JP2008084392A - Magnetic disk device, failure prediction device, and method for predicting failure of magnetic disk device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a magnetic disk device, a failure prediction device, and a failure prediction method, by which a failure occurrence timing and a replacing timing due to the change with time of a disk drive can be predicted. <P>SOLUTION: The failure prediction device 30 is provided with a seek testing means 31 for periodically performing a seek test, and a failure sign determination part 36 for calculating and determining the predicted replacing timing tp of a disk drive 20 on the basis of a difference (Tn-Tn-1)/tc between current seek time (Tn) and previous seek time (Tn-1) with respect to a test execution time interval (tc), and failure sign seek time (Tsc)35. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a magnetic disk device, a failure prediction device, and a failure prediction method for a magnetic disk device, and in particular, a magnetic disk device capable of predicting the life of a mechanical part of a movable part and predicting replacement time and failure occurrence time. The present invention relates to a failure prediction device and a failure prediction method for a magnetic disk device.

  The magnetic disk device has a lifetime. In particular, the mechanical parts constituting the movable part of the magnetic disk device have a lifetime due to deterioration due to operating time and aging.

  Conventionally, there is a magnetic disk unit replacement system that detects a failure sign of each disk unit of a magnetic disk storage device and replaces an operation disk unit with a spare disk unit (see, for example, Patent Document 1).

  Patent Document 1 provides a failure monitoring process for monitoring a failure sign, and the failure monitoring process performs a failure sign based on the operation status of the magnetic disk (read / write operation retry count). This failure sign is recognized after the read / write operation fails. Therefore, in the case of the deterioration of the movable part, since the failure sign is recognized when the degree of progress of the deterioration becomes very large, there is almost no remaining life, and there is a disadvantage that the magnetic disk device generates a failure before replacement. is there.

  In addition, some life prediction / management devices for rotating body storage devices accurately predict and manage the remaining life of the rotation body storage device (see, for example, Patent Document 2).

  Patent Document 2 predicts the remaining life by sensing and analyzing rotational noise or vibration generated from a bearing portion of a spindle motor. The voice analysis unit and the vibration analysis unit diagnose the remaining life state in three stages: normal, a state with some sign of life, and abnormality. However, the above technology for predicting the remaining life due to grease deterioration of the bearing part from the rotation sound or vibration and the failure in the actual usage state (seek operation, read / write operation, etc.) There is a disadvantage that there is no correspondence with the error occurrence) and the remaining life state and the failure do not correspond exactly. Also, the rotational sound and vibration of the spindle motor vary in the assembled state at the time of manufacture, and unless it is compared with the individual initial values of the spindle motor, it cannot be accurately analyzed whether it is a change due to secular change. . In addition, the vibration frequency that occurs due to grease deterioration is not constant, and there are generally multiple, and since the relationship between the rotation synchronization frequency level and the grease life is generally low, the analysis of rotational sound and vibration should identify the frequency of interest. Is actually difficult.

  In addition, the failure prediction system for a magnetic disk device can predict the failure occurrence of the magnetic disk device in advance, and can predict the failure occurrence in advance even when the specifications of the manufacturer and model number of the magnetic disk device are various. (For example, refer to Patent Document 3).

  In Patent Document 3, the response time is obtained from the difference between the time when the central processing unit issues a read / write request to the magnetic disk device via the control device and the time when the data is received, and the response time is normal response without retry. Determine whether the time is exceeded. When it is determined that it has exceeded, information related to the magnetic disk device is recorded in the database device, and the progress of damage to the magnetic disk device is determined by statistical analysis. In the above technique, when the central processing unit issues a read / write command, the response time is obtained, and the progress of damage is determined from the response time. There is a drawback that it cannot be predicted.

Japanese Patent Laid-Open No. 09-016340 JP 2004-152336 A Japanese Patent Application Laid-Open No. 2004-118397

  The technique of Patent Document 1 recognizes a failure sign when a read / write operation fails. Therefore, since the failure sign is recognized when the progress of the deterioration of the movable portion becomes very large, there is a problem that the magnetic disk device has a failure before the replacement because there is almost no remaining life.

