JP2002108721A - Disk array device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a disk array device which can improve the reliability of array constitution information which is held by a controller and respective disk drives and manages an array. SOLUTION: Storage areas for the array constitution information present on a control part and multiple disk drive is 1st storage areas and in addition to them, the control part and disk drives have storage areas for the same array constitution information. The information in the 1st storage areas is compared to estimate correct information and the information in the 2nd storage areas are compared again instead of information judged to be incorrect. When a match is thus obtained, the information judged to be incorrect is rewritten with the information in the 2nd storage areas. Consequently, even if the information in the 1st storage area is destroyed, it can be restored by using the information in the 2nd storage areas.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk array device in which a plurality of disk devices are arrayed to form a predetermined number of logical disks, and more particularly, to improving the reliability of information for managing an array. The present invention relates to a suitable disk array device.

[0002]

2. Description of the Related Art A disk array device in which a plurality of disk devices are arrayed is, for example, a RAID (redunda).
nt arrays of inexpensive
used with a RAID controller as a disk for disks.

In order to manage the array, the RAID controller holds array ID, RAID level, ID of a magnetic disk device belonging to the array, storage capacity of the array, and the like as array configuration information. The array configuration information may include a stripe size, a write cache method, a read cache method, and the like in addition to the above.

If the array configuration information is stored only on the controller, it is possible to manage the disk devices as an array for the time being. However, the array configuration information is usually stored on the controller and each disk device in consideration of the case of replacement of the controller. Thus, when the controller is replaced, the new controller can acquire the array configuration information from the disk device. In addition, even when the array configuration information on any disk device is destroyed, the destroyed array configuration information can be copied and restored from another disk device.

[0005] The controller checks whether the array configuration information on the controller and each disk device are consistent at the time of system startup, in order to guarantee these operations.

In some cases, a plurality of arrayed disk devices may belong to an array under the controller.
In that case, each disk device is identified to which array it belongs by the above array ID.

[0007]

Whether or not the array configuration information on the disk device has been destroyed is determined by CRC (cyc
(like redundancy check). In other words, if CRC is performed by regarding writing and reading to the disk as a transmission system, a change in data during transmission, that is, destruction of the written data can be detected.

However, when it is detected that the array configuration information of a certain disk device has been destroyed, it becomes impossible to know which array the disk device belongs to due to the loss of the array configuration information on the disk device. There may be cases. In this case, it is not always possible to restore the array configuration information such that the disk device whose array configuration information has been destroyed belongs to the original array by the above-described information holding method.

Further, depending on the CRC, the destruction of the array configuration information which is not detected as the destruction of the data may happen by chance. In this case, a check at the time of system startup may result in that the information on the controller does not match the information on the disk even though the array configuration information has not been destroyed. In this case, the cause of inconsistency must be separately investigated and dealt with.

[0010] The present invention has been made in view of the above circumstances, and it is possible to improve the reliability of array configuration information for managing an array, which is held by a controller and each disk device. It is an object to provide a disk array device.

[0011]

In order to solve the above problems, a disk control device according to the present invention has a plurality of disk devices constituting a disk array, and a control unit for controlling the disk array. The plurality of disk devices have first and second areas for storing information for managing the disk array, respectively, and the control unit is configured to store the information for managing the disk array. First and second storage units, and the information of the first area for each of the plurality of disk devices is stored in the second storage unit.
And means for rewriting the information in the first storage unit to information in the second area (claim 1).

The storage area of the array configuration information existing in the control unit and the plurality of disk devices is defined as a first storage area, and the storage area of the same array configuration information is separately stored in the control unit and the plurality of disk devices in the second storage area. As a storage area.

Thus, even if the information in the first storage area is destroyed, it can be restored using the information in the second storage area. Therefore, the reliability of the array configuration information for managing the array can be improved.

Further, the present invention has a plurality of disk devices constituting a disk array, and a control unit for controlling the disk array, wherein the plurality of disk devices store information for managing the disk array. First and second areas for storing information; the control unit includes first and second storage units for storing the information for managing the disk array; and a control unit configured to store the information in the first area. First comparing means for comparing each piece of information and information stored in the first storage unit, means for estimating correct information by a majority decision when the compared results do not match; The comparison is performed using the information of the corresponding second area or the second storage unit instead of the information of the area or the first storage unit that does not have the estimated correct information. When the comparison result of the second comparison means and the comparison result of the second comparison means match, the information of the first area or the first storage unit that does not have the estimated correct information is corresponded. Means for rewriting the information in the second area or the second storage section (claim 2).

