a storage medium, which may include a floppy disk, a CD, a DVD, a hard disk, a flash memory, a U disk, a CF card, a SD card, an MMC card, a SM card, a Memory Stick, an XD card, etc.

[0091] It should be noted that parameter boundaries are stored in the form of key-value pairs.

[0092] In some embodiments of the present invention, Figure 4 As shown, step S105 includes:

[0093] S401, generating initial parameters based on parameter boundaries, and generating parameters to be confirmed based on the initial parameters and the target parameter recommendation model;

[0094] S402, configuring the parameters to be confirmed to CEPH, and collecting and storing the indicator information of CEPH based on the indicator collector;

[0095] S403: Determine whether the indicator information meets the requirements. If the indicator information meets the requirements, the parameters to be confirmed are recommended parameters. If the indicator information does not meet the requirements, the parameters to be confirmed are updated.

[0096] The specific process of updating the parameters to be confirmed is: modify the parameters to be confirmed, generate modified parameters, and then configure the modified parameters to CEPH, and repeat steps S402-S403 until the indicator information meets the requirements. When the indicator information meets the requirements, the corresponding parameters are recommended parameters.

[0097] It should be understood that in order to facilitate querying and tracing the indicator information, after the indicator information is obtained, the parameter boundaries are also persistently stored.

[0098] It should be noted that: since the parameters and indicator information to be confirmed include multiple dimensions, their magnitudes, positive and negative numbers and units are all different, and the parameters and indicator information to be confirmed may include multiple related indicators, in order to improve the model operation success rate of the target parameter recommendation model, the parameters and indicator information to be confirmed need to be preprocessed. The preprocessing process includes but is not limited to feature analysis, dimensionality reduction and standardization.

[0099] First, the parameters and indicators are read from the storage medium, temporarily saved to a local file, and the types of parameters and indicators are processed, such as unifying the storage space to MB, converting the string values ​​of multiple options into defined enumeration values ​​of 0 and 1, and subdividing the values ​​into integer and floating-point types; secondly, it is necessary to process parameters of different sizes, positive and negative numbers, and numbers of different units. Here, the standardization algorithm in machine learning is used to compress the values ​​to the range of [0,1] so that they can be calculated in the matrix and described on the scatter plot; thirdly, it is necessary to perform feature engineering on the correlation between indicators and parameters. Here, the FA / PCA related algorithm of machine learning is used to connect and distinguish principal component analysis (PCA) and factor analysis (FA), and at the same time perform dimensionality reduction on the data, which can convert a large number of related variables into a group of very few unrelated variables; finally, the principal components of the correlation are clustered. When the parameters corresponding to this type of components are adjusted, the parameters will depend on each other and change at the same time.

[0100] Among them, the indicator information and verification indicators are the same, and will not be described here one by one.

[0101] It should be noted that: when an exception occurs in the target parameter recommendation model during training, the program's rollback mechanism is triggered. Depending on the type of error, it determines whether to roll back the recommended parameter settings, whether to throw an exception to stop training due to insufficient computing power, whether to push the exception to the user if the specification configuration is not met, and whether to retry the connection to continue training if the connection is abnormal.

[0102] In order to better implement the CEPH parameter recommendation method in the embodiment of the present invention, based on the CEPH parameter recommendation method, the embodiment of the present invention also provides a CEPH parameter recommendation device, such as Figure 5 As shown, the CEPH parameter recommendation device 500 includes:

[0103] A CEPH docking unit 501 is used to dock with CEPH to ensure that CEPH can execute command operations;

[0104] An indicator collector establishing unit 502 is used to establish an indicator collector, and collect and store configuration parameters of CEPH based on the indicator collector;

[0105] A parameter boundary determination unit 503, configured to perform reference modulation on the configuration parameters to determine the parameter boundary of the CEPH;

[0106] A parameter recommendation model determination unit 504 is used to obtain the resource configuration of the host machine and the user demand, and determine the target parameter recommendation model from the general configuration model based on the sequence model and the deep deterministic policy gradient model based on the resource configuration and the user demand;

[0107] The recommended parameter configuration unit 505 is used to determine the recommended parameters of CEPH based on the target parameter recommendation model and the parameter boundary, and configure the recommended parameters to CEPH.