  The technique of Patent Document 2 has a problem that there is no correspondence between the rotating sound or vibration and the failure (error occurrence) in the actual use state of the rotating body storage device, and the remaining life state and the failure do not correspond accurately. is there.

  The technique of Patent Document 3 has a problem that when the central processing unit does not access the magnetic disk device, a failure of the magnetic disk device cannot be predicted in advance.

  An object of the present invention is to perform a seek test periodically in order to solve the above-mentioned problems, and from the difference (seek time change rate) between the current seek time and the previous seek time with respect to the test execution time interval, the magnetic disk device It is an object of the present invention to provide a magnetic disk device, a failure prediction device, and a failure prediction method capable of recognizing a deterioration state of the disk and performing a predictive diagnosis to predict a failure occurrence time and replacement time.

  A magnetic disk device according to the present invention is a magnetic disk device having a disk drive having a disk medium and a magnetic head for writing / reading information to / from the disk medium, and a failure prediction device for predicting failure of the disk drive. The failure prediction device includes a seek test unit that performs a seek for failure sign diagnosis, a test result storage unit that stores a test result by the seek test unit and an operation time during the test execution, and a failure predictor A failure sign reference time storage unit that sets and stores a failure sign seek time, and a failure sign determination unit that predicts a replacement time from a test result, an operation time, and a trouble sign seek time.

  A magnetic disk device according to the present invention is a magnetic disk device having a disk drive having a disk medium and a magnetic head for writing / reading information to / from the disk medium, and a failure prediction device for predicting failure of the disk drive. The failure prediction device includes a seek test unit that performs a seek for failure occurrence diagnosis, a test result storage unit that stores a test result by the seek test unit and an operation time during the test execution, and a failure determination unit A failure occurrence reference time storage unit that sets and stores a failure occurrence seek time, and a failure occurrence determination unit that predicts a failure time from a test result, an operation time, and a failure occurrence seek time are characterized.

  The seek test means has a timer counter for measuring the seek time.

  The seek test means has means for periodically executing a test and means for repeating a specific seek pattern for a plurality of cycles.

  The seek pattern is characterized by having a random seek pattern.

  The failure predictive seek time is set in advance corresponding to the initial seek time during operation.

  The failure sign determination unit includes a means for predicting a replacement time from a difference in seek time with respect to an operation time interval in which a test is executed.

  The failure occurrence seek time is characterized in that a seek timeout time value is set in advance.

  The failure occurrence determination unit includes means for predicting a failure time from a difference in seek time with respect to an operation time interval at which the test is executed.

  A failure prediction apparatus according to the present invention is a failure prediction apparatus that performs failure prediction of a disk drive including a disk medium and a magnetic head that writes / reads information to / from the disk medium, and performs a seek for failure sign diagnosis. Seek test means, test result storage section for storing test results by seek test means and operation time at the time of test execution, and failure predictive reference time storage section for setting and storing a failure predictive seek time for determining a failure sign And a failure sign determination unit for predicting the replacement time from the test result, the operation time, and the failure sign seek time.

  The failure prediction apparatus according to the present invention is a failure prediction apparatus that performs failure prediction of a disk drive that includes a disk medium and a magnetic head that writes / reads information to / from the disk medium, and performs seek for failure occurrence diagnosis. Seek test means, test result storage section for storing test results by seek test means and operation time at the time of test execution, and failure occurrence reference time storage section for setting and storing a failure occurrence seek time for determining the occurrence of a failure And a failure occurrence determination unit that predicts failure time from the test result, operation time, and failure occurrence seek time.

  The seek test means has a timer counter for measuring the seek time.

  The seek test means has means for periodically executing a test and means for repeating a specific seek pattern for a plurality of cycles.

  The seek pattern is characterized by having a random seek pattern.

  The failure predictive seek time is set in advance corresponding to the initial seek time during operation.

  The failure sign determination unit includes a means for predicting a replacement time from a difference in seek time with respect to an operation time interval in which a test is executed.

  The failure occurrence seek time is characterized in that a seek timeout time value is set in advance.

