WO2024250628A1 - 一种磁盘状态监控及故障预测方法、装置及设备 - Google Patents
一种磁盘状态监控及故障预测方法、装置及设备 Download PDFInfo
- Publication number
- WO2024250628A1 WO2024250628A1 PCT/CN2023/139637 CN2023139637W WO2024250628A1 WO 2024250628 A1 WO2024250628 A1 WO 2024250628A1 CN 2023139637 W CN2023139637 W CN 2023139637W WO 2024250628 A1 WO2024250628 A1 WO 2024250628A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- disk
- prediction
- fault
- status
- prediction model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/3037—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system component is a memory, e.g. virtual memory, cache
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Definitions
- disk failure prediction model A is always used to predict disk failures, but in some cases, disk failure prediction model B is actually used to predict disk failures, and its disk failure prediction results are more accurate.
- disk failure prediction model B is always used to predict disk failures, but in some cases, disk failure prediction model A is actually used to predict disk failures, and its disk failure prediction results are more accurate.
- determining the operating status of each disk in the selected area specifically comprises the following steps:
- An operation status determination module used to determine the operation status of each disk in the selected area, wherein the operation status includes a normal state and a fault state;
- a ratio value calculation module used for taking a disk in a normal operation state as a first disk, taking a disk in a fault operation state as a second disk, and calculating a ratio value between the first disk and the second disk;
- the operating status determination module specifically includes:
- the first prediction model is an integrated learning prediction model in which a neural network model and a time series network model are set in parallel; and the second prediction model is a deep learning prediction model in which a neural network model and a time series network model are set in series.
- the positioning module is used to determine the actual position of the corresponding first disk through the positioning data when the disk failure prediction result is that the disk is about to fail.
- FIG1 is a flow chart of a disk status monitoring and fault prediction method provided in an embodiment of the present specification
- FIG2 is a schematic diagram of the structure of a disk status monitoring and fault prediction device provided in an embodiment of the present specification
- FIG. 3 is a schematic diagram of the structure of the device provided in the embodiment of this specification.
- L1 Determine the operating status of each disk in the selected area, where the operating status includes a normal status and a fault status.
- Step L1 needs to first determine the operating status of each disk, and the operating status of each disk can be displayed through the display interface, where a disk in a normal state can be displayed in green, and a disk in a fault state can be displayed in red.
- the size of the selected area can be adjusted and determined according to actual conditions.
- the normal value of the attribute to be detected corresponding to the maximum read/write rate may be 30MB/s, and the normal value of the attribute to be detected corresponding to the available storage space may be 2GB.
- the maximum read/write rate in the state determination data obtained by disk No. 1 is 5MB/s, and the available storage space is 2GB.
- the maximum read/write rate in the state determination data obtained by disk No. 2 is 30MB/s, and the available storage space is 1GB.
- the maximum read/write rate in the state determination data obtained by disk No. 3 is 30MB/s, and the available storage space is 2GB.
- the operating status of the corresponding disk is determined to be a normal status; when at least one of the attributes to be detected of the disk is different from the corresponding normal value of the attribute to be detected, the operating status of the corresponding disk is determined to be a fault status.
- the operation status of disk 1 is determined to be a faulty state. Because the storage space (1GB) of disk 2 is different from the corresponding normal value of the attribute to be detected (2GB), the operation status of disk 2 is determined to be a faulty state. Because the maximum read/write rate and storage space of disk 3 are the same as the corresponding normal value of the attribute to be detected, the operation status of disk 3 is determined to be a normal state.
- a disk in a normal operating state is taken as a first disk
- a disk in a faulty operating state is taken as a second disk
- a ratio value of the first disk to the second disk is calculated.
- the ratio of the first disk to the second disk is 4/16.
- the ratio of the first disk to the second disk is The value is 11/9.
- the ratio of the first disk to the second disk is 18/2.
- the first prediction model is an integrated learning prediction model in which a neural network model and a time series network model are set in parallel;
- the second prediction model is a deep learning prediction model in which a neural network model and a time series network model are set in series.
- the specific type of the neural network model is not limited, for example, it can be one of BP neural network, Hopfield network, ART network or Kohonen network.
- the specific type of the temporal network model is also not limited, for example, it can be one of RNN network, LSTM network or GRU network.
- an integrated learning prediction model i.e., the first prediction model
- the first prediction model By setting the neural network model and the time series network model in parallel, an integrated learning prediction model (i.e., the first prediction model) can be obtained.
- a deep learning prediction model i.e., the second prediction model
- the disk failure prediction model After the disk failure prediction model is determined, it is only necessary to perform a failure prediction on the selected first disk. For example, in step L2, the ratio of the first disk to the second disk is 18/2 (wherein disk No. 1 and disk No. 2 are the second disks, and the remaining disks are the first disks), then in step L3, the first prediction model is used as the disk failure prediction model. Assuming that in this embodiment, it is currently necessary to perform a failure prediction on disk No. 3, it is only necessary to perform a failure prediction on disk No. 3 using the first prediction model.
- the disk failure prediction result of disk No. 3 is that the disk is about to fail
- the corresponding disk failure prediction result can be sent to the relevant staff so that the staff can take solutions in advance; it can also be displayed through the display interface, and the disk originally displayed in green can be changed to be displayed in orange.
- this embodiment also includes the following steps:
- the default fault prediction mode in this embodiment is the comprehensive prediction mode, that is, if the fault prediction mode is not selected, the comprehensive prediction mode is used.
- the first prediction cycle after the disk failure prediction start button is activated, steps L1 to L4 are performed for the first time.
- step L4 assuming that the ratio of the first disk to the second disk is 18/2 at this time, wherein the operating status of disks 1 and 2 is a failure status, and the operating status of disks 3 to 20 is a normal status, and the first prediction model is used as the disk failure prediction model
- the current mode is a comprehensive prediction mode
- the first prediction model is used to predict the failure of disk 3
- the first prediction model is used to predict the failure of disk 4
- the first prediction model is used to predict the failure of disk 4... until the failure prediction of disk 20 is completed, at which time the first prediction cycle ends.
- a second prediction cycle may be performed after a certain period of time.