Also in this case, the storage area of the array configuration information existing in the control unit and the plurality of disk devices is used as the first storage area. Apart from this, the same array configuration information is stored in the control unit and the plurality of disk devices. An area is provided as a second storage area.

Further, the information in the first storage area is compared to estimate correct information, and the information not determined to be correct is replaced with the information in the second storage area and the comparison is repeated. As a result, if a match is obtained, the information determined to be incorrect is rewritten with the corresponding information in the second storage area. That is, even if the information in the first storage area is destroyed, it can be restored using the information in the second storage area. Therefore, the reliability of the array configuration information for managing the array can be improved.

Further, the control unit includes the first and second control units.
And a control means for rewriting all of the area and the first and second storage sections and controlling a procedure of the rewriting, wherein the control means comprises a first storage section, the first area, The contents are rewritten by a procedure of each of the second storage unit and the second area (claim 3).

This takes into account the case where the contents of the storage area are rewritten when the above-described first storage area and second storage area are provided. If the rewriting procedure is determined in this way, even if the rewriting is not completed due to a power failure during the rewriting, the rewriting is completed after the power is restored without losing the information for managing the array. Or return to the state before rewriting. Therefore, it is possible to improve the reliability of the array configuration information for managing the array in consideration of a situation such as a power failure.

Further, according to the present invention, the first and second areas and the first and second storage sections have first and second storage areas, respectively, and the first and second storage areas are provided.
And a storage area for storing a pointer pointing to any of the storage areas described above, and when rewriting the contents of the first and second areas or the contents of the first and second storage units, the storage areas are added. Rewriting the contents of the storage area on the side different from the storage area indicated by the pointer, rewriting the accompanying pointer after the rewriting, whereby the control means rewrites the first storage unit, , Rewriting of the pointer attached to the first storage unit, and rewriting of the pointer attached to each of the first areas, and further rewriting of the second storage unit and the second area. The order of each rewriting, the rewriting of the pointer attached to the second storage unit, and the rewriting of the pointer attached to each of the second areas. Rewriting and performing (claim 4).

According to this, even if rewriting is not completed due to a power failure while the information for managing the array is being rewritten, if the pointer is not rewritten, the current information is not rewritten. Has no effect. Therefore, rewriting may be performed again. Further, when the power is cut off while the pointer is being rewritten, since the pointer does not match between the control unit and each disk device, this can be detected and the pointer can be correctly rewritten by majority decision. The rewriting is completed by the rewritten pointer. Therefore, it is possible to improve the reliability of the array configuration information for managing the array in consideration of a situation such as a power failure.

In this case, even if the second storage area (the second area, the second storage section) does not exist, the array configuration information for managing the array in the event of a power failure similarly. Can be improved in reliability (claim 5).

[0022]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing a configuration of a disk array device according to one embodiment of the present invention. As shown in FIG. 1, the disk array device has a disk array 12 and a controller 11 for controlling the disk array. here,
A plurality of disk arrays may be provided below the controller 11.

The controller 11 has an array configuration information storage memory 13.
Includes master information 14 corresponding to the disk array 12.
And slave information 1 having the same contents as master information 14
5 are stored. The master information 14 is array configuration information of the disk array 12, and the controller 11
Normally, this content is recognized as array configuration information, so that it will be referred to as master information here. On the other hand, the slave information 15 is not normally used, and the master information 1 is not used.
4 is used when an inconvenience such as data destruction occurs, and is referred to herein as slave information.

The contents of the master information 14 and the slave information 15 include, for example, the ID of the disk array 12 and the RAID.
The information includes the level, the number of disk devices belonging to the disk array 12, the storage capacity of the disk array 12, the stripe size, the write cache method, and the read cache method. Note that a nonvolatile semiconductor read / write memory can be used as the array configuration information storage memory 13, and the contents can be rewritten as needed.

The disk array 12 is located below the controller 11.
If there is another disk array, master information and slave information corresponding to the disk array are stored in the array configuration information storage memory 13.

The disk array 12 is, for example, 3
Although it is assumed that the disk drive is composed of two disk devices 16, 17, and 18, the number may be other than three. Each of the disk devices 16, 17, and 18 has a disk array 1
2 are stored as master information 19, 21, and 23, and slave information 20, 22, and 24 having the same contents as the master information 19, 21, and 23, respectively, are stored. Therefore, the master information 19, 21, 23
Has the same content as the master information 14, and the slave information 20, 22, and 24 has the same content as the slave information 15.

The respective disk devices 16, 1
The master information 19, 21, and 23 of 7 and 18 have exactly the same contents because they are the array configuration information of the disk array 12.
It may have unique contents such as an ID for each of 6, 17, and 18. This is the same for each of the slave information 20, 22, and 24.

