JP2004227449A - Diagnostic device for trouble in disk array device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To grasp error conditions for every unit disk constituting an RAID and across a plurality of disks, and to prevent data from being lost by preliminarily detecting the danger of data loss due to an error which is likely to be generated in a plurality of disks. <P>SOLUTION: This disk array device diagnostic device is provided with: first diagnostic processing for analyzing the operation history information of every unit disk constituting a disk array device, and for executing the diagnostic processing of every unit disk; second diagnosis processing for specifying a unit disk group belonging to a data group from configuration information showing the configuration of the disk array device and analyzing the conditions of any failure in the unit disk group belonging to the data group from the result of the first diagnostic processing; and a display device for displaying the result of the first diagnostic processing or the result of the second diagnostic processing. Furthermore, this diagnostic device is provided with third diagnostic processing for preliminarily storing the combination conditions of the disks under which an error is likely to be generated in a condition file, and for collating them with the results of the first diagnostic processing on the basis of the combination conditions from the condition file, and displaying the diagnostic result on the display device when any error due to the combination conditions is determined as the result of the third diagnostic processing. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a diagnostic device for a failure in a disk array device (RAID: Redundant Array of Inexpensive Disks), and in particular, manages a disk failure status of a disk array device to detect a sign of a disk failure in advance, and to detect data failure in advance. The present invention relates to a process for diagnosing a failure of a disk array device that prevents loss of the disk array in advance.
[0002]
[Prior art]
The disk array device is composed of a plurality of single hard disks, and data groups are written and read over a plurality of hard disks. Since writing and reading of data are performed by adding redundant data for lost data recovery, even if an error occurs in one single disk, data is read from another disk belonging to the same data group using the redundant data for recovery. Data can be read by generating and restoring. Even when an error has occurred in two disks, if no two disk errors have occurred within the minimum address unit having the redundant data for restoration, data is similarly generated from another disk. It can recover and read data, and has high reliability.
[0003]
On the other hand, even if there is an error in a single disk, the disk array device uses the redundant data for restoration and operates without any apparent failure status. Can not be confirmed from. In order to grasp the internal error occurrence situation, it becomes necessary for the maintenance staff of the disk array device to analyze the log information of the disk array device.
Even if multiple disks fail due to initial failure of the disk and deterioration of parts due to long-term use, errors may potentially be progressing and there is a risk that data may be lost in the future. is there.
[0004]
As general preventive maintenance for a disk array device, maintenance personnel refer to the log information of the disk array device when performing regular maintenance, and determine whether there is a disk that needs preventive replacement based on the disk error status. That is the fact. However, the failure of the disk may be progressing before the next periodic maintenance is performed, and the maintenance staff cannot know such a situation. In the worst case, an error may occur in two disks to which the same data group belongs. In such a case, it is impossible to recover the error by using the redundant data for repair, so the periodic maintenance is performed. Depending on the timing, there is a risk of data loss.
[0005]
As a technique relating to preventive maintenance of a disk device error, for example, there is a technique disclosed in Japanese Patent Application Laid-Open No. H11-353819 (Patent Document 1). According to this technology, the number of times of occurrence of a disk error is stored in a counter, and when the number of times of occurrence of the error exceeds a threshold value, an alarm is issued to detect a disk failure in advance and prevent data loss.
[0006]
[Patent Document 1]
JP-A-11-353819 (pages 2-4, FIG. 2)
[0007]
[Problems to be solved by the invention]
However, the technique described in Japanese Patent Application Laid-Open No. H11-353819 is merely a management based on a disk error threshold, and there is no suggestion on how to apply the technique to, for example, RAID. Even if it is applied to RAID, it is only a hint that the failure of a single disk is detected in advance, and there is no explanation about how to grasp the error state of all the disks constituting the RAID in a cross-sectional manner and how to diagnose it. Not mentioned.
[0008]
SUMMARY OF THE INVENTION An object of the present invention is to provide a disk array device that can prevent data loss by grasping the error situation in the disk array device and detecting in advance the danger of data loss due to the occurrence of an error in a plurality of disks. To provide a diagnostic device.
It is another object of the present invention to automatically perform error diagnosis of a single disk constituting a disk array device, error diagnosis of a plurality of disk groups, and a combination thereof.