- the second prediction cycle that is, the steps L1 to L4 are performed for the second time, and the fault prediction mode is changed to the partial prediction mode before the step L4 (the operator can select it on the display interface).
- the step L4 assuming that the ratio of the first disk to the second disk is 4/16, among which the operation status of the disk No. 5, the disk No. 7, the disk No. 9 and the disk No. 18 is normal, and the operation status of the remaining disks is faulty, and the first prediction model is used as the disk fault prediction model
- the current prediction mode is partial, the operator needs to select the first disk for which fault prediction is required on the display interface (i.e., determine the first disk to be predicted), assuming that disks 5 and 9 are selected.
- the first prediction model is used to predict the fault of disk 5, and then the first prediction model is used to predict the fault of disk 9.
- the second prediction cycle ends.
- the third prediction cycle may be performed after a certain period of time.
- the third prediction cycle that is, the third time to perform steps L1 to L4, if the fault prediction mode is not selected, then the fault prediction mode is changed back to the default comprehensive prediction mode.
- step L4 assuming that the ratio of the first disk to the second disk is 11/9 at this time, wherein the operation status of disks 5 to 7, disks 9 to 13, and disks 18 to 20 is normal, and the operation status of the remaining disks is a faulty state, and the second prediction model is used as the disk fault prediction model
- the fault prediction of disk 5 is first performed by the second prediction model, and then the fault prediction of disk 6 is performed by the second prediction model, and then the fault prediction of disk 7 is performed by the second prediction model.
- the fault prediction of disk 7 stops and enters the partial prediction mode.
- the operator first selects the first disk to be predicted on the display interface, assuming that disk 10 is selected. After the first disk to be predicted is selected, the fault prediction of disk 10 is performed by the second prediction model. When the prediction of disk No. 10 failure is completed, the third prediction cycle ends.
- the fourth prediction cycle will be performed again, and this cycle will continue until the disk failure prediction end button is pressed.
- the disk failure prediction needs to be performed again, just press the disk failure prediction start button again. If you suddenly want to perform disk failure prediction immediately during the interval waiting period, just press the disk failure prediction start button again.
- the disk failure prediction method of this embodiment can, under normal circumstances, automatically and regularly (for example, every hour) predict failures of disks in the selected area; under special circumstances (for example, if you want to immediately predict failures of a first disk), you only need to change the failure prediction mode from the full prediction mode to the partial prediction mode, and then select the corresponding first disk for failure prediction.
- This method can meet various usage requirements and is easy to operate.
- this embodiment also includes the following steps:
- Step M1 Obtain the fault prediction data and positioning data of the first disk.
- Step M1 is arranged after step L2 and before step L4.
- the fault prediction data includes SMART (Self-Monitoring Analysis and Reporting Technology) data. This embodiment does not limit the specific training method of the disk fault prediction model.
- SMART Self-Monitoring Analysis and Reporting Technology
- This embodiment does not limit the specific training method of the disk fault prediction model.
- Step M2 When the disk failure prediction result is that the disk is about to fail, the actual location of the corresponding disk is determined by the positioning data. Step M2 is arranged after step L2.
- the positioning data includes information such as the disk number and geographical location.
- the disk failure prediction result is that the disk is normal, the next disk is directly predicted to fail.
- the disk failure prediction result is that the disk is about to fail, the prediction result of the corresponding first disk and the disk number, geographical location and other information are sent to the relevant staff, so that the staff can quickly find the corresponding first disk when they know that a certain first disk is about to fail.
- the disk status monitoring and fault prediction method of this embodiment uses different disk failure prediction models to predict disk failures when the ratio of normal disks to faulty disks is different. Compared with always using the same disk failure prediction model, the disk failure prediction result is more accurate.
- Embodiment 2 A disk status monitoring and fault prediction device, used to execute the disk status monitoring and fault prediction method described in Embodiment 1, as shown in FIG2 , comprises: an operation status determination module, a ratio value calculation module, a prediction model determination module and a disk fault prediction module.
- the operation status determination module is used to determine the operation status of each disk in the selected area, and the operation status includes a normal state and a fault state.
- the ratio value calculation module is used to take the disk whose operation status is a normal state as the first disk, and the disk whose operation status is a fault state as the second disk, and calculate the ratio value of the first disk and the second disk.
- the prediction model determination module is used to determine the disk failure prediction model according to the ratio value: when the ratio value is less than the first preset ratio threshold or greater than the second preset ratio threshold, the first prediction model is used as the disk failure prediction model; when the ratio value is greater than or equal to the first preset ratio threshold and less than or equal to the second preset ratio threshold, the second prediction model is used as the disk failure prediction model.
- the disk failure prediction module is used to perform failure prediction on the selected first disk through the disk failure prediction model to obtain a disk failure prediction result.
- the operation status determination module specifically includes: a data acquisition unit, an attribute comparison unit and an operation status determination unit.
- the data acquisition unit is used to acquire the state determination data of the disk, and the state determination data of each disk includes at least one attribute to be detected.
- the attribute comparison unit is used to compare the attribute to be detected of each disk with the corresponding normal value of the attribute to be detected.
- the operation status determination unit is used to determine that the operation status of the corresponding disk is a normal state when each attribute to be detected of the disk is the same as the corresponding normal value of the attribute to be detected; when at least one attribute to be detected of the disk is different from the corresponding normal value of the attribute to be detected, the operation status of the corresponding disk is determined to be a fault state.
- the first prediction model is an integrated learning prediction model in which a neural network model and a time series network model are set in parallel;
- the second prediction model is a deep learning prediction model in which a neural network model and a time series network model are set in series.
- the disk status monitoring and fault prediction device of this embodiment further includes: a prediction mode determination module.
- the prediction mode determination module is used to determine the fault prediction mode of the disk, and the fault prediction mode includes a comprehensive prediction mode and a partial prediction mode;
- the fault prediction mode is a comprehensive prediction mode, fault prediction is performed on all the first disks in sequence according to the order of the first disks;
- the fault prediction mode is a partial prediction mode, the first disk to be predicted is determined, and fault prediction is performed on the first disk.
- the disk status monitoring and fault prediction device of this embodiment further includes: a data acquisition module and a positioning module.