The controller 11 includes a disk array 12
It always controls writing and reading of data to and from the master information 14 and the slave information 15 of the controller 11 and the master information 19 and 21 of the disk devices 16, 17, 18 when necessary.
23, and also controls rewriting of the slave information 20, 22, and 24. In addition, as will be described later, it is also in charge of prior processing required for these rewrites. Such rewriting and processing are performed, for example, by using a ROM (read only memory).
ory) (not shown), and software can be realized by preparing and executing the software.

The disk array device shown here is usually used by being connected to a host system (not shown). Therefore, the master information 14, 1
9, 21, 23, slave information 15, 20, 22, 24
When necessary, rewriting when necessary and prior processing required for this rewriting can be performed on the host system side.

According to the disk array device shown here, the controller 11 and each disk device 16, 1
7 and 18 store slave information of the same content in addition to the master information. Therefore, even when data is destroyed in the master information, it can be repaired by the corresponding slave information. Therefore, it is possible to restore a disk device in which data corruption has occurred in the master information so that the disk device always belongs to the original disk array.

Note that such a repair function is not achieved in an extremely rare case where both the master information and the corresponding slave information are destroyed, but the possibility of such both destruction is reduced at all. To do so, the master information 1 in each of the disk devices 16, 17, 18
For example, a method of setting the physical storage positions of the slave information 9, 21, and 23 and the slave information 20, 22, and 24 sufficiently separated from each other can be adopted.

Next, an operation example in the case where data destruction occurs in one or more of the master information 14, 19, 21, and 23 in this embodiment will be described with reference to FIG.
FIG. 11 is a flowchart showing a data repair operation to cope with a case where data destruction occurs in one or more of the master information 14, 19, 17, and 18.

First, when the system is started, the controller 11 reads out the master information 14, 19, 21, and 23 and checks whether or not all of them match (step 31). If all match, this processing is terminated without performing data recovery operation, not data destruction.

If they do not all match, grouping is performed each time they match (step 32). In the case where the results of the grouping are all groups for each unit, there is no point in performing the subsequent processing such as finding no basis for determining which is correct. , Master information 14, 19, 2
Slave information 15, 20, 22, 2 corresponding to 1, 23
4 and the slave information is set as new master information (step 33), and the grouping of step 32 is performed.

Using the result of the grouping in step 32, correct information is assumed by majority decision (step 3).
5). Here, for information having information not adopted in the majority decision, the master information and the slave information are exchanged (the same).

It is checked whether the information assumed to be correct and the information replaced by the master slave all match again (step 36). If they match, the assumed information is regarded as correct, and the other information is rewritten to the corresponding slave information (step 39). As a result, the master information in which the data destruction has occurred is repaired by the corresponding slave information.

In step 35, the result of the majority decision may result in the same number of candidates. In that case, it is not known at that time which of the candidates is likely to be correct. Therefore, assuming that the specific candidate is correct, the master / slave is replaced for the information of the other candidates to check for a match (step 35). If a match is obtained, the candidate is determined to be correct, and data destruction is repaired by rewriting in step 39.

If no match is obtained in step 36, it is considered that the candidate is not correct. Therefore, another candidate is picked up in step 37 and the process returns to step 35, and the master-slave is replaced for information other than the other candidate. Look for a match. By executing such a loop, it is possible to select a correct one in a case where a plurality of candidates are generated, and to repair the master information considered to have caused the data destruction by using the corresponding slave information.

If no match is obtained in step 36 and there are no candidates that can be assumed to be correct in step 37, the master information cannot be automatically restored as described above. A message prompting the configuration is issued (step 38), and this process ends.

Next, in the embodiment shown in FIG. 1, the master information 14 and the slave information 1
5. Master information 19, 21, 2 on each disk device
3. The operation when all the slave information 20, 22, and 24 are rewritten will be described with reference to FIG. FIG.
4 is a flowchart showing such a rewriting operation. In addition,
Such rewriting is necessary when completely recreating the configuration of the disk array, but is also necessary when only reconfiguring the disk array (for example, setting the read cache method).

As shown in FIG. 3, first, the controller 11
The above master information 14 is rewritten. Next, the master information 19, 21, 2 on each of the disk devices 16, 17, 18
Rewrite 3. Here, the order may be made in a specific order. For example, each disk device 16, 17,
If an ID number is assigned to the reference numeral 18, the ID is rewritten in ascending order of the ID number.

