JP2005250551A - Disk array apparatus and transfer data control method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To prevent the data breakdown of a disk by error-detecting the write transfer of correct data to any erroneous disk device caused by the failure of a bus or the internal failure of a disk array control part. <P>SOLUTION: This disk device 20 which has received transfer data from a disk array control part 10 performs the calculation of a data assurance code by including not only the transfer data but also its own unique ID 25. When data transfer to a correct disk device 20 is executed, ID used for generating the data assurance code by the disk array control part 10 is matched with the unique ID 25 of the disk device 20 which executes the check of the data assurance code, and write data are written to the disk. When the transfer to the erroneous disk device 20 is executed, and the check error of the data assurance code is detected, the write data are not written to the disk but discarded so that the data writing to the erroneous disk device can be prevented. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a disk array device and a transfer data control method, and more particularly to a disk array device and a transfer data control method in which each disk device has a unique identification number and each disk device has a data guarantee check function.

  In general, a disk array device is composed of a plurality of disk devices and a disk array control unit for controlling the plurality of disk devices, and the disk device is a disk unit for storing data and a disk control for controlling data storage to and reading from the disk unit. The transfer data is divided into physical sector units of the disk unit by the disk array control unit, and the transfer data divided into physical sector units is transferred to a plurality of disk devices and stored in the disk unit, RAID etc. Perform the configuration.

  Further, data and a data guarantee code are stored in the physical sector of each disk device, and a CRC (Cyclic Redundancy Check) code is usually used as the data guarantee code.

  In a disk array device that performs a data guarantee code check only when reading data from the disk unit, if the data is destroyed due to a bus failure or the like when writing data, the corrupted illegal data is written to the disk unit, causing a large amount at once. When data writing occurs, there is a problem that data recovery is impossible even if RAID is used.

  In the technique described in Patent Document 1, each disk control unit is provided with a data guarantee code check function (eg, CRC check function), so that data guarantee code check is performed before writing data to the disk part. When the data guarantee code check error occurs, the problem that the data cannot be recovered by writing illegal data is solved by not performing the writing.

However, in general, the data guarantee code guarantees only the contents of the associated transfer data. If correct data is transferred to the wrong transfer destination due to a bus failure or a failure in the disk array control unit, Patent Document 1 In the technology described in the above, it is impossible to detect an error with the data guarantee code check function, and the above-mentioned problem cannot be solved.
JP 2000-105675 A

  In a disk array device that does not perform a data guarantee code check when writing data to the disk unit, it is impossible to detect illegal data when writing to the disk unit, and thus writing illegal data cannot be avoided.

  In addition, in the technique described in Patent Document 1, it is possible to detect an error in the data that is the guaranteed range of the data guarantee code, but the correct data generated due to a bus failure or a failure in the disk array control unit is incorrect. Data writing to the disk device cannot detect errors.

  An object of the present invention is to enable detection of incorrect data even when correct write data is transferred to an incorrect disk device due to a bus failure or an internal failure of the disk array control unit. To avoid data corruption.

  The disk array device of the present invention is composed of a disk array controller (10 in FIG. 1) and a plurality of disk devices (20, 30, 40, 50 in FIG. 1), and each disk device has a unique ID (FIG. 1). 25, 35, 45, 55), and when the disk array control unit (10 in FIG. 1) calculates the data guarantee code, the unique ID of the data transfer destination disk device (see FIG. 1) 25, 35, 45, and 55), and a data guarantee code generator (12 in FIG. 1) for calculating the data guarantee code, and each disk controller (21, 31, 41, 51 in FIG. 1). When the data guarantee code check is performed, the data guarantee code is calculated and generated from the transfer data and the own unique ID (25, 35, 45, 55 in FIG. 1), and from the disk array control unit (10 in FIG. 1). Transferred A data guarantee code check unit (22, 32, 42, 52 in FIG. 1) for comparing and checking the generated data guarantee code and the data guarantee code comparison check result is inconsistent. Includes a disk access control unit (23, 33, 43, and 53 in FIG. 1) that discards transfer data without writing it to the disk unit (24, 34, 44, and 54 in FIG. 1).

<Action>
When performing data write transfer from the disk array control unit (10 in FIG. 1) to the disk device (20, 30, 40, 50 in FIG. 1), the data is the disk unit (24, 34, 44, 54 in FIG. 1). The transfer data is accompanied by a data guarantee code.

  When the disk array controller (10 in FIG. 1) generates and calculates a data guarantee code associated with the transfer data, a transfer destination disk device unique ID (25, 35, 45, 55 in FIG. 1) is added to the transfer data. Thus, not only the transfer data but also the transfer destination disk device unique ID is included in the guarantee range of the data guarantee code.