[0009]
[Means for Solving the Problems]
The present invention is directed to a system including a disk array device that writes and reads data in a data group by adding redundant data for recovering lost data to a single disk group including a plurality of single disks. A storage device for collecting and storing the operation history information of each single disk and the configuration information indicating the configuration of the disk array device, and analyzing the operation history information of each single disk read from this storage device and analyzing each single disk A first diagnostic processing unit for performing a diagnostic process, a single disk group belonging to a data group configuring the disk array device is identified from the configuration information read from the storage device, and a single disk group is identified from a diagnosis result of the first diagnostic processing unit. A second diagnostic processing unit that analyzes and processes a failure situation in the disk group; and a first diagnostic processing unit. It constitutes a diagnostic apparatus and a display means for displaying information related to the diagnostic result of the information or the second diagnosis processing means associated with the cross-sectional results.
Here, the operation history information includes, for example, information regarding the error content, error address, and occurrence time of the single disk, and the configuration information includes, for example, information configuring a single disk group corresponding to the data group. Information is periodically collected from the disk array device and stored in the storage device. The first diagnostic processing means periodically calculates the number of error occurrences, the error increase rate, and the error address distribution of each disk from the operation history information. The second diagnostic processing means may calculate a failure state in the data group having the same repair redundant data when an error is detected in a plurality of disks according to the processing result of the first diagnostic processing means. preferable.
Also, a third diagnostic processing unit having a combination condition file in which a combination condition in which an error of the disk is likely to occur is stored in advance, and a processing result of the first diagnostic processing unit is collated based on the combination condition file. Have. When an error due to the combination condition is determined as a result of the diagnosis by the third diagnosis processing means, information relating to the diagnosis result is visually displayed on the display means to notify a failure.
Further, when the second diagnostic processing unit detects a failure in a plurality of disks belonging to the same data group based on the failure analysis result, the second diagnosis processing unit includes a batch processing unit for performing diagnosis of the plurality of disks by batch processing.
The present invention also considers the diagnostic processing in the first, second, and third diagnostic processing means as first to third diagnostic processing modes, respectively, and refers to a diagnostic method appropriately including these diagnostic processing modes. be able to.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a diagram showing a configuration of a disk array device (RAID) and its diagnostic device. The RAID 10 is connected to the diagnostic device 20 via a communication line 30. The communication line 30 is, for example, an interface conforming to the LAN standard. The illustrated example shows an example in which one RAID 10 is connected for the sake of explanation, but it may be considered that a plurality of RAIDs are actually connected to the communication line 30. Therefore, the diagnostic device 20 diagnoses a plurality of RAIDs.
[0011]
The RAID 10 includes a plurality of single disk groups 12 and a controller 11. The plurality of single disk groups 12 is a group of m single disks in the horizontal direction and a plurality of disk groups (n × m single disk groups in total) having n rows in the vertical direction. The data, including the redundant data for restoration, is recorded in groups of m individual disks in the group unit in the horizontal direction.
[0012]
The controller 11 has a memory 13 and a processing unit (not shown), controls writing and reading of data to and from the single disk group 12, accesses the single disk group 12, and collects status information. The status information includes operation history information 13 and RAUD configuration information 14 for each disk, and the collected information is temporarily stored in the memory 13.
The operation history information 13 includes information on an error type, an error occurrence address, and an occurrence time for each error as error history information at the time of accessing each disk. The error type indicates an error type such as a hardware error, a disk medium error, and an interface error. The error address indicates an address on the disk for specifying the location where the error has occurred. The occurrence time indicates the time when the error occurred. Each time an error occurs in each disk, operation history information including such information is stored in the memory 13 as log information.
Since the operation history information 13 is collected from each single disk of each RAID, if there are m × n single disks, the operation history information 13 is also stored in the memory 13 corresponding to each single disk. Become.
The configuration information 14 is information relating to the RAID device configuration, and includes the RAID level (specifying the format of data storage), the total number of individual disks constituting each RAID, the size of the data group (the storage capacity of the group), that is, the head of the data group And information indicating the address of the single disk and its width (the number of disks in the horizontal direction), and the storage capacity of the data group (vertical depth: G-byte capacity, etc.). The RAID level defines the format of group data storage.
These pieces of information 14 and 15 are periodically transmitted as RAID state information 31, for example, every 24 hours from each RAID 10 to the diagnostic apparatus 20 via the communication line 30.
[0013]
Next, the diagnostic device 20 will be described.
The diagnostic device 20 diagnoses an error situation based on the operation history information 14 and the configuration information 15 transmitted from each RAID 10 and performs a process for diagnosing and recovering an error part in the target RAID according to the diagnosis result. It has a function to emit.