- the data acquisition module is used to acquire the fault prediction data and positioning data of the first disk, and the disk fault prediction model performs fault prediction on the selected first disk through the fault prediction data;
- the positioning module is used to determine the actual position of the corresponding first disk through the positioning data when the disk fault prediction result is that the disk is about to fail.
- the disk status monitoring and fault prediction system of this embodiment uses different disk failure prediction models to predict disk failures when the ratio of normal disks to faulty disks is different. Compared with always using the same disk failure prediction model, the disk failure prediction results are more accurate.
- Embodiment 3 A device comprises a memory for storing a program; and a processor for running the program stored in the memory to execute the disk status monitoring and fault prediction method described in Embodiment 1.
- the device of this embodiment may include an input device, an input interface, a processor, a memory, an output interface, and an output device.
- the input interface, the processor, the memory, and the output interface are interconnected through a bus, and the input device and the output device are connected to the bus through the input interface and the output interface, respectively, and then connected to other components of the device of this embodiment.
- the input device is used to receive input information from the outside and transmit the input information to the processor through the input interface; the processor processes the input information based on the computer executable instructions stored in the memory to generate output information, and temporarily or permanently stores the output information in the memory, and then transmits the output information to the output device through the output interface; the output device outputs the output information to the outside of the device of this embodiment for user use.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Software Systems (AREA)
- Debugging And Monitoring (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
一种磁盘状态监控及故障预测方法、装置及设备,涉及磁盘故障预测技术领域。方法包括以下步骤:确定所选区域内每一磁盘的运行状态,运行状态包括正常状态和故障状态;将运行状态为正常状态的磁盘作为第一磁盘,将运行状态为故障状态的磁盘作为第二磁盘,并计算第一磁盘与第二磁盘的比例值;根据比例值确定磁盘故障预测模型:当比例值小于第一预设比例阈值或大于第二预设比例阈值时,将第一预测模型作为磁盘故障预测模型。在正常磁盘与故障磁盘的比例不同时,采用不同的磁盘故障预测模型来对磁盘进行故障预测,使得磁盘的故障预测结果更为准确。
Description
本说明书实施例主要涉及磁盘故障预测技术领域,具体为一种磁盘状态监控及故障预测方法、装置及设备。
为了存储和处理大规模的数据,数据中心聚集着数以万计的服务器以及存储设备,磁盘故障直接影响着服务器以及存储设备的运行状况,如果不能及时发现有可能导致数据丢失,进而造成极大的经济损失。现有技术通常使用磁盘故障预测模型来预测磁盘故障,并通过优化磁盘故障预测模型的算法来提高磁盘故障预测结果的准确度。
然而发明人发现,如果一直使用同一个故障预测模型来预测磁盘故障,会存在在某些情况下磁盘故障预测结果准确度并不是很高的问题。例如,如果一直使用磁盘故障预测模型A来预测磁盘故障,但是在某一些情况下,其实使用磁盘故障预测模型B来预测磁盘故障,其磁盘故障预测结果更为准确。又例如,一直使用磁盘故障预测模型B来预测磁盘故障,但是在某一些情况下,其实使用磁盘故障预测模型A来预测磁盘故障,其磁盘故障预测结果更为准确。
发明内容
本说明书实施例针对现有技术存在的问题,提出了一种磁盘状态监控及故障预测方法、装置及设备,用于提高磁盘故障预测结果的准确度。
在第一方面,本说明书实施例提供了一种磁盘状态监控及故障预测方法,包括以下步骤:
确定所选区域内每一磁盘的运行状态,所述运行状态包括正常状态和故障状态;
将运行状态为正常状态的磁盘作为第一磁盘,将运行状态为故障状态的磁盘作为第二磁盘,并计算所述第一磁盘与所述第二磁盘的比例值;
根据所述比例值确定磁盘故障预测模型:当所述比例值小于第一预设比例阈值或大于第二预设比例阈值时,将第一预测模型作为磁盘故障预测模型;当所述比例值大于等于第一预设比例阈值且小于等于第二预设比例阈值时,将第二预测模型作为磁盘故障预测模型;
通过所述磁盘故障预测模型对所选第一磁盘进行故障预测以得到磁盘故障预测结果。
作为优选,确定所选区域内每一磁盘的运行状态具体包括以下步骤:
获取磁盘的状态确定数据,每一所述磁盘的状态确定数据均包括至少一个待检测属性;
将每一所述磁盘的待检测属性与对应的待检测属性常态值进行比对;
当所述磁盘的每一所述待检测属性均与对应的待检测属性常态值相同时,判定对应磁盘的运行状态为正常状态;当所述磁盘有至少一个所述待检测属性与对应的待检测属性常态值不相同时,判定对应磁盘的运行状态为故障状态。
作为优选,所述第一预测模型为神经网络模型和时序网络模型并行设置的集成学习式预测模型;所述第二预测模型为神经网络模型和时序网络模型串行设置的深度学习式预测模型。
作为优选,还包括以下步骤:
确定磁盘的故障预测模式,所述故障预测模式包括全面预测模式和部分预测模式;当所述故障预测模式为全面预测模式时,根据所述第一磁盘的排序先后依次对全部所述第一磁盘进行故障预测;当所述故障预测模式为部分预测模式时,确定待预测的第一磁盘,并对所述第一磁盘进行故障预测。
作为优选,还包括以下步骤:
获取所述第一磁盘的故障预测数据和定位数据,所述磁盘故障预测模型通过所述故障预测数据对所选第一磁盘进行故障预测;
当磁盘故障预测结果为磁盘将要故障时,通过所述定位数据确定对应第一磁盘的实际位置。在第二方面,本说明书实施例提供了一种磁盘状态监控及故障预测装置,包括:
运行状态确定模块,用于确定所选区域内每一磁盘的运行状态,所述运行状态包括正常状态和故障状态;
比例值计算模块,用于将运行状态为正常状态的磁盘作为第一磁盘,将运行状态为故障状态的磁盘作为第二磁盘,并计算所述第一磁盘与所述第二磁盘的比例值;
预测模型确定模块,用于根据所述比例值确定磁盘故障预测模型:当所述比例值小于第一预设比例阈值或大于第二预设比例阈值时,将第一预测模型作为磁盘故障预测模型;当所述比例值大于等于第一预设比例阈值且小于等于第二预设比例阈值时,将第二预测模型作为磁盘故障预测模型;
磁盘故障预测模块,用于通过所述磁盘故障预测模型对所选第一磁盘进行故障预测以得到磁盘故障预测结果。
作为优选,所述运行状态确定模块具体包括:
数据获取单元,用于获取磁盘的状态确定数据,每一所述磁盘的状态确定数据均包括至少一个待检测属性;
属性比对单元,用于将每一所述磁盘的待检测属性与对应的待检测属性常态值进行比对;
运行状态确定单元,用于当所述磁盘的每一所述待检测属性均与对应的待检测属性常态值相
同时,判定对应磁盘的运行状态为正常状态;当所述磁盘有至少一个所述待检测属性与对应的待检测属性常态值不相同时,判定对应磁盘的运行状态为故障状态。
作为优选,所述第一预测模型为神经网络模型和时序网络模型并行设置的集成学习式预测模型;所述第二预测模型为神经网络模型和时序网络模型串行设置的深度学习式预测模型。
作为优选,还包括:
预测模式确定模块,用于确定磁盘的故障预测模式,所述故障预测模式包括全面预测模式和部分预测模式;当所述故障预测模式为全面预测模式时,根据所述第一磁盘的排序先后依次对全部所述第一磁盘进行故障预测;当所述故障预测模式为部分预测模式时,确定待预测的第一磁盘,并对所述第一磁盘进行故障预测;
数据获取模块,用于获取所述第一磁盘的故障预测数据和定位数据,所述磁盘故障预测模型通过所述故障预测数据对所选第一磁盘进行故障预测;
定位模块,用于当磁盘故障预测结果为磁盘将要故障时,通过所述定位数据确定对应第一磁盘的实际位置。
在第三方面,本说明书实施例提供了一种设备,包括:存储器,用于存储程序;处理器,用于运行所述存储器中存储的所述程序,以执行本说明书实施例第一方面的磁盘状态监控及故障预测方法。
本说明书实施例,在正常磁盘(即第一磁盘)与故障磁盘(即第二磁盘)的比例不同的大环境下,采用不同的磁盘故障预测模型来对磁盘进行故障预测,具体的,当第一磁盘与第二磁盘的数量相差较大时使用第一预测模型进行磁盘故障预测,当第一磁盘与第二磁盘的数量相差不大时使用第二预测模型进行磁盘故障预测,相比一直使用同一个磁盘故障预测模型,使得磁盘的故障预测结果更为准确。
图1为本说明书实施例提供的磁盘状态监控及故障预测方法的流程示意图;
图2为本说明书实施例提供的磁盘状态监控及故障预测装置的结构示意图;
图3为本说明书实施例提供的设备的结构示意图。
下面将参照附图更详细地描述本说明书的实施例。
实施例1:一种磁盘状态监控及故障预测方法,如图1所示,包括以下步骤:
L1.确定所选区域内每一磁盘的运行状态,所述运行状态包括正常状态和故障状态。
假设本实施例中所选区域内的磁盘有20个,分别为1号磁盘、2号磁盘、3号磁盘......以及20号磁盘。步骤L1需要先确定每一个磁盘的运行状态,且每一个磁盘的运行状态可以通过显示界面显示,其中,正常状态的磁盘可以通过绿色显示,故障状态的磁盘可以通过红色显示。所选区域的大小可根据实际情况调节确定。
进一步的,本实施例中,步骤L1具体可以包括以下步骤:
L11.获取磁盘的状态确定数据,每一所述磁盘的状态确定数据均包括至少一个待检测属性。一个磁盘可以有多种待检测属性,例如,读写权限,最大读写速率,物理位置,可用存储空间等。本实施例可以先确定好将哪些待检测属性作为状态确定数据。例如,本实施例可以最大读写速率和可用存储空间作为状态确定数据。
L12.将每一所述磁盘的待检测属性与对应的待检测属性常态值进行比对。
本实施例中,与最大读写速率对应的待检测属性常态值可以是30MB/s,与可用存储空间对应的待检测属性常态值可以是2GB。假设本实施例中,1号磁盘获取到的状态确定数据中的最大读写速率是5MB/s,可用存储空间为2GB。2号磁盘获取到的状态确定数据中的最大读写速率是30MB/s,可用存储空间为1GB。3号磁盘获取到的状态确定数据中的最大读写速率是30MB/s,可用存储空间为2GB。
L13.当所述磁盘的每一所述待检测属性均与对应的待检测属性常态值相同时,判定对应磁盘的运行状态为正常状态;当所述磁盘有至少一所述待检测属性与对应的待检测属性常态值不相同时,判定对应磁盘的运行状态为故障状态。
因为1号磁盘的最大读写速率(5MB/s)与对应的待检测属性常态值(30MB/s)不同,所以判定1号磁盘的运行状态为故障状态。因为2号磁盘的存储空间(1GB)与对应的待检测属性常态值(2GB)不同,所以判定2号磁盘的运行状态为故障状态。因为3号磁盘的最大读写速率和存储空间均与对应的待检测属性常态值相同,所以判定3号磁盘的运行状态为正常状态。
