JP3845239B2 - Disk array device and failure recovery method in disk array device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a disk array device, and more particularly to a failure recovery method in a disk array device.
[0002]
[Prior art]
The disk array device has a scale of several tens of TB (terabytes) due to the mounting of a large-capacity hard disk drive (HDD) that is a physical disk. Emphasizing transferability from conventional devices, logical disks having the same capacity as conventional ones are provided. This is realized by a technique that logically divides one array rank composed of a plurality of physical disks into a plurality.
[0003]
[Problems to be solved by the invention]
In such a disk array device, when a failure of a hard disk drive that is a physical disk occurs, all the logical disks related to the failed hard disk drive in the array rank are in a failure state, and for a long time until the completion of recovery, There is a problem that the performance is lowered or the reliability is lowered with respect to the access from the upper level.
[0004]
In addition, recent disk array devices are equipped with large-capacity HDDs, and it takes a long time to complete recovery after a failure occurs. Since the capacity of HDDs will continue to increase in the future, the time until recovery will continue to increase. Time tends to be longer.
[0005]
FIG. 7 is a block diagram of a conventional disk array device. The host 1 and the disk array device 902 are connected by an upper I / F 3. A READ / WRITE access command from the host 1 is received and recognized by the host I / F control unit 201. The logical disk access control unit 202 recognizes the logical disk, the address, and the transfer length of the READ / WRTIE access target according to an instruction from the higher-level I / F control unit 201 and instructs the array control unit 203.
[0006]
The array control unit 203 controls the physical disk 205 that stores the logical disk to be read / written according to an instruction (access type, address, transfer length) from the logical disk access control unit 202. Therefore, the HDD control unit 204 To give instructions.
[0007]
The HDD control unit 204 controls the physical disk 205 according to an instruction from the array control unit 203.
[0008]
After preparation for data transfer is completed on the physical disk 205, READ or WRITE data transfer processing is performed. During the data transfer, the HDD control unit 204 monitors the end of the data transfer.
[0009]
After completing the data transfer, the HDD control unit 204 reports the end of the data transfer to the logical disk access control unit 202 via the array control unit 203.
[0010]
After receiving the data transfer end, the logical disk access control unit 202 reports the data transfer end to the upper I / F control unit 201.
[0011]
The host I / F control unit 201 receives the end of data transfer and reports the end of READ / WRITE access to the host 1.
[0012]
With the above operation, a series of I / O operations (READ / WRITE access) are performed.
[0013]
Such a disk array device affects a plurality of logical disks stored in the HDD when the HDD fails.
[0014]
As an example, consider a case where logical disks “A”, “B”, “C”, and “D” are arranged as shown in FIG. In this case, the logical disk “A” is composed of A1 to A3, the logical disk “B” is composed of B1 to B3, the logical disk “C” is composed of C1 to C3, and the logical disk “D” is It consists of D1 to D3. HDD # 1, HDD # 2, HDD # 3, and HDD # 4 are physical disks, HDD # 1 is composed of A1, B2, and C3, HDD # 2 is composed of A2, B3, and D1, and HDD # 3 is A3. , C1, and D2, and HDD # 4 includes B1, C2, and D3. When a failure occurs in HDD # 2, logical disk “A” data A2, logical disk “B” data B3, logical disk “D” data D4 fails, and logical disk “A” “,“ B ”,“ D ”are affected (see FIG. 3).
[0015]
In the failure recovery, an HDD in which no failure has occurred is read and restored, and then written back to a spare HDD prepared in advance or a replaced HDD. However, until all the data on the HDD is written back, even if the data related to some of the logical disks has been recovered, the recovery is not completed. Become.
[0016]
As a result, there is a problem that a plurality of logical disks are affected by performance degradation and reliability degradation for a long time.
[0017]
FIG. 8 is a diagram in which the process from the start of recovery to the completion of recovery is written in the configuration of FIG. 3. Even if the data A2 is recovered, the recovery process is not completed until the data B3 and D1 are recovered. Show.
[0018]
An object of the present invention is to provide a disk array apparatus and method for accessing a specific logical disk from the host before all logical disks have been restored.
