JP4920390B2 - Storage device - Google Patents

Description

  The present invention relates to a storage apparatus suitable for data backup.

  In the field of information systems, a backup that saves the storage contents of a storage device to a magnetic tape mounted in a tape library in preparation for the destruction or loss of programs and data due to computer hardware failure or software failure Has been done. In this case, the magnetic tape can be stored externally from the tape library, and by rotating the magnetic tape stored externally, it is possible to back up data exceeding the allowable amount of the tape library. Also, it is possible to prepare for disaster recovery (disaster recovery) by storing the magnetic tape in a safe or the like.

In recent years, along with the price reduction of magnetic disks such as hard disks, cases of using magnetic disks instead of magnetic tape to back up stored contents are increasing. In this case, RAID (Redundant Array of Inexpensive Disks) that reads and writes data in parallel on a plurality of magnetic disks is often used. For example, a system has been proposed in which a backup magnetic disk device is installed at a location distant from the magnetic disk device to be backed up and the stored contents are synchronized (for example, see Patent Document 1).
Japanese Patent Laid-Open No. 2002-7304

  However, when the storage contents are backed up using a magnetic disk as described above, the magnetic disk cannot be stored externally and therefore cannot be rotated like a magnetic tape. For this reason, backup data exceeding the maximum capacity of the magnetic disk cannot be stored, and disaster recovery (disaster recovery) cannot be supported.

  In addition, when a backup magnetic disk device is installed at a remote place and backup data is stored in the magnetic disk device as in the system described in Patent Document 1, the cost required for the magnetic disk device and the distance between both magnetic disk devices The cost burden required for system construction, such as the cost required for the network, increases.

  An object of the present invention is to provide a storage apparatus that can cope with disaster recovery and store large-capacity backup data while reducing the cost burden.

A storage apparatus according to the present invention includes a master logical unit formed on a magnetic disk that holds data written to the apparatus main body, and a backup of data in the master logical unit formed on a magnetic disk that can be attached to and detached from a housing. A backup logical unit that holds data; and a controller that controls the writing of the backup data of the master logical unit to the backup logical unit by switching the state of the backup logical unit, the controller comprising the backup logical unit Is synchronized with the master logical unit, is divided to separate the backup logical unit from the master logical unit, and is a copy that copies data in the master logical unit to the backup logical unit. And a first management table for managing the offline state in which the magnetic disk constituting the backup logical unit is removed from the apparatus housing, and the backup logical unit is configured to determine whether the controller can use the backup logical unit. The controller includes a second management table for registering identification information for determining the backup logical unit, and when the controller requests export and import of the backup logical unit, the controller sets the backup logical unit in a divided state and When the notification that the backup logical unit has been exported and imported has been received, the state in the first management table is a divided state, and the internal information is stored in the case according to the identification information in the second management table. Does not exist in the backup The state of the backup logical unit is set to the offline state, the state in the first management table is the offline state, and the state of the backup logical unit existing in the enclosure is determined by the identification information in the second management table. and split state, characterized rotation to Rukoto the said backup logic unit.

  According to the present invention, it is possible to provide a storage apparatus that can cope with disaster recovery and store large-capacity backup data while reducing the cost burden.

  FIG. 1 is an overall configuration diagram of a storage apparatus 10 according to an embodiment of the present invention. As shown in FIG. 1, a storage apparatus 10 according to an embodiment of the present invention includes a controller 12, a master LU 13, and a plurality of backup LUs 14-1 to 14-n in a casing 11. Here, an LU (Logical Unit) is a logical unit on one RAID group composed of, for example, a plurality of magnetic disk devices (hereinafter referred to as “HDD” as appropriate).

  The controller 12 controls the entire apparatus, and writes data to the master LU 13 and backup LUs 14-1 to 14-n according to instructions from a host computer or the like to which the storage apparatus 10 is connected. Data is read from the master LU 13 and backup LUs 14-1 to 14-n. The master LU 13 holds data written in the storage apparatus 10 under the control of the controller 12.

  The backup LUs 14-1 to 14-n hold backup data of the master LU 13, and each backup LU 14-1 to 14-n is used to hold backup data having different backup timings. The backup data in each of the backup LUs 14-1 to 14-n is held by performing synchronization and disconnection described later between the master LU 13 and each of the backup LUs 14-1 to 14-n under the control of the controller 12.

  Although FIG. 1 shows the case where n backup LUs 14 are provided, the number of backup LUs 14 can be changed as appropriate. Further, as shown in FIG. 1, each backup LU 14-1 to 14-n shows a case where a logical unit on one RAID group is configured by two HDDs 15a and 15b.