  The failure occurrence determination unit includes means for predicting a failure time from a difference in seek time with respect to an operation time interval at which the test is executed.

  A failure prediction method for a magnetic disk device according to the present invention includes a disk drive having a disk medium and a magnetic head for writing / reading information to / from the disk medium, and a failure prediction device for predicting a failure of the disk drive. A failure prediction method for a device, wherein the failure prediction device performs a seek for failure sign diagnosis by seek test means, a test result by the seek test means, and an operation time at the time of test execution to a test result storage unit The step of storing, the step of setting the failure predictive seek time for determining the failure predictor and storing it in the failure predictive reference time storage unit, and the replacement timing by the failure predictor determining unit based on the test result, operation time and failure predictive seek time And a step of performing prediction.

  A failure prediction method for a magnetic disk device according to the present invention includes a disk drive having a disk medium and a magnetic head for writing / reading information to / from the disk medium, and a failure prediction device for predicting a failure of the disk drive. A failure prediction method for a device, wherein the failure prediction device performs a seek for failure occurrence diagnosis by a seek test unit, a test result by the seek test unit, and an operation time at the time of test execution to a test result storage unit The failure occurrence determination unit determines the failure time from the step of storing, the step of setting the failure occurrence seek time for determining failure occurrence and storing it in the failure occurrence reference time storage unit, and the test result, operation time, and failure occurrence seek time. And a step of performing prediction.

  The failure sign determination unit has a step of predicting the replacement time from the difference in seek time with respect to the operation time interval at which the test is executed.

  The failure occurrence determination unit includes means for predicting a failure time from a difference in seek time with respect to an operation time interval at which the test is executed.

  In the present invention, a seek test is periodically performed, and the deterioration status of the magnetic disk device is determined based on the difference between the current seek time and the previous seek time with respect to the test execution time interval. It is possible to perform a predictive failure diagnosis corresponding to the above.

  Next, the best mode for carrying out the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a schematic configuration diagram showing a first embodiment of a magnetic disk apparatus 10 according to the present invention.

  Referring to FIG. 1, the magnetic disk device 10 includes a disk drive 20 and a failure prediction device 30.

  The disk drive 20 includes a disk medium 21 and a magnetic head 22 for writing / reading information to / from the disk medium 21.

  The failure prediction apparatus 30 includes seek test means 31 that performs a seek for failure sign diagnosis of the disk drive 20, a test result storage unit 33, a failure sign reference time storage unit 34, and a failure sign determination unit 36.

  The seek test means 31 includes a timer counter 32 that measures a seek time, a means (not shown) that periodically executes a random seek test, and a means (not shown) that repeats a random seek pattern for a plurality of cycles. The seek test means 31 does not accept an external I / O (INPUT / OUTPUT) command during execution of the random seek test, and always measures the seek time under the same conditions.

  The test result storage unit 33 stores the test result (not shown) executed by the seek test means 31 and the operation time (not shown) of the disk drive 20 at the time of the test execution.

  The failure sign reference time storage unit 34 stores a failure sign seek time (Tsc) 35 set in advance for determining a sign of failure. The failure predictive seek time (Tsc) 35 is set corresponding to the initial seek time (To) when the disk drive 20 is operating. The test result storage unit 33 and the failure sign reference time storage unit 34 are provided in a nonvolatile storage unit (not shown).

  The failure sign determination unit 36 performs prediction by calculating the predicted replacement time (tp) of the disk drive 20 from the test result, the operation time, and the failure sign seek time (Tsc) 35.

  Next, the operation of the magnetic disk device 10 configured as described above will be described with reference to FIG. 2 and FIG.

  FIG. 2 is a diagram for explaining prediction by calculating the predicted replacement time (tp) of the disk drive 20, and is a diagram showing a change in seek time with respect to the operating time of the disk drive 20.