[0019]
[Means for Solving the Problems]
The disk array device according to the present invention includes means for detecting that a physical disk has failed and a physical disk assigned to a priority logical disk among logical disks to which the failed physical disk is assigned. A first reading unit that reads all data related to the priority logical disk from a physical disk of which no failure has occurred, and data of the priority logical disk from the data read by the first reading unit First recovery means for recovering all data stored in the physical disk in which the failure occurred, and first storage means for temporarily storing the data recovered by the first recovery means; , the logical disk is not a priority logical disk of the logical disk that the failed physical disk is allocated A second reading unit that reads all data related to a logical disk that is not the priority logical disk from a physical disk that has not failed among the allocated physical disks, and is read by the second reading unit. Second recovery means for recovering all data stored in the failed physical disk from among the data of the logical disk that is not the preferred logical disk from the stored data, and temporarily stored in the first storage means And a second storage means for storing the stored data and the data restored by the second restoration means .
[0020]
The disk array device according to the present invention further comprises second storage means for storing an identifier of a logical disk designated by the host in the above disk array device, and the reading means is stored in the second storage means. The logical disk having the identifier is used as the priority logical disk.
[0021]
Furthermore, in the disk array device according to the present invention, in the above disk array device, when the priority logical disk is accessed from the host, a failure occurs in the physical disks allocated to the priority logical disk. It further comprises access means for accessing a physical disk that has not been accessed and the first storage means.
[0022]
Furthermore, the disk array device according to the present invention is characterized in that, in the above disk array device, the first storage means is a nonvolatile semiconductor memory.
[0023]
A failure recovery method in a disk array device according to the present invention includes a step of detecting that a failure has occurred in a physical disk, and a logical disk assigned to the priority logical disk among the logical disks to which the physical disk in which the failure has occurred is assigned. a first reading step of reading all the data relating the physical disks in said priority logical disk failure of the physical disk is not generated there, the priority logical disk from data read by said first reading step one o'clock and first recovery step to recover all data failure of the data has been stored in the physical disk has occurred, the data that has been restored by the first restoration step in a first storage means a first storing step of storing manner, with the failed physical disk is allocated A second reading step of reading all data related to the logical disk that is not the priority logical disk from the physical disks that have not failed among the physical disks that are assigned to the logical disks that are not the priority logical disks And a second recovery step for recovering all data stored in the physical disk in which a failure has occurred among the data of the logical disk that is not the priority logical disk from the data read in the second read step. And a second storage step for storing in the second storage means the data temporarily stored in the first storage means and the data recovered by the second recovery step. To do.
[0024]
Furthermore, error recovery method in a disk array device according to the invention, in the error recovery method in the disk array device, a second storage step of storing the identifier of the logical disk specified by the host, the first read In the step, the logical disk having the identifier stored in the second storage step is set as the priority logical disk.
[0025]
Further, the failure recovery method in the disk array device according to the present invention is the above-described failure recovery method in the disk array device, wherein the physical allocated to the priority logical disk when the priority logical disk is accessed from the host. It further has an access step of accessing a physical disk in which no failure has occurred among the disks and the first storage means.
[0026]
DETAILED DESCRIPTION OF THE INVENTION
The present invention provides a non-volatile semiconductor memory in a disk array device that uses a large-capacity physical disk and has a plurality of logical disks in one array rank. Referring to a preset logical disk value, data lost due to a failure in the logical disk indicated by the logical disk value is preferentially restored to the nonvolatile semiconductor memory.
[0027]
As a result, a specific logical disk can be preferentially restored without waiting for a long time until the recovery of the entire area of the failed physical disk is completed, and the performance degradation and reliability degradation time can be shortened. .
[0028]
FIG. 1 shows a block diagram as an embodiment of the present invention.
[0029]
The disk array device 2 according to the present embodiment is connected to the host 1 via a host I / F 3.
[0030]
The disk array device 2 of this embodiment includes a higher-level I / F control unit 201, a logical disk access control unit 202, an array control unit 203, an HDD control unit 204, a physical disk 205, a priority recovery disk instruction unit 206, and a recovery control unit 207. , An HDD failure detection unit 208, and a priority disk memory unit 209.
[0031]
The host I / F control unit 201 controls transfer with the host 1 which is a host device. The logical disk access control unit 202 controls access to the logical disk according to an instruction from the upper I / F control unit 201. The array control unit 203 instructs the HDD control unit 204 to control the physical disk 205 storing the logical disk to be accessed according to the instructions (access type, address, transfer length) from the logical disk access control unit 202. I do.
[0032]
The HDD control unit 204 controls the physical disk 205 according to an instruction from the array control unit 203.