[0108] The CEPH parameter recommendation device 500 provided in the above embodiment can implement the technical solution described in the above CEPH parameter recommendation method embodiment. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above CEPH parameter recommendation method embodiment, which will not be repeated here.

[0109] like Figure 6 As shown, the present invention also provides a CEPH parameter recommendation device 600. The CEPH parameter recommendation device 600 includes a processor 601, a memory 602 and a display 603. Figure 6 Only some components of the CEPH parameter recommendation device 600 are shown, but it should be understood that it is not required to implement all of the shown components, and more or fewer components may be implemented instead.

[0110] In some embodiments, the processor 601 may be a central processing unit (CPU), a microprocessor or other data processing chip, and is used to run program codes or process data stored in the memory 602, such as the CEPH parameter recommendation method of the present invention.

[0111] In some embodiments, the processor 601 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, the processor 601 may be local or remote. In some embodiments, the processor 601 may be implemented in a cloud platform. In one embodiment, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-cloud, etc., or any combination thereof.

[0112] In some embodiments, the memory 602 may be an internal storage unit of the CEPH parameter recommendation device 600, such as a hard disk or memory of the CEPH parameter recommendation device 600. In other embodiments, the memory 602 may also be an external storage device of the CEPH parameter recommendation device 600, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the CEPH parameter recommendation device 600.

[0113] Furthermore, the memory 602 may include both an internal storage unit and an external storage device of the CEPH parameter recommendation device 600. The memory 602 is used to store application software for installing the CEPH parameter recommendation device 600 and various data.

[0114] In some embodiments, the display 603 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 603 is used to display information on the CEPH parameter recommendation device 600 and to display a visual user interface. The components 601-603 of the CEPH parameter recommendation device 600 communicate with each other via a system bus.

[0115] In one embodiment, when the processor 601 executes the CEPH parameter recommendation program in the memory 602, the following steps may be implemented:

[0116] Connect to CEPH to ensure that commands can be performed on CEPH;

[0117] Establish an indicator collector, based on which the configuration parameters of CEPH are collected and stored;

[0118] Benchmark the configuration parameters to determine the parameter boundaries of CEPH;

[0119] Obtain the resource configuration of the host machine and user requirements, and determine the target parameter recommendation model from the general configuration model based on the sequence model and the deep deterministic policy gradient model based on the resource configuration and user requirements;

[0120] The recommended parameters of CEPH are determined based on the target parameter recommendation model and the parameter boundary, and the recommended parameters are configured to CEPH.

[0121] It should be understood that: when the processor 601 executes the CEPH parameter recommendation program in the memory 602, in addition to the above functions, other functions may also be implemented, and details may refer to the description of the above corresponding method embodiments.

[0122] Furthermore, the embodiment of the present invention does not specifically limit the type of the CEPH parameter recommendation device 600 mentioned. The CEPH parameter recommendation device 600 may be a portable CEPH parameter recommendation device such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop computer, etc. Exemplary embodiments of portable CEPH parameter recommendation devices include but are not limited to portable CEPH parameter recommendation devices equipped with IOS, Android, Microsoft or other operating systems. The above-mentioned portable CEPH parameter recommendation device may also be other portable CEPH parameter recommendation devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the CEPH parameter recommendation device 600 may not be a portable CEPH parameter recommendation device, but a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0123] Accordingly, an embodiment of the present application also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the steps or functions in the CEPH parameter recommendation method provided in the above-mentioned method embodiments can be implemented.

[0124] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0125] The above is a detailed introduction to a CEPH parameter recommendation method, device, equipment and storage medium provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A CEPH parameter recommendation method, characterized in that: include: Connect to CEPH to ensure that the CEPH can execute command operations; Establishing an indicator collector, and collecting and storing the configuration parameters of the CEPH based on the indicator collector; Performing a reference modulation on the configuration parameters to determine the parameter boundaries of the CEPH; Obtain resource configuration of the host machine and user requirements, and determine a target parameter recommendation model from a general configuration model based on a sequence model and a deep deterministic policy gradient model based on the resource configuration and the user requirements; Determining recommended parameters of the CEPH based on the target parameter recommendation model and the parameter boundary, and configuring the recommended parameters to the CEPH; The performing a reference modulation on the configuration parameter to determine the parameter boundary of the CEPH includes: Determine an independent parameter in the configuration parameters and an associated parameter associated with the independent parameter; Determining an association relationship between the independent parameter and the associated parameter; An independent parameter boundary of the independent parameter is determined, and an associated parameter boundary of the associated parameter is determined based on the independent parameter boundary and the associated relationship.