  The transfer data transferred from the disk array control unit (10 in FIG. 1) to the transfer destination disk device (20, 30, 40, 50 in FIG. 1) is only the write data and the data guarantee code, and the transfer destination disk device ( IDs 20, 30, 40, and 50) in FIG. 1 are not transferred.

  The disk control unit (21, 31, 41, 51 in FIG. 1) that has received the transfer data from the disk array control unit (10 in FIG. 1) assigns its own unique ID (25, 35, 45, 55 in FIG. 1) to the transfer data. ) Is added to calculate the data guarantee code, and is compared with the data guarantee code transferred from the disk array control unit.

  When data is transferred to the wrong disk device due to a bus failure or an internal failure of the disk array control unit, the ID used for data guarantee code generation by the disk array control unit (10 in FIG. 1) and the disk control unit (21 in FIG. 1) , 31, 41, 51), the IDs (25, 35, 45, 55 in FIG. 1) do not match, a data guarantee code mismatch error is detected, and if a data guarantee code mismatch error is detected, disk access By discarding the transfer data by the control unit (21, 31, 41, 51 in FIG. 1), it becomes possible to avoid illegal data writing to the disk device.

  According to the present invention, in the write transfer from the disk array controller to the disk device, the ID indicating the disk device that is the transfer destination is also included in the guaranteed range of the data guarantee code, so that the bus failure or the disk array controller can If the correct data is transferred to the wrong disk device due to the failure of the data, it can be detected as a data guarantee code mismatch error, and it is possible to prevent the problem that illegal data is written to the disk device. Become.

  Next, embodiments of the present invention will be described in detail with reference to the drawings.

<Description of configuration>
FIG. 1 shows an embodiment of a disk array apparatus according to the present invention. The disk array apparatus of this embodiment is connected to an upper host 00 and performs disk array control and data transfer control in accordance with commands from the upper host 00. The disk array controller 10 and the disk array controller 10 are connected to the disk array controller 10 to transfer data to and from the disk array controller 10. The disk array controller 10 and each disk device 20, 30, 40 and 50 are connected by buses 26, 36, 46 and 56.

  The disk array control unit 10 divides data into physical sector units of the disk units 24, 34, 44, and 54 and divides the ID selected by the transfer destination disk device unique ID selection unit 11 when performing data write / transfer. The data guarantee code generation unit 12 calculates and generates a data guarantee code using the data with the ID added, and the data transfer unit 13 transfers the data and the data guarantee code to the destination disk device. To do.

  The data transferred from the disk array control unit 10 to the disk devices 20, 30, 40, 50 is only the write data and the data guarantee code, and the ID used for the data guarantee code calculation generation is not transferred.

  In the disk devices 20, 30, 40, 50, the data guarantee code is calculated and generated by adding the unique IDs 25, 35, 45, 55 to the data transferred from the disk array controller 10, and the generated data guarantee code is The data guarantee code transferred from the disk array control unit 10 is compared with the data guarantee code check units 22, 32, 42, 52, and the disk access control units 23, 33, 43, 53 check the comparison of the data guarantee code. As a result, if a mismatch error has occurred, the write data transferred from the disk array control unit 10 is discarded, and if no error has occurred, writing to the disk units 24, 34, 44, 54 is performed.

<Description of operation>
Next, the operation of FIG. 1 will be described.

  An operation in which data of one physical sector unit length that should be transferred from the disk array control unit 10 to the disk device 20 is erroneously transferred to the disk device 30 and a data guarantee code mismatch error occurs in the disk device 30 will be described. To do.

  Further, in describing the operation of FIG. 1, the following settings are assumed.

  It is assumed that one physical sector unit length is 520 bytes, 512 bytes of data and 8 bytes of CRC code are included, that is, an 8-byte CRC code is used as the data guarantee code, and the CRC code is calculated from the same data. Assuming that CRC codes that can be expected to be 0 when the generated CRCs are calculated using the same CRC generation algorithm are used, each disk device unique ID 25, 35, 45, 55 is an 8-byte value. Assume that

  First, the ID selection unit 11 of the disk array control unit 10 selects the unique ID of the write data transfer destination disk device 20 (all values A in hexadecimal and 8 bytes wide).

  Next, the selected ID (all values A of 8 byte width in hexadecimal) is added after the 512 bytes of transfer data, resulting in a total data length of 520 bytes (A0 in FIG. 2), and ID (in hexadecimal) CRC calculation is performed on the 520-byte data to which all values A having an 8-byte width are added, and the data guarantee code generation unit 12 generates a CRC.

  The data transfer unit 13 adds the CRC generated after the 512-byte data to generate transfer data (B0 in FIG. 2), and transfers the data to the transfer destination disk device 20.