  1 and 2, the seek test means 31 periodically performs a seek test. In this embodiment, as shown in FIG. 2, the random seek test is executed at a constant operation time interval (tc). The random seek test is executed, for example, for 3 cycles (when the number of cycles is set to 10 or more, the time required for the test becomes long, and when 1 cycle is used, the variation cannot be absorbed when the operation variation occurs. Cycle execution). The seek pattern may be a repeat seek or an incremental seek. However, all the operation positions of the mechanical parts of the movable part (positioner mechanism including the magnetic head 22 and spindle mechanism including the disk medium 21) that affect the life are affected. Random seek is good for covering.

  The timer counter 32 measures the seek time, and records the measurement result and the operation time at the time of measurement (the elapsed time from the initial operation time of the disk drive 20) in the test result storage unit 33.

  Next, the failure sign determination unit 36, the failure sign seek time (Tsc) 35, the seek time (Tn) at the current measurement with respect to the operation time interval (tc) at which the test was executed, and the seek time at the previous measurement ( Prediction replacement time (tp) is calculated from the difference from Tn−1) and predicted. That is, the predicted replacement time (tp) is calculated from the linear function having a gradient of (Tn−Tn−1) / tc and the failure predictive seek time (Tsc) 35. The failure predictive seek time (Tsc) 35 is set in advance with respect to the initial seek time (To) for each individual disk drive 20 based on field performance data, experimental data, and the like. Alternatively, it is set to a time value (1.1 To) that increases by, for example, 10% with respect to the initial seek time (To).

  The failure sign determination unit 36 displays the information on a display unit (not shown) or predicts to the operator when the failure sign seek time (Tsc) 35 is reached if the gradient is gentle and the gradient difference from the previous measurement is small. It is notified that it is time to replace (tp). If the gradient is large and the gradient difference (change in gradient) from the previous measurement increases rapidly, there is a possibility that the prediction replacement time (tp) may be reached suddenly. Notification is made before the seek time (Tsc) 35 is reached. Note that a determination value for determining whether or not to perform notification depending on the magnitude of the gradient and the difference in the measured value of the gradient between the current time and the previous time is set in advance using a field performance value, an experimental value, an experience value, or the like.

  As described above, the magnetic disk device 10 determines the deterioration status of the disk drive 20 from the difference between the current seek time and the past seek time with respect to the test execution time interval (tc), and performs a predictive diagnosis. There is an effect that the predicted exchange time (tp) corresponding to the situation can be obtained.

  FIG. 3 is a schematic configuration diagram showing a second embodiment of the magnetic disk device 50 of the present invention.

  Referring to FIG. 3, the magnetic disk device 50 includes a disk drive 20 and a failure prediction device 60. In addition, the same component as 1st Example is displayed with the same code | symbol. The second embodiment is different from the first embodiment only in that the failure prediction device 60 calculates a predicted failure time (tr) instead of the predicted replacement time (tp), and only the different points are described mainly. To do.

  The failure prediction device 60 includes seek test means 31 that performs a seek for failure sign diagnosis of the disk drive 20, a test result storage unit 33, a failure occurrence reference time storage unit 67, and a failure occurrence determination unit 69.

  The failure occurrence reference time storage unit 67 stores a failure occurrence seek time (Tsr) 68 set in advance for determining occurrence of a failure. A failure occurrence seek time (Tsr) 68 sets a seek timeout time value of the disk drive 20.

  The failure occurrence determination unit 69 performs prediction by calculating the predicted failure time (tr) of the disk drive 20 from the test result, the operation time, and the failure occurrence seek time (Tsr) 68.

  Next, the operation of the magnetic disk device 50 configured as described above will be described with reference to FIG. 3 and FIG.

  FIG. 4 is a diagram for explaining the calculation of the predicted failure time (tr) of the disk drive 20 and the prediction of the predicted replacement time (tp), and shows the change in seek time with respect to the operating time of the disk drive 20. is there.

  Hereinafter, with reference to FIGS. 3 and 4, only the operation different from the first embodiment will be described.

  The failure occurrence determination unit 69 includes a failure occurrence seek time (Tsr) 68, a seek time (Tn) at the current measurement with respect to an operation time interval (tc) at which the test is executed, and a seek time (Tn-1) at the previous measurement. The predicted failure time (tr) is calculated based on the difference from () and predicted. That is, the predicted failure time (tr) is calculated from the linear function having a gradient of (Tn−Tn−1) / tc and the failure occurrence seek time (Tsr) 68.

