JP2018173701A - Data management system, storage device, backup device, method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize differential backup of a block storage device with simple logic, and with less processing load and/or less resource load.SOLUTION: A storage device 61 comprises: a rest point counter update unit 611 which updates the value of a rest point counter, when a rest point is detected; a final update period update unit 612 which updates a final update period to the value of the rest point counter, when data are written in a block; and a rest point information provision unit 613 which outputs the rest point counter and the final update period. Moreover, a backup device 62 comprises: a rest point information acquisition unit 621 which acquires the rest point counter and the final update period, when performing backup; and an update block detection unit 622 which detects an update block, and holds the acquired rest point counter as a reference point counter, on the basis of the acquired final update period and the reference point counter.SELECTED DRAWING: Figure 12

Description

  The present invention relates to a data management system, a storage device, a backup device, a storage device backup method, and a storage device backup program for managing data stored in a storage device.

  As a method for backing up data stored in the storage device, a method called differential backup is widely used in order to reduce the amount of data transferred to and from the storage device and the amount of data stored on the backup device side. The differential backup is a method of backing up only the difference from the timing of the previous backup when the backup device performs backup.

  In differential backup, the backup device backs up (acquires and stores) all the data in the target storage device at the first time, and then after the previous backup in the target storage device after the second time. Only the updated part, that is, the differential data is backed up. When the backup device restores data (returns to the state of the previous backup), it sequentially applies the differential data from the second time to the specified backup time to the data acquired at the time of the first backup. Then, restore to the data state at the specified time.

  In order to prepare for a failure that makes the storage device unusable, the backup device is usually provided outside the storage device.

  When the backup device is provided outside the storage device, it is preferable that the updated data can be identified without knowing the detailed internal structure of the storage device to be backed up.

  For example, when the file system that controls the storage device manages the last update date and time of each file, such as a file system of a NAS (Network Attached Storage) storage device, the difference data is identified by referring to them. it can. For example, the backup device compares its own previous backup date and time with the last update date and time of each file, and determines whether or not backup of each file is necessary depending on whether or not the last update date and time of the file is newer than the previous backup date and time. Good. As described above, if the storage system manages the file update date and time, the difference data can be identified relatively easily. For example, the xcopy command widely used for the backup of Microsoft Windows (registered trademark) is an example of backup software that executes such an algorithm.

  However, many storage apparatuses such as a SAN (Storage Area Network) storage apparatus including a storage unit that manages a storage area in a predetermined block unit do not have a mechanism for managing the last update date and time. For this reason, for such a storage apparatus, the difference data cannot be specified from the data blocks stored in the storage unit using the last update date and time as described above.

  Hereinafter, in the present invention, a storage apparatus including one or more such storage units may be referred to as a “block storage apparatus”. It does not matter whether the storage unit in the block storage apparatus is a physical unit or a logical unit (LU). Hereinafter, each storage unit included in the block storage apparatus may be referred to as a block device.

  As a method for realizing the differential backup for the block storage device, for example, if all data is acquired every time and compared with the state at the time of the previous backup, the differential data can be specified. However, this method requires reading of all data and calculating the checksum, which not only incurs a load and time cost on the backup device side, but also places a load on the storage device side, which reduces the performance of the storage device. There is a problem of being connected.

  As another method, there is a method in which the block storage apparatus side has a mechanism for recording the last update time in association with each block (for example, see Patent Document 1).

  As another method, there is a method in which the backup device analyzes the updated portion by referring to internal data for snapshot management or the like held by the block storage device (see, for example, Patent Document 2). .

JP 2008-217209 A JP 2008-225615 A

  However, the method described in Patent Document 1 requires that the time on the storage device side and the time on the backup device side be strictly synchronized, and it is difficult to guarantee the consistency of backup data. There is. For example, due to time lag between both devices, if it is erroneously determined that one updated block is not actually updated, inconsistency occurs in the backup data, but restoration is performed without knowing that. The entire block device may not be available.

  As a countermeasure, a method such as setting the reference time to be 10 minutes before the previous backup can be considered in order to allow a time lag, but the amount of time lag depends on the storage device side. For this reason, it is difficult to obtain the optimum value, and if the deviation amount is allowed to be large, the possibility of specifying the data of the block that has not been updated as differential data increases, and the load at the time of data transfer and the backup device side The load of memory capacity increases.

  Further, in the method described in Patent Document 1, a new area for managing the last update time is required on the block storage apparatus side for the number of blocks. For this reason, there is a problem that the memory usage on the storage device side increases as the number of blocks increases.

  For example, if it is assumed that the system is operated for 10 years and the year is 365 days without considering the leap year and the leap second, 10 years becomes 315, 360,000 seconds. When it is attempted to implement the last update time with a counter value that counts in units of one second, the number of seconds does not fit in 16 bits, and a data capacity of more bits is required. Considering the calculation efficiency, the last update date and time is managed as 32-bit data. For example, when a block device of 1 TiB is divided into blocks of 32 KiB, if a 32-bit memory is required for each block, 31250 blocks × 32 bits = 1 MiB are consumed for the control memory of the storage apparatus.

  Also, in Patent Document 2, the server device connected to the block storage device acquires the access count, snapshot creation count, mount count, backup count, restore count, and the like in the block storage device, and based on the file, The fact that group migration is determined is described. However, Patent Document 2 only describes that the server that acquired these data performs migration comparison to compare the size with a specific variable, and specific logic for differential backup is disclosed. Not.

  If you want to perform differential backup without inconsistency, you need to know how these counts are managed on the block storage side and how the processing that measures these counts is performed. . However, if the differential backup must be controlled based on such different internal processing for each storage device, there is a problem that the implementation of the backup device becomes complicated and the versatility is lost.

  In view of the above-described problems, the present invention provides a data management system, storage device, backup device, and storage device that can implement differential backup of a block storage device with simple logic and with less processing load and / or less resource load. An object of the present invention is to provide a backup method and a storage apparatus backup program.

  A data management system according to the present invention includes a storage apparatus including one or more storage units that manage storage areas in units of predetermined blocks, and a backup apparatus that backs up data stored in the storage units of the storage apparatus. When the storage device detects a stationary point at a predetermined timing while the data stored in the storage unit is stationary, the storage device updates the value of the stationary point counter held by the storage device. When data is written to either the counter update unit or a block that is a plurality of block-specific data areas of the storage unit, among the last update period held by the device in association with each block A last update period update unit that updates the last update period corresponding to the block to the value of the static point counter; The backup device includes a quiesce point information providing unit that outputs a quiesce point counter and a last update period in response to a request from the backup device. Based on the stationary point information acquisition unit that acquires the last update period of the unit, the acquired last update period, and the reference point counter that indicates the reference point of backup for each of the blocks held by the own device, The block of the storage unit that is the backup target detects the updated block that has been updated since the previous backup, and the acquired quiesce counter is the current block of the block identified as the updated block And an update block detector that holds as a backup reference point counter That.

  The storage apparatus according to the present invention has one or more storage units that manage storage areas in a predetermined block unit, and a quiesce point that is a predetermined timing in a state where data stored in the storage unit is stationary in the own apparatus. When detected, when data is written to any of a block that is a plurality of data areas in block units of the storage unit and a stationary point counter update unit that updates the value of the stationary point counter held by the own device A last update period update unit that updates the last update period corresponding to the block among the last update periods held by the own apparatus in association with each block to the value of the quiescent point counter, and a request from the backup device And a quiesce point information providing unit for outputting a quiesce point counter and a last update period.

  The backup device according to the present invention provides a backup device that stores data in a storage unit of a storage device having one or more storage units that manage storage areas in units of predetermined blocks. A stationary point counter whose value is updated when a stationary point at a predetermined timing is detected while the data stored in the storage unit is stationary, and a final update period of the storage unit to be backed up A quiesce point information acquisition unit that acquires a last update period that is updated to the value of the quiesce point counter when data is written to any of the blocks that are data areas of the storage unit, and the acquired last update period And a reference point counter indicating the reference point of backup for each block held by the device. A block in which a storage unit that is a backup target has an update block that is a data update from the previous backup, and the acquired quiesce counter is identified as an update block And an update block detection unit that holds as a reference point counter for the current backup.

  According to the storage apparatus backup method of the present invention, a storage apparatus including one or more storage units that manage storage areas in a predetermined block unit is in a state where data stored in the storage unit is stationary in the own apparatus. When a quiesce point at a predetermined timing is detected, the value of the quiesce point counter held by the own device is updated, and the storage device applies to any of the blocks that are a plurality of data areas in block units of the storage unit. When the data is written, the last update period corresponding to the block is updated to the value of the quiesce point counter among the last update periods held by the own device in association with each block, and the storage device stores the data. In response to a request from the backup device that backs up the data stored in the unit, When the backup device performs backup, the backup device acquires a quiesce point counter and the last update period of the storage unit to be backed up, and the backup device acquires the last update period acquired, Based on the reference point counter indicating the backup reference point for each block held by the local device, data was updated from the previous backup from among the blocks of the storage unit that was the backup target An update block that is a block is detected, and the acquired still point counter is held as a reference point counter for the current backup of the block identified as the update block.

  A storage apparatus backup program according to the present invention provides a storage system for backing up data stored in a storage unit of a storage apparatus having one or more storage units that manage storage areas in predetermined block units. When a quiesce point at a predetermined timing is detected in a state where data stored in the storage unit is stationary in the storage device, a quiesce point counter whose value is updated and a storage unit to be backed up A quiesce point information acquisition process that acquires a last update period that is updated to the value of the quiesce point counter when data is written to any of the blocks that are data areas of the storage unit. Updated last update period and a block for each block held by the device. Based on the reference point counter indicating the reference point of backup, an update block that is a block whose data has been updated since the previous backup is detected and acquired from among the blocks of the storage unit that is the backup target An update block detection process is performed in which the static point counter is held as a reference point counter for the current backup of a block identified as an update block.

  According to the present invention, differential backup of a block storage device can be realized with simple logic and with less processing load and / or less resource load.

It is a block diagram which shows the structural example of the data management system of 1st Embodiment. 3 is a flowchart showing an example of the operation of the block storage device 1. It is explanatory drawing which shows the example of the still point counter 141 hold | maintained at the meta-information management part 14, and the last update period 142. FIG. It is explanatory drawing which shows the example after the update of the last update period. 3 is a flowchart showing an example of the operation of the block storage device 1. 3 is a flowchart showing an example of the operation of the block storage device 1. It is explanatory drawing which shows the example after the update of the stationary point counter. 4 is a flowchart illustrating an example of a backup operation of the backup device 2. 4 is a flowchart illustrating an example of a backup operation of the backup device 2. It is explanatory drawing which shows the example of the recording content of a backup file. It is a schematic block diagram which shows the structural example of the computer concerning embodiment of this invention. It is a block diagram which shows the outline | summary of the data management system of this invention.

Embodiment 1. FIG.
Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram illustrating a configuration example of a data management system according to the first embodiment. A data management system 500 illustrated in FIG. 1 includes a block storage device 1, a backup device 2, and an application device 3.

  The block storage device 1 includes one or more storage units 11, a data input unit 12, a data output unit 13, a meta information management unit 14, a meta information providing unit 15, a last update period update unit 16, a stationary A point detection unit 17, a stationary point counter update unit 18, and a snapshot storage unit 111 are included.

  As described above, the storage unit 11 is a storage unit that manages its own storage area in units of predetermined blocks. Each of the storage units 11 of this embodiment has a plurality of data areas (hereinafter simply referred to as blocks) in block units. Hereinafter, data stored in a block may be referred to as block data.

  The data input unit 12 accepts a data write request to the storage unit 11 from the outside. The data output unit 13 receives a data read request from the storage unit from the outside.

  The meta information management unit 14 manages (holds) the meta information of the block storage device 1. Examples of the meta information include a name, capacity, access authority, etc. for each of the storage units 11, but the meta information is not limited to these.

  The meta information management unit 14 of the present embodiment further holds a still point counter 141 and a last update period 142 as meta information.

  The quiesce point counter 141 is a counter whose value is updated (increased or decreased) each time a quiesce point is obtained, which is a point in time when the state of the block storage device 1 or the state of the storage unit 11 satisfies a predetermined condition. The stationary point counter 141 is provided for each storage unit 11 when a stationary point is taken independently for each storage unit 11. For example, the meta information management unit 14 may hold the still point counter 141 in association with the identifier of the storage unit 11.

  In many block storage device backups, there is a characteristic of taking a quiesce point to ensure consistency. Also in the present embodiment, such a characteristic that takes a stationary point can be used. Note that the quiesce point of the present embodiment is not limited to a point taken to ensure consistency in backup of the block storage device.

  In the present embodiment, for example, a quiesce point is defined as a point in time when the state of the block storage device 1 or the state of the storage unit 11 satisfies a predetermined condition. The predetermined condition includes a stationary condition that data is stationary at least in the target (block storage device 1, storage unit 11, group thereof, or the like). In other words, the stationary point may be a point in time when the data is stationary in a predetermined target and a state that satisfies a predetermined condition is satisfied. In addition, a quiesce point can be defined as a point in time when a predetermined request (for example, a quiesce point securing request) from the backup device 2 is received. However, in the present embodiment, the stationary points are generated at a time interval longer than at least 1 second.

  In the present embodiment, the time point when the block storage device 1 creates a snapshot will be described as an example of a quiesce point. However, the quiesce point may be an arbitrary time point that satisfies the above conditions. Other examples of the quiesce point include, for example, a time point when the application device 3 is notified of the shutdown and a time point when the file system is unmounted from the application device 3. Note that the rest point may be in any of these states. That is, the predetermined condition can allow a plurality of different states.

  Further, the last update period 142 is provided for each block of each storage unit 11. For example, the meta information management unit 14 may hold the last update period 142 in association with the identifier of the storage unit 11 and the identifier of the block.

  Here, the last update period 142 is a period in which data was last updated in the corresponding block. More specifically, the last update period 142 is information of a predetermined time interval including the time when the data was last updated in the corresponding block.

  In the present embodiment, such a last update period 142 is expressed using a stationary point counter 141. More specifically, in the last update period 142, it is determined which section of the time section sandwiched between two stationary points that are temporally adjacent to each other when the data is last updated in the corresponding block. Information to be shown. The last update period 142 may be, for example, information (still point counter) indicating a still point corresponding to the start point of a period (the above-described time interval) in which data is last updated in the corresponding block.

  The meta information providing unit 15 provides the meta information held in the meta information management unit 14 to the outside.

  The last update period update unit 16 updates the corresponding last update period 142 in response to a data write request to the data input unit 12. For example, in response to a data write request to the data input unit 12, the final update period update unit 16 sets the final update period 142 associated with the requested block to the value of the quiesce point counter 141 corresponding to the block. Update. Here, the stationary point counter 141 corresponding to the block is, for example, the stationary point counter 141 associated with the storage unit 11 having the block. If there is only one stationary point counter 141, the only stationary point counter 141 may be the stationary point counter 141 corresponding to the block. In addition, when the stationary point counter 141 is provided for each group of the predetermined storage unit 11, the stationary point counter 141 associated with the group to which the storage unit 11 having the block belongs corresponds to the stationary point corresponding to the block. A point counter 141 may be used. In the present embodiment, it is assumed that the relationship between the unit in which the stationary point counter 141 is provided (that is, the stationary point detection unit) and the block corresponding to the unit is known in the last update period update unit 16.

  The stationary point detector 17 detects that a stationary point has been taken.

  The stationary point counter update unit 18 updates the stationary point counter 141 based on the detection result by the stationary point detection unit 17. Each time the stationary point counter updating unit 18 detects that a stationary point has been taken by the stationary point detecting unit 17, the stationary point counter updating unit 18 updates the corresponding stationary point counter 141. The stationary point counter updating unit 18 updates the stationary point counter 141 associated with the detection source storage unit 11 every time the stationary point detecting unit 17 detects that a stationary point has been taken, for example. In the quiescent point counter updating unit 18 as well, it is assumed that the relationship between the detected quiesce point and the updated quiesce point counter 141 is already known.

  The snapshot storage unit 111 stores a snapshot.

  The backup device 2 includes a backup unit 21, a restore unit 22, a backup data storage unit 23, and a quiesce point counter storage unit 24.

  The backup unit 21 backs up data (more specifically, block data of each block included in the storage unit 11) stored in each storage unit 11 or the designated storage unit 11 of the block storage device 1.

  The restore unit 22 uses each storage unit 11 of the block storage device 1 or a designated storage unit 11 based on a backup file, which will be described later, or information stored in the backup data storage unit 23, the quiesce point storage unit 24, or the like. Restore data to.

  The backup data storage unit 23 stores backup data of each storage unit 11 included in the block storage device 1 that is a backup target.

  The backup unit 21 includes a stationary point securing unit 211 and an update block detection unit 212.

  The stationary point securing unit 211 secures a stationary point to be backed up at the time of backup. The quiesce point securing unit 211 issues a predetermined event accompanied by the occurrence of a quiesce point to the block storage device 1 in a state where writing of data to at least the storage unit 11 to be backed up in the block storage device 1 is stopped. Thus, a stationary point may be secured. For example, when the quiesce point is the time when the snapshot is updated, the quiesce point securing unit 211 creates a snapshot in the backup target and generates a quiesce point. In addition, for example, if the stationary point is a reception point of a predetermined request (for example, a stationary point securing request) from the backup device 2, the stationary point securing unit 211 transmits the request specifying the backup target, Generate a quiesce point.

  When performing a new backup, the update block detection unit 212 detects a block whose data has been updated from the previous backup time to be backed up.

  The quiesce point counter storage unit 24 holds a quiesce point counter 241 serving as a backup reference point in association with each minimum unit (for example, the storage unit 11) to be backed up. Note that the stationary point counter storage unit 24 may hold the stationary point counter 241 associated with each of the acquired difference data 231 for each backup, with the above as the stationary point counter 241 in the initial state. The stationary point counter 241 may be referred to as a “reference point counter” in the sense of indicating a backup reference point.

  Here, the quiesce point counter 241 held by the backup device 2 is set with the quiesce point counter 141 corresponding to the target acquired from the block storage device 1 at the backup time immediately before the target.

  The application device 3 is a device that makes a data write request or a data read request to the block storage device 1 and inputs / outputs data to / from the block storage device 1.

  In this embodiment, the block storage apparatus 1 may be a SAN storage apparatus provided with an interface such as Fiber Channel, iSCSI, or SAS (Serial Attached SCSI). The block storage device 1 may be a general-purpose server into which software that communicates using iSCSI or a unique protocol is introduced, for example. Further, the block storage device 1 may be, for example, a flash storage that communicates with NVM Express or a unique protocol.

  In addition, the backup device 2 is a general-purpose server that includes, for example, software in which the backup unit 21 and the restore unit 22 are installed, and includes storage devices corresponding to the backup data storage unit 23 and the quiescent point counter storage unit 24 It may be. The backup device 2 may be a physical server or a virtual server. Further, the backup data storage unit 23 may be realized by an external storage connected to the backup device 2.

  The application device 3 may be an arbitrary server, a personal computer, or an information processing device such as a CPU provided in them. The application apparatus 3 may be a general-purpose server that creates a file system in the storage unit 11 of the block storage apparatus 1 and stores application programs or data files in the file system. The application device 3 may be a physical server or a virtual server. The application apparatus 3 may be a sensor device such as a camera that records a still image or a moving image.

  Next, the operation of this embodiment will be described. First, the operation on the block storage device 1 side will be briefly described.

  The block storage device 1 of this embodiment divides each storage unit 11 into logical blocks of a certain size. For example, the size of each block may be 32 KiB. If the block is made smaller, the data amount reduction effect by differential backup is enhanced in the backup device 2, but the memory usage is increased.

  When the storage unit 11 is newly created, the block storage device 1 adds a quiesce point counter 141 and a last update period 142 corresponding to the new storage unit 11 to the meta information management unit 14, and sets these values as initial values. Turn into. In this embodiment, the initial value of the still point counter 141 is 1 and the initial value of the last update period 142 of each block is 0.

  FIG. 2 is a flowchart showing an example of the operation of the block storage device 1 of the present embodiment. FIG. 2 is a flowchart showing an example of the operation when the block storage apparatus 1 accepts a data write request. In the example illustrated in FIG. 2, first, the data input unit 12 receives a data write request from the application device 3 (step S <b> 101).

  Next, the data input unit 12 writes data in the block in response to the request (step S102). Note that the block to which data is written is not particularly limited.

  When the data writing by the data input unit 12 is successful (Yes in Step S103), the final update period update unit 16 updates the final update period 142 of the corresponding block (Step S104). If the writing has failed (No in step S103), the process proceeds to step S105 without performing step S104.

  Finally, the data input unit 12 returns the write result to the request source (step S105).

  FIG. 3 is an explanatory diagram illustrating an example of the stationary point counter 141 and the last update period 142 held in the meta information management unit 14. FIG. 3A is an explanatory diagram illustrating an example of the stationary point counter 141, and FIG. 3B is an explanatory diagram illustrating an example of the final update period 142.

  In the block storage device 1, when there is a write request from the application device 3 to the data input unit 12 and the write is successful, the final update period update unit 16 updates the final update period 142 of the corresponding block.

  FIG. 4 is an explanatory diagram showing an example after the last update period 142 is updated. For example, when the data in the block identified by block number 3 is updated by writing in the initial state shown in FIG. 3, the last update period update unit 16 corresponds to the block as shown in FIG. The last update period 142 to be set is set to 1, which is the value of the stationary point counter 141.

  FIG. 5 is a flowchart showing an example of the operation of the block storage device 1 when a stationary point is detected. In the example shown in FIG. 5, first, the stationary point detection unit 17 detects a stationary point (step S111).

  Next, the stationary point counter updating unit 18 updates the stationary point counter 141 corresponding to the detected stationary point (step S112).

  FIG. 6 is a flowchart showing an example of the operation when the block storage apparatus 1 accepts a snapshot request. In the example shown in FIG. 5, first, the block storage apparatus 1 receives a snapshot request (step S121).

  Next, a snapshot function unit (not shown) of the block storage device 1 creates a snapshot of each storage unit 11 in response to the received snapshot request (step S122). Thereby, the block data of each storage unit 11 is in a state where a stationary point is taken.

  In the block storage device 1, after step S <b> 122 in which a stationary point is taken, the operations of steps S <b> 111 to S <b> 112 are performed.

  If access to the block storage device 1 is frozen before taking a snapshot in step S122, unwritten data is written in the storage unit 11 before taking the snapshot. That is, the operations in steps S101 to S105 are performed.

  For example, when the application device 3 has created a file system in the storage unit 11 of the block storage device 1, when the file system is frozen, a new write request to the file system is suspended, and the storage unit 11 Unwritten data is written, and the data in the storage unit 11 is in a state where a stationary point is taken. When the freeze of the file system is canceled, the write request is suspended, and the data in the storage unit 11 is in a state where no quiesce point has been taken. Many file systems have a file system freeze function.

  FIG. 7 is an explanatory diagram showing an example after the stationary point counter 141 is updated. With the creation of the above snapshot, the stationary point counter 141 is updated to the state shown in FIG. 7, for example. In this example, there is shown an example in which there is no unwritten data in the storage unit 11, but when there is unwritten data, the last update corresponding to the block in which the data is written. The period 142 is set to 1, which is the value of the quiesce point counter 141 at the time before taking the snapshot.

  Next, the operation of the backup device 2 will be described. 8 to 9 are flowcharts illustrating an example of the backup operation of the backup device 2.

  In the example shown in FIGS. 8 to 9, first, the quiesce point securing unit 211 obtains (secures) the quiesce point of the storage unit 11 to be backed up. A method of taking a stationary point is, for example, shutting down the application apparatus 3. For example, the file system is unmounted from the application apparatus 3. Further, for example, the file system of the application apparatus 3 is frozen, the snapshot of the storage unit 11 is created in the block storage apparatus 1, and the freeze of the file system of the application apparatus 3 is released.

  In this example, an example using the third method will be described. That is, a method of creating a snapshot by temporarily freezing the file system that does not require the application device 3 to be shut down and the file system unmounted will be described. Although it is possible to perform backup without freezing the file system, it should be noted that in this case, unwritten data is not reflected and some files may not be restored.

  First, the stationary point securing unit 211 freezes data access at the application level to the block storage device 1 (step S201).

  For example, the stationary point securing unit 211 may log in to the application apparatus 3 using Secure Shell and execute a command for file system freeze.

  Next, the stationary point securing unit 211 causes the block storage device 1 to create a snapshot (step S202). For example, the quiesce point securing unit 211 may log in to the management interface of the block storage device 1 using Secure Shell and execute a command for creating a snapshot. Further, for example, the stationary point securing unit 211 may execute a remote procedure call for creating a snapshot of the block storage device 1.

  Next, the stationary point securing unit 211 cancels the frozen state of the data access frozen in step S201 (step S203). Since the writing to the file system is suspended while the file system is frozen, it is desirable to release the freeze as soon as possible.

  If the application device 3 does not create a file system in the storage unit 11 of the block storage device 1, the above operation may be performed without going through the file system.

  Next, the update block detection unit 212 makes a request to the meta information providing unit 15 and receives the quiesce point counter 141 and the last update period 142 (step S204). For example, the update block detection unit 212 may issue a SCSI command for requesting the block storage apparatus 1 to read meta information, and receive data including the meta information from the block storage apparatus 1 as a response. Further, for example, the update block detection unit 212 may obtain the information by logging in to the management interface of the block storage device 1 using Secure Shell and executing a command for outputting meta information. Further, for example, the update block detection unit 212 may obtain the information by executing a remote procedure call for obtaining meta information of the block storage device 1.

  Next, the update block detection unit 212 stores the still point counter 141 acquired in step S204 for the current backup (step S205). The update block detection unit 212 may store the acquired still point counter 141 as the still point counter 241 in the still point counter storage unit 24, for example.

  Note that the backup destination of the stationary point counter 141 may not be a hard disk drive built in the backup device 2, for example. The backup destination may be an external storage device connected to the backup device 2, for example.

  In this example, an example is shown in which one backup file is created for one backup in the still point counter storage unit 24 as a file system created in a hard disk drive built in the backup device 2. At this time, the update block detection unit 212 may determine the file name of the backup file from the identification ID of the storage unit 11 and the quiesce point counter 141. For example, when the identification ID of the storage unit 11 is 001114FB and the quiesce point counter 141 is 1, the file name of the backup file is, for example, 001114FB-00001. Let it be BAK. Note that the update block detection unit 212 creates an empty file in this step.

  Next, the update block detection unit 212 confirms whether or not information related to the previous backup (for example, the backup file described above) exists (step S206). If not, the process proceeds to step S207. If it exists, the process proceeds to step S208. move on.

  In step S207, the update block detection unit 212 substitutes 0 for a variable b indicating a backup reference point for obtaining an update block. Since there is no information about the previous backup, this is the first backup. In this example, in the first backup, only the blocks that have been written are backed up, and all zero blocks that are not written and are in the initial state of the block are not backed up. By not backing up all zero blocks, read processing from the block storage device 1 can be omitted, and the capacity for storing data can be saved.

  If there is information related to the previous backup, the update block detection unit 212 substitutes the quiesce point counter 241 at the previous backup into the variable b (step S208).

  Next, the processing of steps S210 to S212 is repeated for each target block of the snapshot stored in the snapshot storage unit 111 (loop L209).

  That is, the update block detection unit 212 first substitutes the variable a for the last update period of the target block acquired in step S204 (step S210).

  Next, the values of the variable a and the variable b are compared (step S211). If a is large, the process proceeds to step S212. If b is large or equal, the loop L209 is continued without executing step S212.

  On the other hand, if a is large, the target block is assumed to be an updated block whose data has been updated since the previous backup, and the block data of the target block is backed up from the snapshot of the block storage device 1 (step S212). In this example, the update block block data is written as difference data in the backup file.

  FIG. 9 is an explanatory diagram showing an example of the recorded contents of a backup file. As an example of the contents to be additionally recorded (recorded) in the backup file, as shown in FIG. 10, there is a set of a block number and a block data set as an update block.

  The update block detection unit 212 repeats the loop L209, and when a is larger than b, the update block detection unit 212 adds a combination of the block number and the block data to the backup file.

  When the loop L209 ends, the quiesce point securing unit 211 causes the block storage device 1 to delete the snapshot of the storage unit 11 to be backed up (step S213). When the block storage apparatus 1 adopts the copy-on-write snapshot method, it is expected that the write performance of the storage unit 11 is improved by deleting the snapshot created at the start of this backup. The However, deleting a snapshot is not essential for implementing the present invention. For example, in order to restore the snapshot data within the block storage device 1 in a short time, the snapshot may be left without being deleted.

  In order to restore the data backed up as described above, the restore unit 22 may read out the old backup file sequentially and write it into the storage unit 11. Since the restore operation for restoring based on the differential backup data is widely known, detailed description thereof is omitted.

  As described above, the backup device 2 according to the present embodiment can specify a block updated by a simple algorithm that compares the quiesce point counter at the time of the previous backup and the last update period of each newly acquired block. For this reason, the backup device 2 does not need to worry about other internal data and the snapshot / replication function of the block storage device 1, and can implement differential backup in a short period of time and at low cost.

  Further, according to the present embodiment, even when the backup device 2 performs the initial backup, it is not necessary to read out all data and compare it with the backup. When backing up from a snapshot, the snapshot can be deleted after the backup is completed. Therefore, the performance of the block storage device 1 can be improved, the increase in data capacity of the backup device 2 can be prevented, and the transfer load between the block storage device 1 and the backup device 2 can be reduced.

  Furthermore, time synchronization is not required and the memory usage on the block storage device 1 side can be reduced as compared with the case of storing the last update date and time.

  That is, when the block storage device 1 holds the last update date and time, it is considered that 32 bits per block are necessary when calculated from the general backup frequency and product life, whereas the last update period 142 of the present invention is used. Is about 16 bits per block.

  In addition, users of block storage devices can back up in a short time and can obtain a backup device with a low load during backup at a low cost. You can build a system that is less likely to lose data than taking

  The number of bits in the last update period 142 was calculated under the following assumption. That is, in general, many backups are taken outside business hours, and nights or weekends are set every day. In addition to this, even if backup is taken for irregular maintenance, it is considered sufficient to make backups at a frequency of several times / day. Also, considering the general maintenance period of storage products, it is considered sufficient if the numbers are not overflowing for 20 years. According to the assumed case, it is only necessary to support 20 years × 365 days × 3 times = 219,000 backups, and 16 bits (up to 65,535) are sufficient in consideration of processing efficiency. it is conceivable that.

  In the modification described later, 32 bits per block or 64 bits may be assumed in some cases, but it is more complicated than the method of comparing the last update time or the method of referring to the number of snapshots directly. There is no need for simple control and time synchronization, and there can still be obtained an effect that it can be realized only by an algorithm based on a simple comparison and that data consistency can be guaranteed. That is, it is possible to reduce the processing load (the amount of data read is small) and / or the resource load (a small amount of memory used) with simple logic.

Modified example.
Note that at least a part of the configuration and method of the above-described embodiment can also be applied to incremental backup. When performing incremental backup, an updated block may be detected by using a backup quiesce point counter 241 serving as a base point of incremental backup instead of the quiesce point counter 241 at the time of the previous backup.

  Further, when performing differential backup only from a snapshot, the snapshot identification ID may be substituted for the quiesce point counter 141. For example, when the snapshot identification ID increases in order of 1, 2, 3,... Each time a snapshot is taken, the identification ID can be used as the stationary point counter 141. In the present invention, such a snapshot identification ID is also regarded as one aspect of the stationary point counter 141.

  The stationary point counter 141 may be a counter that increases (or decreases) at regular time intervals. For example, a timer value that increases every 10 minutes may be used as the stationary point counter 141. In this case, it is preferable that the process of updating the stationary point counter 141 is performed in a state where data rewriting does not occur. In this case, the backup device 2 also backs up a block equal to the quiesce point counter 241 in addition to a block larger than the quiesce point counter 241 at the time of the previous backup. In the present invention, such a counter is also regarded as one aspect of the stationary point counter 141.

  Further, the quiescent point counter 141 may be a counter that increases (or decreases) for each fixed amount of writing. For example, a counter value that increases by 1 every time 32 MiB is written may be used as the stationary point counter 141. In this case, the backup device 2 also backs up a block equal to the quiesce point counter 241 in addition to the block larger than the quiesce point counter 241 at the time of the previous backup. In the present invention, such a counter is also regarded as one aspect of the stationary point counter 141.

  The stationary point counter 141 may be a counter that increases (or decreases) for each writing. Note that this example requires a longer number of bits than using a counter that increases (or decreases) only when a static point is taken, but the effect of realizing differential backup with a simple algorithm is obtained. It is done. In the present invention, such a counter is also regarded as one aspect of the stationary point counter 141.

  Next, a configuration example of a computer according to the embodiment of the present invention will be shown. FIG. 11 is a schematic block diagram illustrating a configuration example of a computer according to the embodiment of the present invention. The computer 1000 includes a CPU 1001, a main storage device 1002, an auxiliary storage device 1003, an interface 1004, a display device 1005, and an input device 1006.

  The block storage device, the backup device, and the application device described above may be mounted on the computer 1000, for example. In that case, the operation of each device may be stored in the auxiliary storage device 1003 in the form of a program. The CPU 1001 reads out the program from the auxiliary storage device 1003 and develops it in the main storage device 1002, and executes the predetermined processing in the above embodiment according to the program.

  The auxiliary storage device 1003 is an example of a tangible medium that is not temporary. Other examples of the tangible medium that is not temporary include a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, and a semiconductor memory connected via the interface 1004. When this program is distributed to the computer 1000 via a communication line, the computer that has received the distribution may develop the program in the main storage device 1002 and execute the predetermined processing in the above embodiment.

  The program may be for realizing a part of predetermined processing in each embodiment. Furthermore, the program may be a difference program that realizes the predetermined processing in the above-described embodiment in combination with another program already stored in the auxiliary storage device 1003.

  The interface 1004 transmits / receives information to / from other devices. The display device 1005 presents information to the user. The input device 1006 accepts input of information from the user.

  Further, depending on the processing contents in the embodiment, some elements of the computer 1000 may be omitted. For example, if the device does not present information to the user, the display device 1005 can be omitted.

  In addition, some or all of the constituent elements of each device are implemented by general-purpose or dedicated circuits (Circuitry), processors, and the like or combinations thereof. These may be constituted by a single chip or may be constituted by a plurality of chips connected via a bus. Moreover, a part or all of each component of each device may be realized by a combination of the above-described circuit and the like and a program.

  When some or all of the constituent elements of each device are realized by a plurality of information processing devices and circuits, the plurality of information processing devices and circuits may be centrally arranged or distributedly arranged. Also good. For example, the information processing apparatus, the circuit, and the like may be realized as a form in which each is connected via a communication network, such as a client and server system and a cloud computing system.

  Next, the outline of the data management system of the present invention will be described. FIG. 12 is a block diagram showing an outline of the data management system of the present invention. A data management system 600 illustrated in FIG. 12 includes a storage device 61 and a backup device 62.

  The storage device 61 (for example, the block storage device 1) includes one or more storage units 611 that manage storage areas in predetermined block units, a quiesce point counter update unit 612, a last update period update unit 613, a quiesce point An information providing unit 614.

  When the stationary point counter updating unit 612 (for example, the stationary point counter updating unit 18) detects a stationary point at a predetermined timing in a state where data stored in the storage unit is stationary in the own device, the own device holds the stationary point counter. The value of the stationary point counter 615 being updated is updated.

  The last update period update unit 613 (for example, the last update period update unit 16) associates each block when data is written in any of the blocks that are a plurality of data areas in block units of the storage unit. The last update period 616 corresponding to the block in the last update period 616 held by the own apparatus is updated to the value of the stationary point counter 615.

  The quiesce point information providing unit 614 (for example, the meta information providing unit 15) outputs a quiesce point counter 615 and a last update period 616 in response to a request from the backup device.

  Further, the backup device 62 includes a still point information acquisition unit 621 and an update block detection unit 622.

  The quiesce point information acquisition unit 621 (a part of the update block detection unit 212) acquires the quiesce point counter 615 and the last update period 616 of the storage unit to be backed up when performing backup.

  Based on the last update period 616 acquired by the update block detection unit 622 (part of the update block detection unit 212) and the reference point counter 623 indicating the reference point of backup for each block held by the own device. A block in which an update block that is a block whose data has been updated since the previous backup is detected from among blocks of the storage unit that is the backup target, and the acquired quiesce point counter 615 is identified as an update block Is stored as the reference point counter 623 for this backup.

  While the present invention has been described with reference to the present embodiment and examples, the present invention is not limited to the above embodiment and examples. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

  The present invention can be suitably applied to backup applications for storage units that store data on a block basis.

500 Data Management System 1 Block Storage Device 11 Storage Unit 111 Snapshot Storage Unit 12 Data Input Unit 13 Data Output Unit 14 Meta Information Management Unit 141 Quiet Point Counter 142 Last Update Period 15 Meta Information Providing Unit 16 Last Update Period Update Unit 17 Stationary Point detection unit 18 Quiet point counter update unit 2 Backup device 21 Backup unit 211 Quiet point securing unit 212 Update block detection unit 22 Restore unit 23 Backup data storage unit 231 Difference data 24 Quiet point counter storage unit 241 Quiet point counter 3 Application device 1000 Computer 1001 CPU
1002 Main storage device 1003 Auxiliary storage device 1004 Interface 1005 Display device 1006 Input device 61 Storage device 611 Storage unit 612 Quiet point counter update unit 613 Last update period update unit 614 Quiesce point information provision unit 615 Still point counter 616 Final update period 62 Backup Device 621 Still point information acquisition unit 622 Update block detection unit 623 Reference point counter

Claims (10)

A storage device including one or more storage units that manage storage areas in predetermined block units, and a backup device that backs up data stored in the storage units of the storage device,
The storage device
A stationary point counter updating unit that updates a value of a stationary point counter held by the own device when detecting a stationary point at a predetermined timing in a state where the data stored in the storage unit is stationary in the own device; ,
When data is written to any of the blocks that are a plurality of data areas of the block unit of the storage unit, the block in the last update period held by the own device in association with each of the blocks A last update period update unit that updates the last update period corresponding to the value of the quiescent point counter;
In response to a request from the backup device, including a stationary point information providing unit that outputs the stationary point counter and the last update period,
The backup device is
When performing the backup, the quiesce point counter, and a quiesce point information acquisition unit that acquires the last update period of the storage unit that is the backup target;
Based on the acquired last update period and the reference point counter indicating the reference point of backup for each of the blocks held by the own device, the previous block among the blocks included in the storage unit to be backed up An update block detection that detects an update block that has been updated from the time of the backup and holds the acquired still point counter as a reference point counter for the current backup of the block identified as the update block A data management system characterized by comprising a department. The data management system according to claim 1, wherein the update block detection unit detects an update block by comparing the acquired last update period with the reference point counter. The backup device is
2. A quiesce point securing unit that issues a predetermined event accompanying the occurrence of the quiesce point to the storage device in a state where data writing to at least a storage unit that is a backup target of the storage device is stopped. Or the data management system of Claim 2. The quiesce point is either the time when the snapshot is updated, the timing when a certain time elapses, the timing when a certain amount of writing is performed in the target recording unit, or the timing when one writing is performed in the target recording unit. The data management system according to any one of claims 1 to 3. The data management system according to any one of claims 1 to 4, wherein a bit length of a stationary counter held by the storage device is 16 bits or less. In each of the recording units, each of the blocks is initialized with a predetermined value,
The stationary point counter is held for each recording unit,
The data management system according to any one of claims 1 to 5, wherein the update block detection unit excludes, from the backup target, a block in which the acquired last update period remains the initial value during the initial backup. One or more storage units that manage storage areas in predetermined block units;
A stationary point counter updating unit that updates a value of a stationary point counter held by the own device when detecting a stationary point at a predetermined timing in a state where the data stored in the storage unit is stationary in the own device; ,
When data is written to any of the blocks that are a plurality of data areas of the block unit of the storage unit, the block in the last update period held by the own device in association with each of the blocks A last update period update unit that updates the last update period corresponding to the value of the quiescent point counter;
A storage apparatus, comprising: a quiesce point information providing unit that outputs the quiesce point counter and the last update period in response to a request from the backup device. When performing backup of data stored in a storage unit of a storage apparatus having one or more storage units that manage storage areas in predetermined block units, the storage apparatus stores the data in the storage unit. A stationary point counter whose value is updated when a stationary point at a predetermined timing is detected in a state where the data to be stationary is stationary, and a final update period of the storage unit to be backed up, the storage unit A quiesce point information acquisition unit that acquires a final update period that is updated to the value of the quiesce point counter when data is written to any of the blocks that are data areas of
Based on the acquired last update period and the reference point counter indicating the reference point of backup for each of the blocks held by the own device, the previous block among the blocks included in the storage unit to be backed up An update block detection that detects an update block that has been updated from the time of the backup and holds the acquired still point counter as a reference point counter for the current backup of the block identified as the update block And a backup device. A storage device having one or more storage units that manage storage areas in a predetermined block unit detects a stationary point at a predetermined timing in a state where data stored in the storage unit is stationary in the own device. Then, the value of the static point counter held by the own device is updated,
When the storage device writes data to any of the blocks that are a plurality of data areas of the block unit of the storage unit, the storage device stores the last update held in association with each of the blocks Update the last update period corresponding to the block in the period to the value of the quiesce point counter,
In response to a request from a backup device that backs up data stored in the storage unit, the storage device outputs the quiesce point counter and the last update period,
When the backup device performs the backup, it acquires the quiesce point counter and the last update period of the storage unit to be backed up,
The backup device has a storage unit that is a backup target based on the acquired last update period and a reference point counter indicating a backup reference point for each of the blocks held by the own device. An update block that is a block whose data has been updated since the previous backup is detected from among the blocks, and the acquired still point counter is used as a reference point counter for the current backup of the block identified as the update block A storage apparatus backup method, characterized by comprising: On the computer,
When performing backup of data stored in a storage unit of a storage apparatus having one or more storage units that manage storage areas in predetermined block units, the storage apparatus stores the data in the storage unit. A stationary point counter whose value is updated when a stationary point at a predetermined timing is detected in a state where the data to be stationary is stationary, and a final update period of the storage unit to be backed up, the storage unit A static point information acquisition process for acquiring a final update period that is updated to the value of the static point counter when data is written to any of the blocks that are the data area of the data area, and the acquired final update period, To a reference point counter indicating a backup reference point for each of the blocks held by the own device Based on this, the update block that is a block in which data has been updated since the previous backup is detected from among the blocks of the storage unit that is the backup target, and the acquired quiescent point counter is identified as the update block A storage device backup program for executing an update block detection process to be held as a reference point counter for the current backup of a block that has been processed.

Images