  Here, if it is operating normally, data transfer is performed to the disk device 20, but in this description, it is assumed that data transfer is performed to the disk device 30 due to a failure of the data transfer unit 13.

  The disk device 30 that has received the transfer data from the disk array control unit 10 has its own unique ID 35 (all values B of 8 bytes in hexadecimal notation) after the 512 bytes of the data part in the transfer data from the disk array control unit 10. Value) is added, and a CRC code transferred from the disk array control unit 10 is added after that to generate a total of 528 bytes of data (C0 in FIG. 2), and for the generated 528 bytes of data The data guarantee code check unit 32 performs CRC calculation.

  Here, since the value of the ID part is different from the data generated by the disk array control unit 10, the calculation result in the data guarantee code check unit 32 is not 0, but a CRC error is detected and the data guarantee is performed. The CRC error detected by the code check unit 32 is notified to the disk access control unit 33, and the disk access control unit 33 discards the transfer data from the disk array control unit 10 without writing it to the disk unit 34.

  With the above operation, the problem that the write data transferred to the wrong disk device is written to the disk unit can be solved.

<Other embodiments of the invention>
Next, another embodiment of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 3, XBARs 60, 70, 80, 90 are connected between the disk array controller 10 and the disk devices 20, 30, 40, 50, and the disk array controller 10 and XBARs 60, 70, 80, 90 are connected. Are connected by data buses 61, 71, 81, 91 and address buses 62, 72, 82, 92.

  The XBARs 60, 70, 80, 90 have a switch function for determining a data transmission route between the disk array control unit 10 and the disk devices 20, 30, 40, 50, and each XBAR 60, 70, 80, 90 It is connected to the disk device, and the specific disk device 20, 30, 40, 50 is designated by the address bus 62, 72, 82, 92 from the disk array control unit 10 and data is sent to the XBAR 60, 70, 80, 90 The XBARs 60, 70, 80, 90 transfer data to the disk devices 20, 30, 40, 50 designated by the address buses 62, 72, 82, 92.

  The XBARs 60, 70, 80, 90 and the disk devices 20, 30, 40, 50 each have a unique unique ID, and the disk array control unit 10 includes all the XBARs 60, 70, 80, 90 and the disk devices 20, 30, 40 and 50 unique ID information is held.

  That is, when transferring write data from the disk array controller 10 to a specific disk device 20, 30, 40, 50, the ID information of the XBARs 60, 70, 80, 90 used as the data transfer route and the final transfer destination It is possible to add the ID of a certain disk device 20, 30, 40, 50 to the data and calculate the data guarantee code using the data with the added ID, and only the data is within the compensation range of the data guarantee code. The unique IDs of the XBARs 60, 70, 80, and 90 and the unique IDs of the disk devices 20, 30, 40, and 50 can be included.

  Furthermore, if each disk device 20, 30, 40, 50 knows the unique ID of XBAR 60, 70, 80, 90 to be used as a write data transfer route from the disk array controller 10 by setting or the like in advance, It is possible to use the data guarantee code check by adding the own unique ID as the target data of the data guarantee code check and the unique ID of the XBAR 60, 70, 80, 90 used as the route after the data part in the transfer data. In this case, the check including the switch route can be performed, and the data guarantee capability is further improved.

  If the XBAR 60, 70, 80, 90 knows the unique ID of the transfer destination disk 20, 30, 40, 50 by setting or the like, the XBAR 60, 70, 80, 90 sends the data from the disk array control unit 10. It is also possible to perform a data guarantee code check on the transfer data. If the result of the data guarantee code check is a data guarantee code mismatch error, the transfer data is transferred to the disk devices 20, 30, 40, 50. It is also possible to discard without.

  Further, in the disk array apparatus having the same configuration as that shown in FIG. 3, it is possible to change the data guarantee code of the transfer data by using the XBARs 60, 70, 80, 90.

  In this case, the disk array control unit 10 has unique ID information of all the XBARs 60, 70, 80, 90, and when the disk array control unit 10 calculates the data guarantee code, the XBARs 60, 70, 80 follow the data. The ID is selected from the 90 unique IDs by the transfer route, the selected ID is added to the data to generate the data guarantee code, and the write data and the unique IDs of the XBARs 60, 70, 80, 90 are guaranteed as the transfer data. The data guarantee code included in is transferred to the XBARs 60, 70, 80, 90 as the transfer destinations.

  In XBAR 60, 70, 80, 90, when checking the data guarantee code of the transfer data from the disk array controller 10, the calculation is performed by adding its own unique ID to the end of the transfer data as the calculation target of the data guarantee code. When the data transmission route from the disk array control unit 10 to the XBARs 60, 70, 80, 90 can be verified and the result of the data guarantee code check is a mismatch error, the transfer from the disk array control unit 10 It is also possible to discard the data without transferring it to the disk units 20, 30, 40, 50.