本实施例通过步骤L1先确定哪些磁盘是正常的,哪些磁盘是已经故障了。只有当磁盘目前是正常的,才需要进一步进行故障预测,来确定哪些磁盘将要故障,进而提前采取解决措施,避免因磁盘后期故障而导致数据丢失。
L2.将运行状态为正常状态的磁盘作为第一磁盘,将运行状态为故障状态的磁盘作为第二磁盘,并计算所述第一磁盘与所述第二磁盘的比例值。
当第一磁盘的数量为4个,第二磁盘的数量为16个时,第一磁盘与第二磁盘的比例值为4/16。当第一磁盘的数量为11个,第二磁盘的数量为9个时,第一磁盘与第二磁盘的比例
值为11/9。当第一磁盘的数量为18个,第二磁盘的数量为2个时,第一磁盘与第二磁盘的比例值为18/2。
L3.根据所述比例值确定磁盘故障预测模型:当所述比例值小于第一预设比例阈值或大于第二预设比例阈值时,将第一预测模型作为磁盘故障预测模型;当所述比例值大于等于第一预设比例阈值且小于等于第二预设比例阈值时,将第二预测模型作为磁盘故障预测模型。
假设本实施例中的第一预设比例阈值为5/15,第二预设比例阈值为15/5。当比例值为4/16时,小于第一预设比例阈值,使用第一预测模型作为磁盘故障预测模型。当比例值为18/2时,大于第二预设比例阈值,使用第一预测模型作为磁盘故障预测模型。当比例值为11/9时,大于等于第一预设比例阈值且小于等于第二预设比例阈值,使用第二预测模型作为磁盘故障预测模型。
发明人发现,当第一磁盘与第二磁盘的数量相差比较大(即比例值小于第一预设比例阈值或大于第二预设比例阈值)时,使用第一预测模型进行磁盘故障预测的准确度相对使用第二预测模型进行磁盘故障预测的准确度更高,因此,本实施例在第一磁盘与第二磁盘的比例值小于第一预设比例阈值或大于第二预设比例阈值时,将第一预测模型作为磁盘故障预测模型。当第一磁盘与第二磁盘的数量相差不大(即比例值大于等于第一预设比例阈值且小于等于第二预设比例阈值)时,使用第二预测模型进行磁盘故障预测的准确度相对使用第一预测模型进行磁盘故障预测的准确度更高,因此,本实施例在第一磁盘与第二磁盘的比例值大于等于第一预设比例阈值且小于等于第二预设比例阈值时,将第二预测模型作为磁盘故障预测模型。
进一步的,本实施例中,所述第一预测模型为神经网络模型和时序网络模型并行设置的集成学习式预测模型;所述第二预测模型为神经网络模型和时序网络模型串行设置的深度学习式预测模型。
本实施例中,神经网络模型的具体类型不做限定,例如可以是BP神经网络、Hopfield网络、ART网络或Kohonen网络中的其中一种。时序网络模型的具体类型也不做限定,例如可以是RNN网络、LSTM网络或GRU网络中的其中一种。
将神经网络模型和时序网络模型并行设置,能够得到集成学习式预测模型(即第一预测模型),发明人发现,当第一磁盘与第二磁盘的数量相差比较大时,通过集成学习式预测模型对磁盘故障进行预测,其磁盘故障预测结果的准确度更高。将神经网络模型和时序网络模型串行设置,能够得到深度学习式预测模型(即第二预测模型),发明人发现,当第一磁盘与第二磁盘的数量相差不大时,通过深度学习式预测模型对磁盘故障进行预测,其磁盘故障预
测结果的准确度更高。
本实施例在不同的磁盘正负比(即正常与故障比)大环境下,采用不同的磁盘故障预测模型来对第一磁盘进行故障预测,相比一直使用同一个磁盘故障预测模型,使得磁盘的故障预测结果更为准确。
L4.通过所述磁盘故障预测模型对所选第一磁盘进行故障预测以得到磁盘故障预测结果。
当磁盘故障预测模型确定好之后,只需要对所选第一磁盘进行故障预测即可。例如,步骤L2中,第一磁盘与第二磁盘的比例值为18/2(其中,1号磁盘和2号磁盘为第二磁盘,其余磁盘为第一磁盘),那么步骤L3中,使用第一预测模型作为磁盘故障预测模型。假设本实施例,当前需要对3号磁盘进行故障预测,那么只需要通过第一预测模型对3号磁盘进行故障预测即可。当3号磁盘的磁盘故障预测结果为磁盘要将故障时,可以将对应的磁盘故障预测结果发送给相关工作人员,以使工作人员能够提前采取解决措施;还可以通过显示界面显示,将原本显示为绿色的磁盘改为通过橙色显示。
进一步的,本实施例还包括以下步骤:
S1.确定磁盘的故障预测模式,所述故障预测模式包括全面预测模式和部分预测模式;当所述故障预测模式为全面预测模式时,根据所述第一磁盘的排序先后依次对全部所述第一磁盘进行故障预测;当所述故障预测模式为部分预测模式时,确定待预测的第一磁盘,并对所述第一磁盘进行故障预测。
本实施例默认的故障预测模式为全面预测模式,即如果没有选择故障预测模式,那么就是采用全面预测模式。
第1次预测循环:启动磁盘故障预测开始按钮后,第一次进行步骤L1至L4,在进行步骤L4时(假设,此时第一磁盘与第二磁盘的比例值为18/2,其中,1号磁盘和2号磁盘的运行状态为故障状态,3号磁盘至20号磁盘的运行状态为正常状态,第一预测模型作为磁盘故障预测模型),因为当前为全面预测模式,所以先通过第一预测模型对3号磁盘进行故障预测,然后通过第一预测模型对4号磁盘进行故障预测,接着通过第一预测模型对4号磁盘进行故障预测......直到20号磁盘故障预测结束,此时第1次预测循环结束。
本实施例可以间隔一段时间再进行第2次预测循环。
第2次预测循环:即第二次进行步骤L1至L4,且在进行步骤L4前将故障预测模式改为部分预测模式(操作人员可以在显示界面上进行选择)。在进行步骤L4时(假设,此时第一磁盘与第二磁盘的比例值为4/16,其中,5号磁盘、7号磁盘、9号磁盘及18号磁盘的运行状态为正常状态,其余磁盘的运行状态为故障状态,第一预测模型作为磁盘故障预测模型),
因为当前为部分预测模式,所以需要操作人员先在显示界面上选择需要进行故障预测的第一磁盘(即确定待预测的第一磁盘),假设选择5号磁盘和9号磁盘。在待预测的第一磁盘选择好之后(操作人员按下选择完成按钮),先通过第一预测模型对5号磁盘进行故障预测,然后通过第一预测模型对9号磁盘进行故障预测。当9号磁盘故障预测结束后,第2次预测循环结束。
本实施例可以间隔一段时间再进行第3次预测循环。
第3次预测循环:即第三次进行步骤L1至L4,如果没有对故障预测模式进行选择,那么故障预测模式又重新变为默认的全面预测模式。在进行步骤L4时(假设,此时第一磁盘与第二磁盘的比例值为11/9,其中,5号磁盘至7号磁盘、9号磁盘至13号磁盘、18号磁盘至20号磁盘的运行状态为正常状态,其余磁盘的运行状态为故障状态,第二预测模型作为磁盘故障预测模型),因为当前为全面预测模式,所以先通过第二预测模型对5号磁盘进行故障预测,然后通过第二预测模型对6号磁盘进行故障预测,接着通过第二预测模型对7号磁盘进行故障预测......。如果在对7号磁盘进行故障预测时,操作人员在显示界面上将故障预测模式改为部分预测模式,那么7号磁盘停止进行故障预测并进入部分预测模式。操作人员先在显示界面上选择需要进行故障预测的第一磁盘,假设选择10号磁盘。在待预测的第一磁盘选择好之后,通过第二预测模型对10号磁盘进行故障预测。当10号磁盘故障预测结束后,第3次预测循环结束。