2. The CEPH parameter recommendation method according to claim 1, characterized in that: The resource configuration includes GPU idle resources and / or CPU idle resources, and the user demand includes an accuracy requirement value; and determining a target parameter recommendation model from a general configuration model based on a sequence model and a deep deterministic policy gradient model based on the resource configuration and the user demand, comprising: When both the GPU idle resources and the CPU idle resources are greater than the preset idle resources, or the accuracy requirement value is greater than the preset requirement value, determining that the deep deterministic policy gradient model is the target parameter recommendation model; When the GPU idle resources or the CPU idle resources are less than or equal to the preset idle resources, and the accuracy requirement value is less than or equal to the preset requirement value, the general configuration model based on the sequence model is determined as the target parameter recommendation model.

3. The CEPH parameter recommendation method according to claim 1, characterized in that: The CEPH parameter recommendation method further includes: defining a benchmark sample, and inputting the benchmark sample into the CEPH configured with the recommended parameters; A verification index after the CEPH executes the benchmark example is obtained, and the recommended parameter is evaluated based on the verification index.

4. The CEPH parameter recommendation method according to claim 1, characterized in that: After performing reference modulation on the configuration parameters to determine the parameter boundary of the CEPH, the method further includes: The parameter boundaries are persistently stored.

5. The CEPH parameter recommendation method according to claim 1, characterized in that: The determining the recommended parameters of the CEPH based on the target parameter recommendation model and the parameter boundary further includes: generating initial parameters based on the parameter boundaries, and generating parameters to be confirmed based on the initial parameters and the target parameter recommendation model; Configuring the parameters to be confirmed to the CEPH, and collecting and storing the indicator information of the CEPH based on the indicator collector; It is determined whether the indicator information meets the requirements. If the indicator information meets the requirements, the parameters to be confirmed are the recommended parameters. If the indicator information does not meet the requirements, the parameters to be confirmed are updated.

6. The CEPH parameter recommendation method according to claim 5, characterized in that: The indicator information includes the number of mon connection sessions, read and write efficiency, OSD write delay count, OSD write total delay, OSD read delay count, OSD read total delay, RBD write delay count, RBD write total delay, RBD read delay count and RBD read total delay.

7. A CEPH parameter recommendation device, characterized in that: include: A CEPH docking unit, used for docking with CEPH to ensure that the CEPH can execute command operations; An indicator collector establishing unit, used to establish an indicator collector, and to collect and store configuration parameters of the CEPH based on the indicator collector; a parameter boundary determination unit, configured to perform reference modulation on the configuration parameter to determine a parameter boundary of the CEPH; A parameter recommendation model determination unit, used to obtain the resource configuration of the host machine and the user demand, and determine the target parameter recommendation model from a general configuration model based on a sequence model and a deep deterministic policy gradient model based on the resource configuration and the user demand; a recommended parameter configuration unit, configured to determine the recommended parameters of the CEPH based on the target parameter recommendation model and the parameter boundary, and configure the recommended parameters to the CEPH; The performing a reference modulation on the configuration parameter to determine the parameter boundary of the CEPH includes: Determine an independent parameter in the configuration parameters and an associated parameter associated with the independent parameter; Determining an association relationship between the independent parameter and the associated parameter; An independent parameter boundary of the independent parameter is determined, and an associated parameter boundary of the associated parameter is determined based on the independent parameter boundary and the associated relationship.

8. A CEPH parameter recommendation device, characterized in that: comprising a memory and a processor, wherein: The memory is used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the CEPH parameter recommendation method described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the CEPH parameter recommendation method described in any one of claims 1 to 6.

Images