  Also, all XBARs 60, 70, 80, 90 have unique ID information of all connected disk devices 20, 30, 40, 50, and the write data from the disk array control unit 10 is a data guarantee code check. If the result is correct (or if the transfer data is not discarded even if the result of the data guarantee code check is an error), a plurality of disk devices 20, 30, 40, 50, the data transfer destination is selected and data is transferred. At this time, the data guarantee code is generated by the XBARs 60, 70, 80, 90, and the generated data guarantee code is attached to the data to the disk devices 20, 30, 40. , 50.

  Specifically, the data guarantee code transferred from the disk array control unit 10 is discarded when the data guarantee code check in the XBAR 60, 70, 80, 90 is completed, and newly generated by the XBAR 60, 70, 80, 90. The data guarantee code is transferred along with the transfer data. At this time, the data guarantee code generated by the XBARs 60, 70, 80, 90 guarantees the unique ID of the disk device 20, 30, 40, 50 which is the final transfer destination.

  In this case, the failure of the data transmission route has occurred between the disk array controller 10 and the XBARs 60, 70, 80, 90, or has occurred between the XBARs 60, 70, 80, 90 and the disk devices 20, 30, 40, 50. It is possible to improve the error resolution ability.

  In addition, the said form is an example for implementing this invention, and is not limited to this. Therefore, various modifications can be made without departing from the scope of the present invention.

It is a block diagram of the Example of the best form of this invention. It is a data format used in each part of the disk array device shown in the embodiment of the best mode of the present invention. It is a block diagram of the Example of the other form of this invention.

Explanation of symbols

00 Host host 10 Disk array control unit 11 Transfer destination ID selector 12 Data guarantee code generation unit 13 Data transfer unit 20, 30, 40, 50 Disk device 21, 31, 41, 51 Disk control unit 22, 32, 42, 52 Data Warranty code check unit 23, 33, 43, 53 Disk access control unit 24, 34, 44, 54 Disk unit 25, 35, 45, 55 Disk device unique ID
26, 36, 46, 56 Data bus 60, 70, 80, 90 XBAR (switch part)
61, 71, 81, 91 Data bus 62, 72, 82, 92 Address bus A0 CRC generation calculation target data format B0 Transfer data format C0 CRC check target data format

Claims (7)

In a disk array device composed of a plurality of disk devices and a disk array control unit that controls the plurality of disk devices, each disk device has a unique ID, and each disk device includes a disk control unit and a disk unit. Configured,
The disk control unit of each disk device includes data check means for checking the validity of the data before writing the data transferred from the disk array control unit to the disk unit,
The disk array control unit holds unique IDs of all the disk devices to be transferred. When the disk array control unit generates a data guarantee code associated with transfer data, the disk array control unit Characterized by comprising means for selecting a unique ID and means for generating a data guarantee code from the selected unique ID and transfer data,
The data check means included in the disk controller of the disk device is characterized by comprising means for generating a data guarantee code from the data written to the disk unit transferred from the disk array controller and its own unique ID,
The data check means included in the disk controller of the disk device includes means for comparing the data guarantee code generated by the disk controller with the data guarantee code transferred from the disk array controller to the disk device. A disk array device characterized by this.   Provided in the disk device, and having means for notifying the occurrence of an error when the comparison result of the data guarantee code is inconsistent in the data checking means, and writing data to the disk unit when the error notification occurs 2. The disk array device according to claim 1, further comprising means for discarding without writing to the disk array device. A plurality of transmission route determination devices for determining a transmission route of data transferred between the disk array control unit and the disk device;
3. The disk array device according to claim 1, wherein the transmission route determination device has its own unique ID.   4. The disk array device according to claim 1, wherein a CRC is used as the data guarantee code. A transfer data control method for transferring data from the disk array control unit to the disk device in a disk array device comprising a plurality of disk devices and a disk array control unit for controlling the plurality of disk devices. ,
In the disk array control unit, comprising a step of selecting a unique ID of a disk device to which data is to be transferred, and a step of generating a data guarantee code from the selected unique ID and transfer data,
The disk device generates a data guarantee code from the data transferred from the disk array controller and its own unique ID, and the data guarantee code generated from the disk device and the disk array controller to the disk. A transfer data control method comprising a step of comparing a data guarantee code transferred to a device.   6. The method according to claim 5, further comprising a step of notifying that an error has occurred when the comparison result of the data guarantee code does not match in the step of comparing, and a step of discarding transfer data when the error notification occurs. Transfer data control method.   7. The transfer data control method according to claim 5, further comprising a route determination step for determining a transmission route for transferring data from the disk array control unit to the disk device.

Images