  Each of the backup LUs 14-1 to 14-n is configured to be detachable from the housing 11 of the storage apparatus 10 under certain conditions. In the following, taking out each backup LU 14-1 to 14-n to the outside of the chassis 11 is called “export”, and the operation is called “export operation”. Also, inserting each of the backup LUs 14-1 to 14-n into the housing 11 is called “import”, and the operation is called “import operation”.

  Here, the state transition in the backup LU 14 will be described with reference to FIG. FIG. 2 is an explanatory diagram of the state transition of the backup LU 14 that the storage apparatus 10 has. As shown in FIG. 2, each of the backup LUs 14-1 to 14-n has four states, SPLIT, COPY, SYNC, and OFFLINE.

  The SPLIT state indicates a state where the backup LU 14 exists in the housing 11 and is separated from the master LU 13. When the synchronization or copying in the SYNC state or the COPY state is interrupted and the master LU 13 and the backup LU 14 are disconnected, the state transits from the SYNC state or the COPY state to the SPLIT state. When a new backup LU 14 is generated in the storage apparatus 10, the generated backup LU 14 is in the SPLIT state.

  The COPY state indicates a state where data is being copied from the master LU 13. When copying of data from the master LU 13 to the backup LU 14 is started, the state transits from the SPLIT state to the COPY state. In this case, only the difference data between the master LU 13 and the backup LU 14 is copied to the backup LU 14. The copy of the difference data will be described later.

  The SYNC state indicates a state in which the master LU 13 is synchronized. When copying of data from the master LU 13 to the backup LU 14 is completed, the state transits from the COPY state to the SYNC state. In this SYNC state, data is also written to the backup LU 14 in synchronization with the writing of data to the master LU 13.

  The OFFLINE state indicates a state in which the backup LU 14 is removed from the housing 11. When exported from the SPLIT state to the outside of the housing 11, the SPLIT state transitions to the OFFLINE state. Conversely, when imported from the outside of the housing 11 to the inside thereof, the state transits from the OFFLINE state to the SPLIT state.

Such state transition in the backup LU 14 is managed by the controller 12. As shown in FIG. 3, the controller 12 has a backup LU status table (first management table) 16 held by itself and backup LUs held by the HDDs 15a and 15b constituting each backup LU 14-1 to 14-n. Configuration table (second management
Table) The state transition in the backup LU 14 is managed with reference to 17a and 17b. Here, the contents of the backup LU status table 16 and the backup LU configuration tables 17a and 17b will be described.

  FIG. 4 is a data configuration diagram showing an example of data registered in the backup LU status table 16. As shown in FIG. 4, the backup LU status table 16 is associated with the numbers assigned to the backup LUs 14-1 to 14-n (hereinafter referred to as “backup LU numbers”), and is associated with the backup LUs 14-1 to 14-. The state of n is registered. The state of the backup LU 14 registered in the backup LU state table 16 is updated one by one according to the state transition.

  FIG. 5 is an explanatory diagram of data registered in the backup LU configuration table 17. As shown in FIG. 5, each backup LU configuration table 17 includes the backup LU number of the backup LU 14 to which the corresponding HDD 15 belongs, the identification information of the HDD, the RAID level of the backup LU 14 configured by the HDD 15, the stripe size, Information on all HDDs 15 constituting the backup LU 14 is held.

  By reading the backup LU configuration table 17 shown in FIG. 5 from each HDD 15, the controller 12 can acquire information about the backup LU 14 to which the HDD 15 belongs, RAID configuration information, and other HDDs 15 constituting the backup LU 14. It can be confirmed whether or not the HDDs 15 constituting the backup LU 14 are all in the housing 11.

  Next, operations in the storage apparatus 10 when exporting the backup LU 14 from the chassis 11 and importing the backup LU 14 into the chassis 11 will be described with reference to FIG. FIG. 6 is a sequence diagram for operations when exporting and importing the backup LU 14.

  FIG. 6 shows the relationship between the user operation and the processing procedure of the controller 12. The import operation and the export operation are performed by the same processing procedure, and in many cases, performed at the same time. In the following, a case where the import operation and the export operation are performed at the same time will be described. However, the import operation and the export operation can be performed separately.

  When the user wishes to export and import the backup LU 14, first, an export / import request is output to the controller 12 (S600). For example, the export / import request can be configured to be output in response to an operation of turning the door lock of the housing 11 storing the backup LU 14 in the opening direction.

  When the export / import request is received, the controller 12 refers to the backup LU status table 16 to determine whether there is a backup LU 14 in the SYNC status or the COPY status, and forcibly disconnects the backup LU status table if it exists. 16 is set to the SPLIT state (S601). Then, after setting all the backup LU status tables 16 to the SPLIT status, the user is notified of permission to export and import (S602). For example, export and import permission can be notified by unlocking the door of the housing 11.

  Upon receiving notification of permission for export and import, the user manually exports and imports the HDDs 15a and 15b constituting the backup LU 14 (S603). When the export and import of the backup LU 14 is completed, the controller 12 is notified of this (S604). For example, completion of export and import can be configured to be output in response to an operation of turning the door 11 in the direction of closing the door of the housing 11.

  After notifying permission of export and import in step S602, the controller 12 waits for an export and import completion notification from the user (S605). Upon receiving notification of completion of export and import, the controller 12 reads the backup LU configuration table 17 of all HDDs 15a and 15b belonging to the backup LU 14 in the chassis 11 (S606).

  Then, the controller 12 determines whether all the HDDs 15 constituting each backup LU 14 exist in the housing 11 (S607), and if any one HDD 15 is missing, notifies the user of an error. (S608). For example, an error to the user can be notified by notifying through a panel or LED.

  Next, the controller 12 determines whether there is a backup LU 14 in the SPLIT state of the backup LU state table 16 that does not exist in the chassis 11. If there is a corresponding backup LU 14, the backup LU 14 is switched to the OFFLINE state (S609).

  Further, the controller 12 determines whether there is a backup LU 14 in the OFFLINE state in the backup LU state table 16 and present in the chassis 11. If there is a corresponding backup LU 14, the backup LU 14 is switched to the SPLIT state (S610).

  When the processing up to S610 is performed, the export operation and the import operation of the backup LU 14 are completed, so the controller 12 notifies the user of the completion (S611). For example, the completion to the user can be notified by notifying with a panel or LED.

  In the embodiment of the present invention, the backup LU 14 composed of two HDDs 15a and 15b is attached to and detached from the housing 11 as a unit. At this time, it is preferable that the HDD 15a, 15b constituting the backup LU 14 is mounted on a holding member (for example, a magazine) that can hold the HDD 15a, and is detached from the housing 11 via the holding member. In this case, it is possible to prevent a situation where the wrong HDD 15 is mounted on the backup LU 14. In particular, it is preferable to arrange an elastic means such as a spring inside the holding member. In this case, it is possible to protect the HDD 15 from an impact that occurs at the time of attachment / detachment, and therefore it is possible to provide the storage apparatus 10 with more excellent portability of the HDD 15.

  Here, processing when the master LU 13 and the backup LUs 14-1 to 14-n are synchronized will be described. The controller 12 holds the difference table of the backup LU 14, and refers to the difference table to synchronize the master LU 13 with each of the backup LUs 14-1 to 14-n. Here, the difference table is held by the controller 12 for every backup LU 14, and the difference between data of the master LU 13 and each of the backup LUs 14-1 to 14-n is registered.

  Hereinafter, a method for managing difference data in the difference table will be described with reference to FIG. FIG. 7 is an explanatory diagram of a method for managing difference data in the difference table provided in the controller 12.

  For example, the master LU 13 is divided into blocks with a fixed size such as 1 Mbyte, and the divided blocks are associated with bits of each difference table. If the size of the master LU 13 is 1 Tbyte and the block size is 1 Mbyte, the size of the difference table is 1 Tbyte / (1 Mbyte * 8) = 125 Kbytes. Since the difference table is held for each backup LU 14, the total size of all the difference tables is 125 Kbytes * the number of backup LUs 14.

  When the backup LU 14 is in the SPLIT state or OFFLINE state and there is a write (write) to the master LU 13, the bit of the difference table corresponding to the block written in the master LU 13 is set to “1”. When the backup LU 14 is newly created, all bits of the difference table are initialized to “1”. In this way, the data difference between the master LU 13 and the backup LU 14 is expressed.

  When the state of the backup LU 14 transitions from the SPLIT state to the COPY state, the master LU 13 block corresponding to the portion where the bit of the difference table is “1” is copied to the backup LU 14. When copying to the backup LU 14 is completed, the bit of the difference table is set to “0”. When all the bits become “0”, the COPY state is changed to the SYNC state, and the synchronization using the difference data is completed. By this method, it is possible to copy differential data even in the backup LU 14 stored externally, and high-speed backup can be realized.

  As described above, in the storage apparatus 10 according to the embodiment of the present invention, the HDD 15 configuring the backup LU 14 can be attached to and detached from the chassis 11, and the controller 11 is OFFLINE with the HDD 15 configuring the backup LU 14 removed from the chassis 11. A backup LU status table 16 capable of managing the status is provided. As a result, since the HDD 15 constituting the backup LU 14 can be stored externally, the backup LU 14 can be rotated like a magnetic tape, so that backup data exceeding the maximum capacity of the storage apparatus 10 can be stored. It becomes. Further, since the backup LU 14 taken out from the casing 11 can be stored in a safe or a remote place, data can be recovered even when the entire casing 11 is destroyed due to a fire or disaster. Become. In this case, it is not necessary to install a backup magnetic disk device at a remote location and transmit backup data via the network, so that the cost burden required for the construction of the backup system can be suppressed. As a result, it is possible to cope with disaster recovery and save a large amount of backup data while reducing the cost burden.

  Further, the storage apparatus 10 includes a backup LU configuration table 17 for registering identification information of the HDD 15 in the HDD 15 configuring the backup LU 14, and the controller 11 stores the backup LU configuration table 17 when the HDD 15 configuring the backup LU 14 is attached or detached. Whether or not the backup LU 14 can be used is determined according to the contents read out. Thereby, even when the HDD 15 constituting the backup LU 14 is configured to be removable, it is possible to prevent the HDD 15 corresponding to the backup LU 14 from being lost.

  Further, since the controller 11 manages the data copied to the backup LU 14 stored externally using the difference table, the data can be backed up at a higher speed even after export and import. Is possible.

  Note that when copying differential data to the backup LU 14, it is also possible to encrypt the data in block units of the differential table. For encryption, for example, a secret key recorded in the firmware of the storage apparatus 10 can be used. In this case, decryption cannot be performed without a combination of the chassis 11 of the storage apparatus 10 and the firmware. However, it is necessary to use an encryption method in which the data size before encryption and the data size after encryption are the same size. As described above, by performing the encryption process when copying the differential data to the backup LU 14, it is possible to safely store the data in the backup LU 14 stored externally.

  In the above description, a case is shown in which a backup LU 14 configured with a plurality of HDDs 15 and functioning as a RAID device is provided. However, the storage medium provided in the storage device 10 can be appropriately changed. For example, the present invention can be applied to a plurality of randomly accessible storage media.

1 is an overall configuration diagram of a storage apparatus according to an embodiment of the present invention. FIG. 3 is an explanatory diagram showing state transition of a backup LU included in the storage apparatus according to the embodiment of the present invention. FIG. 3 is an explanatory diagram of a table in a storage device controller and a backup LU according to an embodiment of the present invention. The data block diagram which shows an example of the data registered into the backup LU state table in embodiment of this invention. Explanatory drawing of the data registered into the backup LU structure table 17 in embodiment of this invention. The sequence diagram about operation at the time of exporting and importing backup LU14 in embodiment of this invention. Explanatory drawing about the management method of the difference data in the difference table with which the controller in embodiment of this invention is provided.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Storage apparatus, 11 ... Case, 12 ... Controller, 13 ... Master LU, 14-1 to 14-n ... Backup LU, 15a, 15b ... Magnetic disk unit (HDD), 16 ... Backup LU status table, 17a, 17b: Backup LU configuration table.

Claims (1)

A master logical unit formed on a magnetic disk that holds data to be written to the device body;
A backup logical unit that is formed on a magnetic disk that can be attached to and detached from a housing and holds backup data of data in the master logical unit;
A controller that controls the writing of the backup data of the master logical unit to the backup logical unit by switching the state of the backup logical unit;
The controller is
A synchronization state in which the backup logical unit is synchronized with the master logical unit;
A split state for separating the backup logical unit from the master logical unit;
A first management table for managing a copy state in which data in the master logical unit is copied to the backup logical unit, and an offline state in which a magnetic disk constituting the backup logical unit is removed from an apparatus housing;
The backup logical unit includes a second management table for registering identification information for the controller to determine whether the backup logical unit can be used,
The controller puts the backup logical unit into a split state when there is a request to export and import the backup logical unit by a user operation, and receives a notification that the user has exported and imported the backup logical unit The state in the first management table is a divided state, and the state of the backup logical unit that does not exist in the enclosure is set to the offline state by the identification information in the second management table, and the first management table The state in the management table is an offline state, and the state of the backup logical unit existing in the chassis is divided according to the identification information in the second management table ,
A storage device which is characterized that you rotate the backup logic unit.

Images