The diagnostic device 20 is, for example, a personal computer or a computer. Although not shown, a processing device for executing a program on an internal bus to process diagnostic data, a memory for storing the program and data, and a state of the RAID indicate a diagnostic result. Display device (reference numeral 400 in FIG. 2), an input device for maintenance personnel to input instructions for diagnosis and various information, and a hard disk (storage device) for storing status information from RAID 10 and information on diagnosis results. It comprises.
[0014]
The program is executed by the processing device to perform the diagnostic processing. The diagnosis process includes an error diagnosis process 21 for each single disk, an error diagnosis process 22 for a plurality of disks, and a combination condition check process 23. As a result of the error diagnosis process 21 for each single disk, the error diagnosis process 22 for a plurality of disks, and the combination condition check process 23, the diagnosis device 20 diagnoses a disk with a precursor of a failure related to the single disk group 12 of the target RAID 10, A batch processing command 32 for performing recovery is issued. The command 32 is transmitted to the target RAID 10 through the line 30.
[0015]
As described above, the operation history information 14 and the configuration information 15 periodically received via the communication line 30 are stored in a memory (not shown) corresponding to each single disk of each RAID. Then, the processing device reads the operation history information 14 stored in the memory, and calculates the number of error occurrences per unit time (for example, 24 hours), the error increase rate, and the error address distribution for each disk. The calculated result is displayed on the display device.
[0016]
Further, the processing device of the diagnostic device 20 determines the number of errors occurring for the same data group having the same repair redundant data in RAID from the number of errors, the rate of error increase, and the address distribution of errors calculated for each single disk. Calculate the error increase rate and error address distribution. Based on the calculation result, a diagnosis is performed in relation to a single disk group that has a risk of data loss due to a failure in the same data group. The status of the diagnosis result is displayed on the display device. This allows the maintenance staff to know the diagnosis status and to know the danger of data loss for each RAID data group.
[0017]
When detecting a failure in a plurality of single disks, the diagnostic device 20 has a function of creating a batch process for diagnosing the disk in the diagnostic device, and issues a command 32 for that to the RAID 10 that is the target.
[0018]
FIG. 2 is a flowchart showing an outline of the error diagnosis processing 21 for each single disk.
The RAID status information (the operation history information 14 and the configuration information 15) 31 automatically and periodically transmitted from the RAID 10 via the communication line 30 includes information on a hard disk device (not shown) in the diagnostic device 20. It is stored in the save area 200 (step 101). Since the state information 31 is periodically transmitted from the RAID every 24 hours and is stored and accumulated in the area 200, the state information for several times is subjected to generation management. The operation history information of each disk is extracted from the state information, and further information relating to the disk error is extracted (step 102). Since the extracted history information 14 includes information on the error type, the error occurrence address, and the occurrence time, it is possible to know which error occurred at which address on which disk.
[0019]
From the operation history information 14 read from the information save area, the processing device calculates the number of times of error occurrence, the rate of increase of errors, and the address distribution of errors in each disk for each unit time (for example, 24 hours). Then, the calculation result is stored in the hard disk device as a result file 300. The result file 300 stores analysis results of past error information analyzed so far. The information of the previous analysis result is read from the result file 300 (step 103), and is compared with the result of the error analysis information calculated this time (step 104). Not only a sudden error occurrence state but also a long-term error state is checked, and the result is recorded in the result file 300 and displayed on the display device 400 together. Based on the result file up to the last time and the current disk error information, it is diagnosed whether there is a disk that has a sign of failure (step 105). In the diagnostic processing, weighting is performed based on the hardware specifications of the disk according to the severity of each error content, and an upper limit value of an error increase rate that may cause a failure in the future is individually set according to the error type. It is checked whether or not each disk has exceeded the above-mentioned upper limit of the error increase rate (step 106). If none exceeds the upper limit, the diagnostic processing is terminated (step 108).
[0020]
On the other hand, if there is a disk exceeding the upper limit, a report of the disk exceeding the upper limit and a warning display are displayed on the display device 400, so that the maintenance staff can detect and notify a failure part in advance. Yes (step 107).
[0021]
FIG. 3 is a flowchart showing an outline of the error diagnosis processing 22 for a plurality of disks.
This processing is executed when an error has occurred in a plurality of disks in the error diagnosis processing 21 for each single disk (step 110). This can be seen from the result of the error diagnosis processing 21 for each single disk shown in FIG. 2, and when an error over a plurality of disks is recognized, the diagnosis processing 22 is started.