间隔一段时间后再进行第4次预测循环,如此不断循环,直到按下磁盘故障预测结束按钮。当需要再进行磁盘故障预测时,再次按下磁盘故障预测开始按钮即可。如果在间隔等待期间,突然想马上进行磁盘故障预测,只需要重新按下磁盘故障预测开始按钮即可。
本实施例的磁盘故障预测方法,在普通情况下,能够自动定期(例如,每隔1个小时)对所选区域内的磁盘进行故障预测;在特殊情况下(例如,想马上对某一第一磁盘进行故障预测),则只需要将故障预测模式从全面预测模式改为部分预测模式,然后选择对应的第一磁盘进行故障预测即可,能够满足各种使用需求,且操作简便。
进一步的,本实施例还包括以下步骤:
M1.获取所述第一磁盘的故障预测数据和定位数据。步骤M1设置在步骤L2之后且位于步骤L4之前。故障预测数据包括SMART(Self-Monitoring Analysis and Reporting Technology,自我监测、分析及报告技术)数据。本实施例对磁盘故障预测模型的具体训练方式不做限定,在需要对某一第一磁盘进行故障预测时,只需要将获取到SMART数据(具体为磁盘内部盘片、马达、内部温度、磁头等属性的状态数据)输入磁盘故障预测模型即可,所述磁盘故障
预测模型通过所述故障预测数据对所选第一磁盘进行故障预测。
M2.当磁盘故障预测结果为磁盘将要故障时,通过所述定位数据确定对应磁盘的实际位置。步骤M2设置在步骤L2之后。定位数据包括磁盘编号、地理位置等信息。当磁盘故障预测结果为磁盘正常时,就直接对下一个磁盘进行故障预测。当磁盘故障预测结果为磁盘将要故障时,就将对应第一磁盘的预测结果以及磁盘编号、地理位置等信息一同发送给相关工作人员,便于工作人员在知道某一第一磁盘将要故障时,能够快速找到对应的第一磁盘。
本实施例的磁盘状态监控及故障预测方法,在正常磁盘与故障磁盘的比例不同时,采用不同的磁盘故障预测模型来对磁盘进行故障预测,相比一直使用同一个磁盘故障预测模型,使得磁盘的故障预测结果更为准确。
实施例2:一种磁盘状态监控及故障预测装置,用于执行实施例1所述的磁盘状态监控及故障预测方法,如图2所示,包括:运行状态确定模块,比例值计算模块,预测模型确定模块和磁盘故障预测模块。
运行状态确定模块用于确定所选区域内每一磁盘的运行状态,所述运行状态包括正常状态和故障状态。比例值计算模块用于将运行状态为正常状态的磁盘作为第一磁盘,将运行状态为故障状态的磁盘作为第二磁盘,并计算所述第一磁盘与所述第二磁盘的比例值。预测模型确定模块用于根据所述比例值确定磁盘故障预测模型:当所述比例值小于第一预设比例阈值或大于第二预设比例阈值时,将第一预测模型作为磁盘故障预测模型;当所述比例值大于等于第一预设比例阈值且小于等于第二预设比例阈值时,将第二预测模型作为磁盘故障预测模型。磁盘故障预测模块用于通过所述磁盘故障预测模型对所选第一磁盘进行故障预测以得到磁盘故障预测结果。
所述运行状态确定模块具体包括:数据获取单元,属性比对单元和运行状态确定单元。数据获取单元用于获取磁盘的状态确定数据,每一所述磁盘的状态确定数据均包括至少一个待检测属性。属性比对单元用于将每一所述磁盘的待检测属性与对应的待检测属性常态值进行比对。运行状态确定单元用于当所述磁盘的每一所述待检测属性均与对应的待检测属性常态值相同时,判定对应磁盘的运行状态为正常状态;当所述磁盘有至少一个所述待检测属性与对应的待检测属性常态值不相同时,判定对应磁盘的运行状态为故障状态。
所述第一预测模型为神经网络模型和时序网络模型并行设置的集成学习式预测模型;所述第二预测模型为神经网络模型和时序网络模型串行设置的深度学习式预测模型。
本实施例的磁盘状态监控及故障预测装置还包括:预测模式确定模块。预测模式确定模块用于确定磁盘的故障预测模式,所述故障预测模式包括全面预测模式和部分预测模式;当所述
故障预测模式为全面预测模式时,根据所述第一磁盘的排序先后依次对全部所述第一磁盘进行故障预测;当所述故障预测模式为部分预测模式时,确定待预测的第一磁盘,并对所述第一磁盘进行故障预测。
本实施例的磁盘状态监控及故障预测装置还包括:数据获取模块和定位模块。数据获取模块用于获取所述第一磁盘的故障预测数据和定位数据,所述磁盘故障预测模型通过所述故障预测数据对所选第一磁盘进行故障预测;定位模块用于当磁盘故障预测结果为磁盘将要故障时,通过所述定位数据确定对应第一磁盘的实际位置。
本实施例的磁盘状态监控及故障预测系统,在正常磁盘与故障磁盘的比例不同时,采用不同的磁盘故障预测模型来对磁盘进行故障预测,相比一直使用同一个磁盘故障预测模型,使得磁盘的故障预测结果更为准确。
实施例3:一种设备,包括存储器,用于存储程序;处理器,用于运行所述存储器中存储的所述程序,以执行实施例1所述的磁盘状态监控及故障预测方法。
具体的,如图3所示,本实施例的设备可以包括输入设备、输入接口、处理器、存储器、输出接口、以及输出设备。其中,输入接口、处理器、存储器、以及输出接口通过总线相互连接,输入设备和输出设备分别通过输入接口和输出接口与总线连接,进而与本实施例设备的其他组件连接。
输入设备用于接收来自外部的输入信息,并通过输入接口将输入信息传送到处理器;处理器基于存储器中存储的计算机可执行指令对输入信息进行处理以生成输出信息,并将输出信息临时或者永久地存储在存储器中,然后通过输出接口将输出信息传送到输出设备;输出设备将输出信息输出到本实施例设备的外部供用户使用。
虽然附图中显示了本说明书的某些实施例,然而应当理解的是,本说明书可以通过各种形式来实现,而且不应该被解释为限于这里阐述的实施例,相反提供这些实施例是为了更加透彻和完整地理解本说明书。应当理解的是,本说明书的附图及实施例仅用于示例性作用,并非用于限制本说明书的保护范围。
Claims (10)
- 一种磁盘状态监控及故障预测方法,包括以下步骤:确定所选区域内每一磁盘的运行状态,所述运行状态包括正常状态和故障状态;将运行状态为正常状态的磁盘作为第一磁盘,将运行状态为故障状态的磁盘作为第二磁盘,并计算所述第一磁盘与所述第二磁盘的比例值;根据所述比例值确定磁盘故障预测模型:当所述比例值小于第一预设比例阈值或大于第二预设比例阈值时,将第一预测模型作为磁盘故障预测模型;当所述比例值大于等于第一预设比例阈值且小于等于第二预设比例阈值时,将第二预测模型作为磁盘故障预测模型;通过所述磁盘故障预测模型对所选第一磁盘进行故障预测以得到磁盘故障预测结果。
- 根据权利要求1所述的磁盘状态监控及故障预测方法,确定所选区域内每一磁盘的运行状态具体包括以下步骤:获取磁盘的状态确定数据,每一所述磁盘的状态确定数据均包括至少一个待检测属性;将每一所述磁盘的待检测属性与对应的待检测属性常态值进行比对;当所述磁盘的每一所述待检测属性均与对应的待检测属性常态值相同时,判定对应磁盘的运行状态为正常状态;当所述磁盘有至少一个所述待检测属性与对应的待检测属性常态值不相同时,判定对应磁盘的运行状态为故障状态。