Next, the slave information 15 on the controller 11 is rewritten. Further, each of the disk devices 16, 1
The slave information 20, 22, and 24 on 7 and 18 are rewritten. The master information 19, 21, and 23 also correspond to this order.
Is the same as

By performing rewriting in the above order, the following effects can be obtained.

In other words, if the rewriting procedure is determined in this way, the array configuration information will not be lost even if the rewriting is not completed due to, for example, a power failure during the rewriting. That is, after the power is restored, the result of the rewriting can be checked in the predetermined order, and the rewriting can be completed based on the checked result or the state before the rewriting can be returned. Therefore, it is possible to improve the reliability of the array configuration information for managing the array in consideration of a situation such as a power failure.

Next, the master information 14, 19, 21, 2 applicable in the embodiment shown in FIG.
3, an example of the data structure of the slave information 15, 20, 22, and 24 and an operation of rewriting the information suitable for having this data structure will be described with reference to FIGS.

FIG. 4 shows the master information 14, 19, 21,
FIG. 23 is a diagram showing an example of the data structure of slave information 23, 15, 20, 22, and 24; This data structure is for further enhancing the resistance to a situation such as a power failure during rewriting and improving the reliability.

In FIG. 4, master information 51 and slave information 52 are master information 14, 19 and 2 respectively.
1, 23 and slave information 15, 20, 22, and 24. Both the master information 51 and the slave information 52 have pointers 53 and 56 and a first array configuration information 5 as a structure.
4, 57, and second array configuration information 55, 58.

The pointers 53 and 56 have contents indicating which of the first and second array configuration information is actually used. Therefore, in the master information 51 and the slave information 52, the array configuration information indicated by the contents of the pointers 53 and 56 is the same as the master information 14, 19, 21, 23 or the slave information 15, 2 described above.
Functions as 0, 22, and 24.

When rewriting the array configuration information that actually functions (uses) with respect to the master information 51 and the slave information 52, the array configuration information different from the array configuration information indicated by the pointer is rewritten. Is rewritten so as to point to the newly rewritten array configuration information.

That is, according to this rewriting,
The array configuration information before being rewritten and used is held inside the master information or the slave information as the other array configuration information.

FIG. 5 is a flow chart showing the rewriting operation of the array configuration information suitable for having such a data structure of master information and slave information.

In the figure, steps 61 to 64 are performed in place of steps 41 and 42 in FIG. 3, and steps 65 to 68 are performed.
Is performed in place of steps 43 and 44 in FIG.

As shown in FIG. 5, first, the controller 1
The master information on No. 1 which is not used is rewritten (step 61). Next, each of the disk devices 16, 1
Rewrite the unused master information in steps 7 and 18 (step 62). Here, the order of the disk devices 16, 17, and 18 can be determined in a predetermined order, for example, in ascending order of their ID numbers.

Next, the pointer in the master information on the controller 11 is rewritten so as to point to a new one (step 63). Further, each of the disk devices 16, 1
The pointers in the master information of Nos. 7 and 18 are rewritten so as to point to the new one (step 64). Also in this case, the order of the disk devices 16, 17, 18 can be determined in a predetermined order, for example, in ascending order of the ID number. Thus, the rewriting of the master information is completed.

Steps 65 to 68 are similarly executed for the slave information. That is, first,
The slave information on the controller 11 which is not used is rewritten (step 65). Next, the slave information on each of the disk devices 16, 17, and 18, which is not used, is rewritten (step 66). here,
The order of the disk devices 16, 17, and 18 can be determined in a predetermined order, for example, in ascending order of their ID numbers.

Next, the pointer in the slave information on the controller 11 is rewritten so as to point to a new one (step 67). Further, each of the disk devices 16, 1
The pointers in the slave information 7 and 18 are rewritten so as to point to the new one (step 68). Also in this case, the order of the disk devices 16, 17, 18 can be determined in a predetermined order, for example, in ascending order of the ID number. Thus, the rewriting of the slave information is completed.

According to the data structure and the rewriting operation described above, the following effects are exhibited. That is, when the rewriting is not completed due to a power failure during the rewriting of the array configuration information, if the pointer has not been rewritten, the currently used information is not affected. Therefore, in order to perform rewriting, a rewriting operation may be performed again.

When the power is cut off while the pointer is being rewritten, the pointers do not match between the controller 11 and the disk units 16, 17, and 18.
This can be detected, whereby the pointer can be correctly rewritten by majority decision. The rewriting is completed by the rewritten pointer.

When it is not possible to determine what is considered correct by majority vote, it is necessary to rewrite the pointers of the disk devices 16, 17, and 18 assuming that the pointer on the controller 11 has been rewritten first. it can.