[0033]
The physical disk 205 stores data in the disk and is controlled by the HDD control unit 204. The priority recovery disk instruction unit 206 stores the value of the priority logical disk at the time of recovery that is transferred in advance from the host 1 via the upper I / F control unit 201.
[0034]
In response to an instruction from the HDD failure detection unit 208, the recovery control unit 207 refers to the value stored in the priority recovery disk instruction unit 206 and controls the array control unit 203 to determine the contents of the failed HDD. Restore.
[0035]
The HDD failure detection unit 208 recognizes that a failure has occurred in the physical disk 205 based on an instruction from the HDD control unit 204.
[0036]
The priority disk memory unit 209 is a non-volatile semiconductor memory, and is a memory used to recover the logical disk data of the value stored in the priority recovery disk instruction unit 206 when the HDD failure is recovered.
[0037]
Next, the operation of this embodiment will be described.
[0038]
First, a case where READ access from the host 1 to the disk array device 2 to the logical disk “A” occurs will be described with reference to FIGS.
[0039]
In this disk array device 2, prior to READ access or WRITE access, the value of the logical disk for priority recovery is transferred from the host 1 in advance and set in the priority recovery disk instruction unit 206.
[0040]
The operation for setting the value of the priority recovery disk will be described as an example where “A” is set in the priority recovery disk instruction unit 206 as the value of the priority recovery disk.
[0041]
First, the host 1 issues a priority recovery disk setting command to the disk array device 2. The transferred command is received by the host I / F control unit 201 and the command is recognized. The host I / F control unit 201 recognizes this as a priority recovery disk setting command and requests the host 1 to transfer a disk value to be set. The priority recovery disk value “A” set from the host 1 is transferred and transferred to the priority recovery disk instruction unit 206 via the upper I / F control unit 201.
[0042]
The priority recovery disk instruction unit 206 receives and stores the priority recovery disk value “A”.
[0043]
With this operation, the priority recovery disk value “A” is set.
[0044]
After setting the priority recovery disk value “A”, the host 1 issues a READ access command to the logical disk “A”.
[0045]
The READ access command is received by the host I / F control unit 201, and the READ command is recognized.
[0046]
After recognizing the READ command, the upper I / F control unit 201 instructs the logical disk access control unit 202 to read the logical disk.
[0047]
The logical disk access control unit 202 recognizes the logical disk “A” to be read-accessed, the address, and the transfer length in accordance with an instruction from the upper I / F control unit 201, and instructs the array control unit 203. The array control unit 203 recognizes the logical disk “A” to be read-accessed, and instructs the HDD control unit 204 to specify the address and transfer length for the physical disks HDD # 1, 2, 3 to be accessed.
[0048]
The HDD control unit 204 reads the physical disks HDD # 1, 2, and 3 to be accessed according to an instruction from the array control unit 203.
[0049]
After the physical disks HDD # 1, 2, and 3 are prepared for data transfer, data transfer processing of data A1, A2, and A3 is performed. The array control unit 203 combines the transferred data A1, A2, and A3, and transfers the data on the logical disk A to the logical disk access control unit 202.
[0050]
The logical disk access control unit 202 transfers the data transferred from the array control unit 203 to the host 1 via the upper I / F control unit 201.
[0051]
When the data transfer from the physical disks HDD # 1, 2, 3 is completed, the HDD control unit 204 reports the end of the data transfer to the logical disk access control unit 202 via the array control unit 203.
[0052]
After receiving the data transfer end, the logical disk access control unit 202 reports the data transfer end to the upper I / F control unit 201.
[0053]
The host I / F control unit 201 receives the end of data transfer and reports the end of READ access to the host 1. Through the above operation, the READ access to the logical disk “A” is performed.
[0054]
On the other hand, a case where a WRITE access of the logical disk “B” occurs from the host 1 to the disk array device 2 will be described.
[0055]
The WRITE access command is received by the host I / F control unit 201, and the WRITE command is recognized. After recognizing the WRITE command, the upper I / F control unit 201 instructs the logical disk access control unit 202 to write the logical disk “B”.
[0056]
The logical disk access control unit 202 recognizes the logical access target logical disk “B”, the address, and the transfer length according to an instruction from the host I / F control unit 201, and gives an instruction to the array control unit 203. The array control unit 203 recognizes the logical disk “B” that is the WRITE access target, and instructs the HDD control unit 204 about the address and transfer length of the physical disk HDD # 4, 1, 2 to be accessed.