  The failure occurrence determination unit 69 determines n hours before the predicted failure time (tr) as the predicted replacement time (tp) and notifies the operator. n is set to a time with a sufficient margin to prevent the occurrence of a failure. For example, n is determined with reference to field performance data, experimental data, etc. of the model-specific disk drive 20.

  In the first and second embodiments, the failure prediction devices 30 and 60 are provided separately from the magnetic disk devices 10 and 50. However, the failure prediction devices 30 and 60 may be included in the magnetic disk devices 10 and 50. Not too long.

  As described above, the magnetic disk device 50 determines the predicted failure time (tr) of the disk drive 20 from the difference between the current seek time (Tn) and the previous seek time (Tn−1) with respect to the test execution time interval (tc). ) And the predicted replacement time (tp) is determined based on the predicted failure time (tr). Therefore, it is possible to predict the predicted failure time (tr) and the predicted replacement time (tp) corresponding to the deterioration state of the disk drive 20.

1 is a schematic configuration diagram showing a first embodiment of a magnetic disk device according to the present invention; FIG. 10 is a diagram for explaining prediction by calculating a predicted replacement time of a disk drive. It is a schematic block diagram which shows 2nd Example of the magnetic disc apparatus of this invention. FIG. 10 is a diagram for explaining a prediction failure time by calculating a predicted failure time of a disk drive.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Magnetic disk apparatus 20 Disk drive 21 Disk medium 22 Magnetic head 30 Failure prediction apparatus 31 Seek test means 32 Timer counter 33 Test result storage part 34 Failure prediction reference time storage part 35 Failure prediction seek time (Tsc)
36 Failure predictor determination unit 50 Magnetic disk device 60 Failure prediction device 67 Failure occurrence reference time storage unit 68 Failure occurrence seek time (Tsr)
69 Failure occurrence determination unit (tc) Time interval (tp) Predicted exchange time (To) Initial seek time (tr) Predictive failure time (Tn) Seek time (Tn-1) Seek time

Claims (22)