- 根据权利要求1所述的磁盘状态监控及故障预测方法,所述第一预测模型为神经网络模型和时序网络模型并行设置的集成学习式预测模型;所述第二预测模型为神经网络模型和时序网络模型串行设置的深度学习式预测模型。
- 根据权利要求1所述的磁盘状态监控及故障预测方法,还包括以下步骤:确定磁盘的故障预测模式,所述故障预测模式包括全面预测模式和部分预测模式;当所述故障预测模式为全面预测模式时,根据所述第一磁盘的排序先后依次对全部所述第一磁盘进行故障预测;当所述故障预测模式为部分预测模式时,确定待预测的第一磁盘,并对所述第一磁盘进行故障预测。
- 根据权利要求1所述的磁盘状态监控及故障预测方法,还包括以下步骤:获取所述第一磁盘的故障预测数据和定位数据,所述磁盘故障预测模型通过所述故障预测数据对所选第一磁盘进行故障预测;当磁盘故障预测结果为磁盘将要故障时,通过所述定位数据确定对应第一磁盘的实际位置。
- 一种磁盘状态监控及故障预测装置,包括:运行状态确定模块,用于确定所选区域内每一磁盘的运行状态,所述运行状态包括正常状态和故障状态;比例值计算模块,用于将运行状态为正常状态的磁盘作为第一磁盘,将运行状态为故障状态的磁盘作为第二磁盘,并计算所述第一磁盘与所述第二磁盘的比例值;预测模型确定模块,用于根据所述比例值确定磁盘故障预测模型:当所述比例值小于第一预设比例阈值或大于第二预设比例阈值时,将第一预测模型作为磁盘故障预测模型;当所述比例值大于等于第一预设比例阈值且小于等于第二预设比例阈值时,将第二预测模型作为磁盘故障预测模型;磁盘故障预测模块,用于通过所述磁盘故障预测模型对所选第一磁盘进行故障预测以得到磁盘故障预测结果。
- 根据权利要求6所述的磁盘状态监控及故障预测装置,所述运行状态确定模块具体包括:数据获取单元,用于获取磁盘的状态确定数据,每一所述磁盘的状态确定数据均包括至少一个待检测属性;属性比对单元,用于将每一所述磁盘的待检测属性与对应的待检测属性常态值进行比对;运行状态确定单元,用于当所述磁盘的每一所述待检测属性均与对应的待检测属性常态值相同时,判定对应磁盘的运行状态为正常状态;当所述磁盘有至少一个所述待检测属性与对应的待检测属性常态值不相同时,判定对应磁盘的运行状态为故障状态。
- 根据权利要求6所述的磁盘状态监控及故障预测装置,所述第一预测模型为神经网络模型和时序网络模型并行设置的集成学习式预测模型;所述第二预测模型为神经网络模型和时序网络模型串行设置的深度学习式预测模型。
- 根据权利要求6所述的磁盘状态监控及故障预测装置,还包括:预测模式确定模块,用于确定磁盘的故障预测模式,所述故障预测模式包括全面预测模式和部分预测模式;当所述故障预测模式为全面预测模式时,根据所述第一磁盘的排序先后依次对全部所述第一磁盘进行故障预测;当所述故障预测模式为部分预测模式时,确定待预测的第一磁盘,并对所述第一磁盘进行故障预测;数据获取模块,用于获取所述第一磁盘的故障预测数据和定位数据,所述磁盘故障预测模型通过所述故障预测数据对所选第一磁盘进行故障预测;定位模块,用于当磁盘故障预测结果为磁盘将要故障时,通过所述定位数据确定对应第一磁盘的实际位置。
- 一种设备,包括:存储器,用于存储程序;处理器,用于运行所述存储器中存储的所述程序,以执行权利要求1-5中任意一权利要求所述的磁盘状态监控及故障预测方法。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310659641.7 | 2023-06-06 | ||
| CN202310659641.7A CN116383016B (zh) | 2023-06-06 | 2023-06-06 | 一种磁盘状态监控及故障预测方法、装置及设备 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2024250628A1 true WO2024250628A1 (zh) | 2024-12-12 |
Family
ID=86975479
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2023/139637 Ceased WO2024250628A1 (zh) | 2023-06-06 | 2023-12-18 | 一种磁盘状态监控及故障预测方法、装置及设备 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN116383016B (zh) |
| WO (1) | WO2024250628A1 (zh) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116383016B (zh) * | 2023-06-06 | 2023-10-10 | 天翼云科技有限公司 | 一种磁盘状态监控及故障预测方法、装置及设备 |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102129397A (zh) * | 2010-12-29 | 2011-07-20 | 深圳市永达电子股份有限公司 | 一种自适应磁盘阵列故障预测方法及系统 |
| US20180060192A1 (en) * | 2016-08-31 | 2018-03-01 | Seagate Technology Llc | Adaptive Failure Prediction Modeling for Detection of Data Storage Device Failures |
| CN108228377A (zh) * | 2017-12-29 | 2018-06-29 | 华中科技大学 | 一种面向磁盘故障检测的smart阈值优化方法 |
| CN109491850A (zh) * | 2018-11-21 | 2019-03-19 | 北京北信源软件股份有限公司 | 一种磁盘故障预测方法及装置 |
| CN111104293A (zh) * | 2018-10-26 | 2020-05-05 | 伊姆西Ip控股有限责任公司 | 用于支持盘故障预测的方法、设备和计算机程序产品 |
| CN116383016A (zh) * | 2023-06-06 | 2023-07-04 | 天翼云科技有限公司 | 一种磁盘状态监控及故障预测方法、装置及设备 |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9535774B2 (en) * | 2013-09-09 | 2017-01-03 | International Business Machines Corporation | Methods, apparatus and system for notification of predictable memory failure |