The device configuration information 15 is extracted from the RAID status information stored in the information save area 200 (step 111). Since the configuration information has information for specifying a single disk group that belongs to the same data group logically having the same repair redundant data, the data group assigned to each disk that has already failed can be sorted out. .
[0022]
In addition, since the error address information of each disk indicated by the operation history information 14 for each disk is the same as the address indicating the physical position in the disk, when considering a logical file image, where the error address information is based on the error address and the configuration information, It is calculated whether an error has occurred in the address of the data group allocated to the file (step 112). That is, the location of the failed disk is physically specified by the “cylinder address”, “track address”, and “sector address”, which are included in the operation history information and stored in the information save area 200. On the other hand, the configuration information includes the size of the data group (storage capacity of the group), the number of used disks, and the RAID level (specifying the format of data storage). Since it is possible to determine which range of the single disk group each data group uses based on the configuration information, it is possible to know which data group includes the physical location where the failure has occurred on the single disk be able to.
Then, a notification of an error occurrence status on a plurality of disks is made (step 113). In this case, as the error information, the position of the error disk, the error occurrence status, and the error address information of the logical file calculated in step 112 are reported.
[0023]
Next, a batch file creation process is performed (step 114). This process is performed when the risk of data loss is notified (step 113). Is created, and the created batch process is stored as a batch process file 500 in the hard disk device.
The batch processing is created by using a command I / F conforming to the I / F standard of the communication line 30 connected to the RAID 10, and automatically creates a read search command sequence for all media surfaces for the disk.
[0024]
By executing the batch process 32 for the RAID 10, the RAID performs a read search on all the media surfaces of the disk, and an error portion detected by the read search is assigned to a reassign block function (a single disk generally has). The error address portion is not used, and the error address portion for a plurality of disks is recovered by a function of pasting data to a replacement sector), thereby eliminating the possibility of data loss and maintaining data integrity.
[0025]
The start timing of the batch processing can be arbitrarily set by the administrator of the diagnostic device 20. The diagnostic device 20 determines the setting status (step 115), and if the setting of the automatic start has been input, immediately after the notification, Batch processing is automatically started (step 116). On the other hand, if the setting of manual activation has been input, the batch processing at the time of notification is not activated, and the administrator of the diagnostic apparatus manually activates the batch processing (step 118). The diagnostic apparatus 20 confirms the content of the execution result of the batch processing by reading the state information 31 from the RAID 10 (steps 117 and 119).
[0026]
FIG. 4 is a flowchart showing an outline of the combination condition check processing 23. The combination condition checking process 23 is a diagnosis process for diagnosing a disk error based on a combination of disk error conditions and detecting the error in advance. The combination condition refers to a combination of an error content and an error occurrence time, a combination of an error content and an address, a combination of a specific condition and an error occurrence timing, and the like. It has been empirically known that an error peculiar to the hard disk device constituting the RAID occurs due to the combination of these. Such a combination condition is stored in the hard disk device of the diagnostic device 20 as a combination condition file 600 in advance.
[0027]
When the combination condition check process is started, combination conditions such as an error timing with a specific condition and an error timing with a plurality of conditions are read from the combination condition file 600 (step 121). Then, a comparison process is performed between the diagnosis process 21 (FIG. 2) for the single disk stored in the result file 300 and the combination condition read from the combination file 600 (step 122). As a result of the comparison processing, when it is determined that there is an error in the single disk based on the combination condition (step 123), the diagnostic device 20 sends a preliminary notice that an error occurs in the single disk based on the combination condition. Perform (124). The notice of advance is displayed on the display device 400 to notify maintenance personnel. On the other hand, as a result of the comparison process, if there is no error in the single disk based on the combination condition, the diagnosis check process based on the combination condition ends.
The combination condition check processing 23 can be executed at any time as long as the conditions are registered in the combination file 600, but it is preferable that this processing is also performed periodically. Therefore, it is preferable to perform the above process periodically when the error diagnosis process 21 for a single disk is performed or when the multiple disk error diagnosis process 22 is performed.
As described above, by performing the diagnosis based on the combination condition of the error status regarding the hard disks configuring the RAID, it is possible to grasp an error state that cannot be obtained from the management of the diagnosis for the single disk, for example, using only the error threshold. This makes it possible to detect in advance the danger of data loss due to the occurrence of a RAID error composed of a plurality of single disks.
[0028]
As described above, according to the above-described embodiment, the operation history information of a single disk collected at regular intervals is stored in the storage device of the diagnostic apparatus several times sequentially, and the history information and the RAID configuration are stored. By diagnosing according to the information, it is possible to detect in advance not only the failure of each single disk but also a long-term variation of the failure from the failure increase rate in a situation where the failure of the disk gradually progresses over a plurality of disks. .
[0029]
【The invention's effect】
As described above, according to the present invention, it is possible to perform not only the error diagnosis of the single disk constituting the disk array device, but also the error diagnosis over a plurality of disks, and the diagnosis under the condition combining these. . As a result, the error status of the disk can be grasped, the danger of data loss due to errors over a plurality of disks can be detected in advance, and data loss can be prevented.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a schematic configuration of a RAID and a failure diagnosis apparatus according to an embodiment of the present invention.
FIG. 2 is a flowchart showing an outline of error diagnosis processing for each single disk.
FIG. 3 is a flowchart showing an outline of error diagnosis processing for a plurality of disks.
FIG. 4 is a flowchart showing an outline of a combination condition check process.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... RAID 11 ... Single disk group 12 ... Controller 13 ... Memory 14 ... Operation history information for each disk 15 ... RAID configuration information 20 ... Diagnosis device 21 ... Error diagnosis processing for each single disk 22 ... Error diagnosis processing for multiple disks 23 ... Combination condition check processing 30. Communication line 31. RAID status information 32. Batch processing command.

Claims (7)

In a system including a disk array apparatus that writes and reads data of a data group by adding redundant data for recovering lost data to a single disk group composed of a plurality of single disks,
A storage device for collecting and storing operation history information of each single disk of the disk array device and configuration information indicating the configuration of the disk array device;
First diagnostic processing means for analyzing the operation history information of each single disk read from the storage device and performing diagnostic processing for each single disk,
A single disk group belonging to a data group constituting the disk array device is specified from the configuration information read from the storage device, and a failure state in the single disk group is analyzed from a diagnosis result of the first diagnostic processing unit. Second diagnostic processing means for processing
Display means for displaying information related to the diagnosis result by the first diagnosis processing means, or information related to the diagnosis result by the second diagnosis processing means,
A diagnostic device comprising: 2. The operation history information according to claim 1, wherein the operation history information includes information on an error content, an error address, and an occurrence time of a single disk, and the configuration information includes information configuring a single disk group corresponding to a data group. And configuration information is periodically collected from the disk array device and stored in the storage device,
The first diagnostic processing means periodically calculates the number of error occurrences, the error increase rate, and the error address distribution of each disk from the operation history information,
The second diagnostic processing means calculates a failure state in a data group having the same redundant data for repair, when an error is detected in a plurality of disks, according to a processing result of the first diagnostic processing means. A diagnostic device characterized by the above-mentioned. 3. The method according to claim 1, wherein the second diagnostic processing means further performs a diagnosis of the plurality of disks by a batch process when a plurality of disks to which the same data group belongs are found from the analysis result of the failure. A diagnostic apparatus comprising: a batch processing unit. The combination condition file according to any one of claims 1 to 3, further comprising: a combination condition file storing a combination condition in which a disk error is likely to occur; and a combination condition read from the combination condition file. A third diagnostic processing unit for collating the processing result of the one diagnostic processing unit, and as a result of the diagnosis by the third diagnostic processing unit, when an error due to the combination condition is determined, information related to the diagnostic result is provided. A diagnostic device for displaying on the display means. In a disk array device diagnostic processing method for writing and reading data of a data group by adding redundant data for recovering lost data to a single disk group composed of a plurality of single disks,
Periodically collecting operation history information of each single disk of the disk array device and configuration information indicating the configuration of the disk array device;
A first diagnostic processing mode for analyzing the collected operation history information of each single disk and performing diagnostic processing for each single disk,
A second diagnostic processing mode for identifying a single disk group belonging to a data group configuring the disk array device based on the configuration information and analyzing and processing a failure situation in the single disk group from the result of the first diagnostic processing; ,
Visually displaying and notifying information related to the result of the first diagnostic processing or information related to the result of the second diagnostic processing,
A diagnostic method comprising: 6. The diagnostic method according to claim 5, wherein the second diagnostic processing mode is performed when an error is found in a plurality of disks as a result of the diagnostic processing in the first diagnostic processing mode. 7. The third diagnostic processing mode according to claim 5, further comprising a third diagnostic processing mode for performing a process of comparing processing results of the first diagnostic processing mode based on a combination condition in which a disk error is likely to occur. A diagnostic method characterized by displaying information related to a result of diagnosis in a diagnostic processing mode on a display device.

Images