[0057]
After the physical disks HDD # 4, 1 and 2 are ready for data transfer, data is received from the host 1 via the higher I / F control unit 201 and logical disk access control unit 202. The array control unit 203 divides the received data and performs WRITE of the data B1, B2, and B3 on the physical disks HDD # 4, 1 and 2.
[0058]
After the WRITE of data is completed, the HDD control unit 204 reports the end of data transfer to the logical disk access control unit 202 via the array control unit 203.
[0059]
After receiving the data transfer end, the logical disk access control unit 202 reports the data transfer end to the upper I / F control unit 201.
[0060]
The host I / F control unit 201 receives the end of the data transfer, and reports to the host 1 the end of the WRITE access of the logical disk “B”.
[0061]
With the above operation, the WRITE access of the logical disk “B” is performed.
[0062]
With reference to FIGS. 1 and 4, the recovery operation when the HDD # 2 fails in the physical disk 205 as shown in FIG. 3 will be described.
[0063]
The HDD control unit 204 recognizes that a failure has occurred in the HDD # 2 in the physical disk 205, and instructs the HDD failure detection unit 208. The HDD failure detection unit 208 detects that a failure has occurred in the physical disk HDD # 2 according to an instruction from the HDD control unit 204. The HDD failure detection unit 208 that has detected the failure of the physical disk HDD # 2 issues a recovery instruction to the recovery control unit 207 that recovers the HDD.
[0064]
The recovery control unit 207 refers to the value set in the priority recovery disk instruction unit 206 (assuming that the logical disk “A” is set) and recovers with priority among the failed logical disks. The logical disk “A” to be processed is recognized. The recovery control unit 207 specifies the logical disk “A” to be recovered with priority to the array control unit 203 and instructs the priority disk memory unit 209 to recover the data A2 lost due to the failure. The array control unit 203 instructs the HDD control unit 204 to access the physical disks HDD # 1, 3 and read the data A1, A3. Based on the read data A 1 and A 3, the array control unit 203 restores the data A 2 lost due to the failure, and writes the data to the nonvolatile semiconductor memory of the priority disk memory unit 209. After restoring the data A2 to the priority disk memory unit 209, the array control unit 203 accesses the logical disk “A” using the priority disk memory unit 209 as shown in FIG. When the host 1 accesses the logical disk “A”, the array control unit 203 passes the HDD control unit 204 through the data A1, A3 and the priority disk memory in the physical disks HDD # 1, 3. It controls to access the data A2 in the unit 209.
[0065]
Through the operation as described above, the logical disk designated by the priority recovery can be recovered preferentially using the non-volatile semiconductor memory of the priority disk memory unit 209, and the data as shown in FIG. There is no need to wait until the recovery of the logical disks “B” and “C” is completed.
[0066]
Subsequently, when a spare HDD is installed in advance, or when a failed HDD is replaced with a normal HDD, a logical disk for which priority recovery is not designated is restored (see FIG. 5).
[0067]
The recovery control unit 207 controls the array control unit 203 to recover the logical disks “B” and “D” other than the value set in the priority recovery disk instruction unit 206 (set the logical disk “A”).
[0068]
The array control unit 203 accesses the HDD control unit 204 to the HDDs # 1, 3, and 4, which are HDDs other than the failed HDD # 2, and reads the data B1, B2, and D2, D3. Instruct. Based on the read data B1, B2, and D2, D3, the array control unit 203 restores the data B3, D1 lost due to the failure, and writes the data to the spare HDD or the replaced HDD. The array control unit 203 accesses the logical disks “B” and “D” using the HDD # 2 as usual after restoring the data B3 and D1 to the spare HDD or the replaced HDD. As described above, the recovery of the logical disks “B” and “D” is performed.
[0069]
Further, the data A2 stored in the priority disk memory unit 209 performs WRITE to the HDD # 2 according to an instruction from the operator when the access to the logical disk “A” is lightly loaded, and thereafter, as usual, the HDD Access using # 2 becomes possible.
[0070]
【The invention's effect】
The first effect is that in a disk array device, when a failure occurs in a physical disk constituting an array rank, a logical disk having a high priority can be preferentially restored.
[0071]
As a result, it is possible to minimize the time until recovery of a logical disk having a high priority, and to shorten the time for performance degradation and reliability degradation.