A magnetic disk device comprising: a disk drive comprising a disk medium and a magnetic head for writing / reading information to / from the disk medium; and a failure prediction device for predicting failure of the disk drive, wherein the failure prediction device comprises: A seek test means for performing a seek for failure sign diagnosis, a test result storage unit for storing a test result by the seek test means and an operation time at the time of test execution, and a trouble sign seek time for determining a trouble sign A magnetic disk apparatus comprising: a failure sign reference time storage unit for setting and storing; and a failure sign determination unit for predicting a replacement time from the test result, the operation time, and the failure sign seek time. A magnetic disk device comprising: a disk drive comprising a disk medium and a magnetic head for writing / reading information to / from the disk medium; and a failure prediction device for predicting failure of the disk drive, wherein the failure prediction device comprises: , Seek test means for performing a seek for failure occurrence diagnosis, a test result storage unit for storing a test result by the seek test means and an operation time at the time of test execution, and a failure occurrence seek time for determining the occurrence of the failure A magnetic disk apparatus comprising: a failure occurrence reference time storage unit for setting and storing a failure occurrence determination unit that predicts a failure time from the test result, the operation time, and the failure occurrence seek time. 3. The magnetic disk apparatus according to claim 1, wherein the seek test means has a timer counter for measuring a seek time. 4. The magnetic disk apparatus according to claim 1, wherein the seek test means includes means for periodically executing a test and means for repeating a specific seek pattern for a plurality of cycles. 5. The magnetic disk device according to claim 4, wherein the seek pattern has a random seek pattern. 6. The magnetic disk device according to claim 1, wherein the failure predictive seek time is set in advance in correspondence with an initial seek time during operation. The said failure sign judgment part has a means to estimate replacement | exchange time from the difference of the seek time with respect to the operation time interval which performed the test, The any one of Claim 1, 3, 4, 5, 6 characterized by the above-mentioned. Magnetic disk unit. 6. The magnetic disk apparatus according to claim 2, wherein a seek timeout time value is set in advance as the failure occurrence seek time. 9. The failure occurrence determination unit according to claim 2, further comprising means for predicting a failure time from a difference in seek time with respect to an operation time interval in which a test is executed. Magnetic disk unit. A failure prediction apparatus for predicting a failure of a disk drive comprising a disk medium and a magnetic head for writing / reading information to / from the disk medium, a seek test means for performing a seek for failure sign diagnosis, and the seek A test result storage unit for storing a test result by the test means and an operation time at the time of test execution, a failure predictive reference time storage unit for setting and storing a failure predictive seek time for determining a failure predictor, and the test result A failure prediction apparatus, comprising: a failure sign determination unit that predicts a replacement time from the operation time and the failure sign seek time. A failure prediction apparatus for predicting a failure of a disk drive comprising a disk medium and a magnetic head for writing / reading information to / from the disk medium, a seek test means for performing a seek for failure occurrence diagnosis, and the seek A test result storage unit for storing a test result by a test means and an operation time at the time of test execution, a failure occurrence reference time storage unit for setting and storing a failure occurrence seek time for determining occurrence of a failure, and the test result A failure prediction apparatus comprising: a failure occurrence determination unit that predicts a failure time from the operation time and the failure occurrence seek time. The failure prediction apparatus according to claim 10 or 11, wherein the seek test means includes a timer counter that measures a seek time. 13. The failure prediction apparatus according to claim 10, wherein the seek test means includes means for periodically executing a test and means for repeating a specific seek pattern for a plurality of cycles. The failure prediction apparatus according to claim 13, wherein the seek pattern has a random seek pattern. The failure prediction apparatus according to any one of claims 10, 12, 13, and 14, wherein the failure predictive seek time is set in advance corresponding to an initial seek time during operation. The said failure sign judgment part has a means to estimate replacement | exchange time from the difference of the seek time with respect to the operation time interval which performed the test, Any one of Claim 10, 12, 13, 14, 15 characterized by the above-mentioned. Failure prediction device. 15. The failure prediction apparatus according to claim 11, wherein a seek timeout time value is set in advance as the failure occurrence seek time. 18. The failure occurrence determination unit includes means for predicting a failure time from a difference in seek time with respect to an operation time interval at which a test is executed, according to any one of claims 11, 12, 13, 14, and 17. The failure prediction apparatus described. A failure prediction method for a magnetic disk device, comprising: a disk drive comprising a disk medium and a magnetic head for writing / reading information to / from the disk medium; and a failure prediction device for predicting a failure of the disk drive, The failure prediction apparatus performs a seek for failure sign diagnosis by a seek test means, a step of storing a test result by the seek test means and an operation time at the time of test execution in a test result storage unit, and a failure sign A step of setting a failure predictive seek time for determination and storing the failure predictive seek time in a failure predictive reference time storage unit; A failure prediction method for a magnetic disk device, comprising: A failure prediction method for a magnetic disk device, comprising: a disk drive comprising a disk medium and a magnetic head for writing / reading information to / from the disk medium; and a failure prediction device for predicting a failure of the disk drive, The failure prediction device performs a seek for failure occurrence diagnosis by a seek test unit, a step of storing a test result by the seek test unit and an operation time at the time of test execution in a test result storage unit, and a failure occurrence A step of setting a failure occurrence seek time for determination and storing the failure occurrence reference time in a failure occurrence reference time storage unit; and a step of predicting a failure time by the failure occurrence determination unit from the test result, the operation time, and the failure occurrence seek time A failure prediction method for a magnetic disk device, comprising: 20. The failure prediction method for a magnetic disk device according to claim 19, wherein the failure sign determination unit includes a step of predicting a replacement time from a difference in seek time with respect to an operation time interval at which a test is executed. 21. The failure prediction method for a magnetic disk device according to claim 20, wherein the failure occurrence determination unit includes means for predicting a failure time from a difference in seek time with respect to an operation time interval at which a test is executed.

Images