| CN110471820B (zh) * | 2019-08-05 | 2023-01-17 | 南开大学 | 一种基于循环神经网络的云存储系统磁盘故障预测方法 |
| CN112764953A (zh) * | 2019-10-21 | 2021-05-07 | 伊姆西Ip控股有限责任公司 | 盘故障控制的方法、电子设备和计算机可读存储介质 |
| CN111752775B (zh) * | 2020-05-28 | 2022-11-18 | 苏州浪潮智能科技有限公司 | 一种磁盘故障预测方法和系统 |
| CN115543762A (zh) * | 2022-11-30 | 2022-12-30 | 苏州浪潮智能科技有限公司 | 一种磁盘smart数据扩充方法、系统及电子设备 |
-
2023
- 2023-06-06 CN CN202310659641.7A patent/CN116383016B/zh active Active
- 2023-12-18 WO PCT/CN2023/139637 patent/WO2024250628A1/zh not_active Ceased
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102129397A (zh) * | 2010-12-29 | 2011-07-20 | 深圳市永达电子股份有限公司 | 一种自适应磁盘阵列故障预测方法及系统 |
| US20180060192A1 (en) * | 2016-08-31 | 2018-03-01 | Seagate Technology Llc | Adaptive Failure Prediction Modeling for Detection of Data Storage Device Failures |
| CN108228377A (zh) * | 2017-12-29 | 2018-06-29 | 华中科技大学 | 一种面向磁盘故障检测的smart阈值优化方法 |
| CN111104293A (zh) * | 2018-10-26 | 2020-05-05 | 伊姆西Ip控股有限责任公司 | 用于支持盘故障预测的方法、设备和计算机程序产品 |
| CN109491850A (zh) * | 2018-11-21 | 2019-03-19 | 北京北信源软件股份有限公司 | 一种磁盘故障预测方法及装置 |
| CN116383016A (zh) * | 2023-06-06 | 2023-07-04 | 天翼云科技有限公司 | 一种磁盘状态监控及故障预测方法、装置及设备 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116383016A (zh) | 2023-07-04 |
| CN116383016B (zh) | 2023-10-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110413227A (zh) | 一种硬盘设备的剩余使用寿命在线预测方法和系统 | |
| US8024609B2 (en) | Failure analysis based on time-varying failure rates | |
| WO2024164713A1 (zh) | 存储器健康状态确定方法、装置、电子设备及存储介质 | |
| US8024613B2 (en) | Method and system for managing apparatus performance | |
| US20240296101A1 (en) | Server fault locating method and apparatus, electronic device, and storage medium | |
| CN115248757A (zh) | 一种硬盘健康评估方法和存储设备 | |
| WO2023185767A1 (zh) | 慢盘检测方法、装置、电子设备及存储介质 | |
| US7769562B2 (en) | Method and apparatus for detecting degradation in a remote storage device | |
| JP2016051425A (ja) | ストレージ制御装置およびストレージ制御プログラム | |
| CN108415810B (zh) | 一种硬盘状态监控方法和装置 | |
| CN114528942A (zh) | 工程机械的数据样本库的构建、故障预测方法及工程机械 | |
| CN115982049A (zh) | 性能测试中的异常检测方法、装置和计算机设备 | |
| CN116383016B (zh) | 一种磁盘状态监控及故障预测方法、装置及设备 | |
| US9430318B2 (en) | Health monitoring and recovery for infrastructure devices | |
| CN108322737A (zh) | 摄像帧率测量方法和装置 | |
| EP4425368A1 (en) | Chip internal voltage prediction model generation method, chip internal voltage prediction method, and related apparatuses | |
| CN109684149A (zh) | 一种NVMe硬盘的硬件信息监控方法、装置及设备 | |
| CN117289143A (zh) | 一种故障预测方法、装置、设备、系统和介质 | |
| CN114048085B (zh) | 一种磁盘故障分析方法、装置、设备及可读存储介质 | |
| CN120256243B (zh) | 一种独立监控的存储系统及方法 | |
| CN106610883B (zh) | 一种检测方法及装置 | |
| CN118795794A (zh) | 一种发电厂计算机监控系统的仿真闭环测试系统及方法 | |
| TWI871872B (zh) | 裝置測試系統及裝置測試方法 | |
| CN118535424B (zh) | 一种服务器及其硬件故障预测方法、程序产品和存储介质 | |
| CN119889418B (zh) | 一种固态硬盘定制化测试方法、系统、电子设备及介质 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23940502 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |