A storage device failure prediction method and system
A storage device and fault prediction technology, applied in neural learning methods, detecting faulty computer hardware, using neural network to detect faulty hardware, etc., can solve problems such as difficult to determine labels, unstable models, and difficult to control predictive time in advance. Achieve the effect of reducing false alarm rate and improving fault detection rate
Active Publication Date: 2020-08-14
HUAZHONG UNIV OF SCI & TECH
View PDF1 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] Aiming at the defects of the prior art, the purpose of the present invention is to solve the technical problems that the labels of the training samples of the prior art disk failure prediction method are difficult to determine, resulting in unstable model construction and difficult control of the advance prediction time
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0038] In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.
[0039] The present invention provides a storage device failure prediction method, the method includes the following steps:
[0040]S1. Receive the input minimum lead time LTMIN and maximum lead time LTMAX, and collect the SMART attribute data of N storage devices of the same storage device series at different time points in real time, ensuring that the collected data includes normal storage device data and faulty storage devices data;
[0041] S2. Randomly shuffle the order of all storage devices, and select the j-th storage device according to the order after the scramble;
[0042] S3. SMART att...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a storage device failure prediction method and system, belonging to the technical field of computer storage. Including: S1. collecting the SMART attribute data of N storage devices at different time points; S2. disturbing the order of all storage devices, and selecting the j=1 storage device; 83. the SMART attribute data of each time point of the storage device as Small batches of samples are input into the fault prediction model for training to obtain output results; S4. According to the state of the storage device at time point tn, output results, LTMIN and LTMAX, dynamically adjust the labels and feedback weights of each sample; S5. Calculate the batch The comprehensive loss Loss j ; S6. Select the next storage device, repeat steps S3‑S5 until all storage devices are taken, and calculate the total loss Loss of all storage devices in this period final ; 87. Judging Loss final Whether it converges, if so, get the trained prediction model, and go to step S8, otherwise, go to step S2; S8. Input the current SMART attribute data of the storage device to be predicted into the trained prediction model, and get the prediction result.
Description
technical field [0001] The invention belongs to the technical field of computer storage, and more particularly relates to a storage device failure prediction method and system. Background technique [0002] Due to the low price per storage capacity and mature technology, disks are widely deployed in data centers and are often used in cold data storage, long-term storage, backup storage and other applications. Once the disk fails, it will cause huge data loss if the data is not backed up, and if there is a backup, restoring the data will incur huge overhead, which will easily cause disk and network I / O bursts, affecting Use of Online Business. [0003] Usually, the data center collects the SMART data of the disk and I / O load statistical data, and uses the machine learning model to build a fault prediction model to evaluate the wear degree of the disk, speculate whether the disk will fail in the near future, and analyze the Potentially high-risk disks take fault handling mea...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06F11/22G06N3/04G06N3/063G06N3/08
CPCG06F11/2263G06N3/08G06N3/063G06N3/048G06N3/044
Inventor 冯丹王芳谢燕文张鑫
Owner HUAZHONG UNIV OF SCI & TECH