Therefore, the reliability of the array configuration information can be improved in consideration of a situation such as a power failure.

In this case, even in the conventional case where there is no slave information, the structure of the master information is set to the pointer and the two array configuration information, and the rewriting is executed as shown in steps 61 to 64. In a situation such as a power failure, the reliability of the array configuration information can be improved.

[0063]

As described in detail above, according to the present invention,
The storage area of the array configuration information existing in the control unit and the plurality of disk devices is a first storage area, and the storage area of the same array configuration information in the control unit and the plurality of disk devices is separately stored in the second storage area. Therefore, even if the information in the first storage area is destroyed, it can be restored using the information in the second storage area. Therefore, the reliability of the array configuration information for managing the array can be improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a disk array device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a data repair operation to cope with a case where data destruction occurs in one or more of the master information 14, 19, 17, and 18 in the disk array device shown in FIG.

3 shows master information 14 and slave information 1 on a controller 11 in the disk array device shown in FIG.
5. Master information 19, 21, 2 on each disk device
3. Flow chart showing the operation when all the slave information 20, 22, 24 are rewritten.

4 shows master information 14, 19, 21, and 23, slave information 15, and the like in the disk array device shown in FIG.
The figure which shows the example of the data structure of 20, 22, 24.

FIG. 5 is a flowchart showing an operation of rewriting array configuration information suitable for having the data structure of master information and slave information as shown in FIG. 4;

[Explanation of symbols]

 11 Controller 12 Disk Array 13 Array Configuration Information Storage Memory 14 Master Information 15 Slave Information 16, 17, 18 Disk Unit 19, 21, 23 Master Information 20, 22, 24 Slave Information

Claims (5)

[Claims] A plurality of disk devices constituting a disk array; and a control unit for controlling the disk array, wherein the plurality of disk devices store information for managing the disk array. And a second area, wherein the control section stores first and second storage sections for storing the information for managing the disk array; and the first area for each of the plurality of disk devices. A means for rewriting the information in the first area with information in the second area; and a means for rewriting the information in the first storage section with information in the second area. 2. A system comprising: a plurality of disk devices constituting a disk array; and a control unit for controlling the disk array, wherein the plurality of disk devices store information for managing the disk array. And a second area, wherein the control unit is configured to store first and second storage units for storing the information for managing the disk array; and each of the information stored in the first area and First comparing means for comparing the information stored in the first storage unit with each other; means for estimating correct information by majority decision when the compared results do not match; the first area or the first area; In which the comparison is performed using the information of the corresponding second area or the second storage unit instead of the information of the storage unit having no estimated correct information among the storage units of And when the comparison result of the second comparison means matches, the information of the first area or the first storage unit that does not have the estimated correct information is stored in the corresponding second area. Or a means for rewriting the information in the second storage unit. 3. The control section further includes control means for rewriting the first and second areas and the first and second storage sections and for controlling the rewriting procedure. 3. The disk array device according to claim 1, wherein the contents are rewritten in the respective procedures of the first storage unit and the first area, the second storage unit, and the second area. 4. The first and second areas and the first and second storage sections have first and second storage areas, respectively, and are provided in the first and second storage areas. When a storage area for storing a pointer indicating one of them is attached, and when the contents of the first and second areas or the contents of the first and second storage units are rewritten, the storage indicated by the attached pointer is added. Rewriting the contents of the storage area on the side different from the area, rewriting the accompanying pointer after the rewriting, whereby the control means can rewrite the first storage unit, rewrite the first area, The rewriting of the pointer attached to the first storage unit and the rewriting of the pointer attached to each of the first areas are performed in this order, and the rewriting of the second storage unit is further performed. 4. The method according to claim 3, wherein rewriting of each of the second areas, rewriting of a pointer attached to the second storage unit, and rewriting of a pointer attached to each of the second areas are performed in this order. Disk array device. 5. A storage device comprising: a plurality of disk devices constituting a disk array; and a control unit for controlling the disk array, wherein the plurality of disk devices have an area for storing information for managing the disk array. The control unit includes: a storage unit that stores the information for managing the disk array; and a control unit that rewrites each of the area and the storage unit and controls a rewriting procedure. Each of the area and the storage unit has a first and a second storage area, respectively, and has a storage area for storing a pointer pointing to one of the first and the second storage areas, respectively, When rewriting each of the areas or the storage unit, the storage area on the side different from the storage area indicated by the accompanying pointer Rewriting containers, and, by applying this rewriting, rewriting of the storage unit,
A disk array device, wherein rewriting is performed in the order of rewriting of each area, rewriting of a pointer attached to the storage unit, and rewriting of a pointer attached to each area.