[0072]
In particular, when the performance degradation is not allowed or important data exists, a great effect is exhibited.
[0073]
The reason is that data related to a logical disk having a high priority among the data of a plurality of logical disks in the failed disk is preferentially restored to the priority disk memory unit which is a nonvolatile semiconductor memory. The other logical disk data is restored as usual.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of a disk array device according to an embodiment of the present invention and a host connected thereto.
FIG. 2 is a diagram illustrating a configuration example of an array rank.
3 is a diagram showing a state in which a failure has occurred in HDD # 2 in the configuration of the array rank of FIG. 2;
4 is a diagram illustrating a state in which a failure has occurred in HDD # 2 and recovery data A2 has been written in a nonvolatile semiconductor memory in the array rank configuration of FIG. 2 according to an embodiment of the present invention.
FIG. 5 is a diagram for explaining the recovery operation of logical disks other than the priority logical disk according to the embodiment of the present invention;
FIG. 6 is a timing chart showing a state from the start of recovery according to the present invention until recovery of a logical disk with a high priority is completed, and after that, another logical disk is recovered.
FIG. 7 is a block diagram showing a configuration of a conventional disk array device and a host connected thereto.
FIG. 8 is a timing chart showing a state from the start of recovery to the completion of recovery according to a conventional example.
[Explanation of symbols]
1 Host 2 Disk array device 3 Host I / F
201 Host I / F Control Unit 202 Logical Disk Access Control Unit 203 Array Control Unit 204 HDD Control Unit 205 Physical Disk 206 Priority Recovery Disk Instruction Unit 207 Recovery Control Unit 209 HDD Failure Detection Unit 209 Priority Disk Memory Unit

Claims (6)

Means for detecting that a physical disk has failed;
Priority information storage means for holding an identifier of a priority logical disk that is preferentially restored in the event of a physical disk failure;
Reading means for reading out data related to the logical disk stored in the physical disk in which no failure has occurred with respect to the logical disk in which data is allocated to the physical disk in which the failure has occurred;
Recovery means for recovering data of the logical disk stored in the failed physical disk from the data read by the reading means;
First and second storage means for storing the data restored by the restoration means,
The recovery means preferentially recovers the data of the priority logical disk corresponding to the identifier held in the priority information storage means at the time of data recovery, stores the data in the first storage means, and then stores the other logical disk The data is sequentially recovered and stored in the second storage means,
When the data recovery of the priority logical disk is completed in the first storage means, access to the priority logical disk from the host device without waiting for the data recovery completion of other logical disks in the second storage means A disk array device characterized by being made possible . 2. The disk array device according to claim 1 , wherein a physical disk in which no failure has occurred among physical disks allocated to the priority logical disk when the priority logical disk is accessed from a host; A disk array device further comprising access means for accessing the first storage means. 3. The disk array device according to claim 1, wherein the first storage means is a non-volatile semiconductor memory. Detecting that a physical disk has failed;
Read all the data related to the priority logical disk from the physical disks that have not failed among the physical disks that are assigned to the priority logical disk of the logical disks to which the failed physical disk is assigned A first reading step;
A first recovery step of recovering all data stored in the physical disk in which a failure has occurred among the data of the priority logical disk from the data read in the first reading step;
A first storage step of temporarily storing the data recovered by the first recovery step in a first storage means;
From the physical disk assigned to the logical disk that is not the preferred logical disk among the logical disks to which the failed physical disk is assigned, to the logical disk that is not the preferred logical disk. A second reading step for reading all such data;
A second recovery step of recovering all data stored in the physical disk in which a failure has occurred among the data of the logical disk that is not the priority logical disk from the data read in the second reading step;
A second storage step of storing the data recovered in the second recovery step in a second storage means;
A third storage step of storing data temporarily stored in the first storage means in the second storage means;
A failure recovery method for a disk array device, comprising: 5. The failure recovery method for a disk array device according to claim 4 , further comprising a fourth storage step for storing an identifier of a logical disk designated by a host, wherein the first read step stores in the fourth storage step. A failure recovery method in a disk array device, wherein a logical disk having a specified identifier is used as the priority logical disk. 6. The failure recovery method for a disk array device according to claim 4 , wherein when a priority is given to the priority logical disk from a host, a failure occurs in a physical disk assigned to the priority logical disk. A failure recovery method in a disk array device, further comprising an access step of accessing a physical disk that has not been accessed and the first storage means.

Images