JP6394059B2 - Storage system, storage method, and program - Google Patents

Description

  The present invention relates to a storage system, a storage method, and a program.

  In a storage system using a file server such as NAS (Network Attached Storage), in order to acquire file system backup data, it is necessary to secure a static point of the file system by temporarily suspending access. For example, when the storage system is a system that operates for 24 hours, it is required to minimize the access stop time for obtaining the backup data of the file system.

  As a technique for acquiring backup data of a file system in such a short time, Patent Literature 1 and Patent Literature 2 disclose methods using snapshots. Snapshot is a technique for saving an image of a file system while ensuring consistency of the file system at a certain point in time. In a snapshot, an image before processing is saved for a file being written (a file for which writing processing has not been completed). Snapshots can be acquired in a very short time. For this reason, by using a snapshot, backup data of the file system can be acquired without making the storage system user recognize the service stop.

JP 2011-204008 A JP 2006-99406 A

  A normal file system is constructed by closing a single file server. Therefore, a snapshot can be created at an arbitrary time, and the file system at that time can be restored using the created snapshot.

  On the other hand, in a distributed file system used in a processing system such as big data, a file system is constructed on a plurality of file servers. In such a distributed file system, for example, file servers are installed in a plurality of regions, and the control device and each file server are connected by an unstable network (a network in which simultaneity between devices is not guaranteed). In this case, in order to acquire a snapshot of the entire file system, for example, it is necessary to stop the service until it is confirmed that the snapshot has been acquired in all the file servers. In addition, when some file servers become unresponsive due to a failure or the like, the service stop time becomes longer.

  As described above, the distributed file system has a problem that it is difficult to obtain backup data of the entire file system without stopping the service.

  An object of the present invention is to provide a storage system, a storage method, and a program that can solve the above-described problems and can acquire backup data of the entire file system in a distributed file system without stopping the service. .

  The storage system of the present invention stores operation information indicating operation related to a file executed on a plurality of storage devices, and states of storage areas for storing the files in the plurality of storage devices However, after restoration based on snapshots obtained by the plurality of storage devices before the restoration time, acquisition of the snapshots is instructed to the plurality of storage devices included in the operation information. And an operation re-execution unit that re-executes the operation related to the file from the time until the restoration time to the plurality of storage devices.

  The storage method of the present invention stores operation information indicating operations related to files executed on a plurality of storage apparatuses, and the state of the storage area storing the files in the plurality of storage apparatuses is before the restoration time. After the restoration based on the snapshots acquired by the plurality of storage apparatuses, the recovery time is included in the operation information from the time when the plurality of storage apparatuses are instructed to acquire the snapshots. The operations related to the file up to are re-executed for the plurality of storage apparatuses.

  The program of the present invention stores, in a computer, operation information indicating an operation related to a file executed on a plurality of storage devices, and a state of a storage area for storing the file in the plurality of storage devices is a restoration time. After being restored based on snapshots previously acquired in the plurality of storage devices, the operation information includes the operation information from the time when the plurality of storage devices are instructed to acquire the snapshots. A process is executed to re-execute the operation related to the file up to the restoration time to the plurality of storage apparatuses.

  The effect of the present invention is that backup data of the entire file system in the distributed file system can be acquired without stopping the service.

It is a block diagram which shows the characteristic structure of embodiment of this invention. It is a block diagram which shows the structure of embodiment of this invention. It is a figure which shows the acquisition timing of the snapshot in embodiment of this invention. It is a figure which shows the pattern of the operation log which concerns on the snapshot acquisition period, and the file state at the time of restoration in an embodiment of the invention. FIG. 5 is a diagram illustrating an operation log pattern and a file state at the time of restoration (continued in FIG. 4) according to a snapshot acquisition period in the embodiment of the present invention. It is a figure which shows the example of the operation information 321 in embodiment of this invention. It is a figure which shows the other example of the operation information 321 in embodiment of this invention. It is a figure which shows the example of the control information 231 in embodiment of this invention. It is a flowchart which shows the backup data generation process in embodiment of this invention. It is a flowchart which shows the file system restoration process in embodiment of this invention. It is a figure which shows the example of the state of a snapshot and a storage area in embodiment of this invention.

  First, an outline of the configuration and operation of the embodiment of the present invention will be described.

  FIG. 2 is a block diagram showing the configuration of the embodiment of the present invention. Referring to FIG. 2, the storage system 100 according to the embodiment of the present invention includes a control device 200, one or more proxy servers (or proxy devices) 300 (300a, b,...), And a plurality of NAS (or Storage device) 400 (400a, b,...). The storage system 100 is connected to one or more client devices 500 (500a, b,...) Via a network such as the Internet. The control device 200, the proxy server 300, and the NAS 400 in the storage system 100 are also connected (distributed and arranged on the network) by a network such as the Internet or a LAN (Local Area Network).

  The client device 500 writes (creates, updates, deletes) and reads files from the storage system 100.

  The control device 200 of the storage system 100 controls acquisition of backup data (operation log and snapshot) of the file system and restoration of the file system.

  The proxy server 300 writes and reads a file to and from the NAS 400 in response to a request from the client device 500.

  A file system (distributed file system) is constructed on a plurality of NAS 400. Each of the plurality of NAS 400 stores a file.

  Here, file writing and reading processing in the embodiment of the present invention will be described.

  In the embodiment of the present invention, it is assumed that the file operation from the client apparatus 500 to the storage system 100 is performed by an HTTP (Hypertext Transfer Protocol) message as in a general big data processing system. In this case, the file operation from the client device 500 is performed by the put / get / delete method.

  File operations received from the client device 500 using the put / get / delete method are allocated to any of the plurality of proxy servers 300 in the storage system 100 according to a predetermined rule. Here, it is assumed that operations for the same file are allocated to the same proxy server 300.

  The proxy server 300 simultaneously performs a file operation according to put / get / delete received from the client device 500 on a plurality of NAS 400 selected according to a predetermined rule.

  Here, when the proxy server 300 receives the put / delete from the client device 500, the proxy server 300 requests the NAS 400 to write a file. For example, the proxy server 300 requests the NAS 400 to generate a file in the file system for the new file put, update the file for the already generated file put, and delete the file for the delete. For example, when the proxy server 300 receives a response from a predetermined number of the NAS 400 that transmitted the request, the proxy server 300 transmits a response to the put / delete to the client device 500.

  Further, when the proxy server 300 receives a get from the client device 500, the proxy server 300 requests the NAS 400 to read the file. For example, the proxy server 300 transmits a file from the NAS 400 that responded first among the NAS 400 that transmitted the request to the client device 500 as a response to the get.

  In response to a request from the proxy server 300, the NAS 400 creates, updates, and deletes a file in the storage area on the storage medium, and reads the file from the storage area.

  As described above, the file operation for the NAS 400 is performed based on the HTTP message from the client device 500. Therefore, the file update in the NAS 400 is a file unit update. Also, there are no dependencies between files. Furthermore, as described above, operations for the same file are processed by the same proxy server 300. For this reason, in the embodiment of the present invention, the NAS 400 does not pass the operation for the same file (the subsequent operation from the client device 500 is executed before the previous operation).

  Next, generation of backup data and restoration of a file system in the embodiment of the present invention will be described.

  In the embodiment of the present invention, in the backup data generation process, each proxy server 300 acquires a log (operation log) of a file operation performed on the NAS 400 in response to an instruction from the control device 200. Further, each NAS 400 acquires a snapshot of the state of the storage area in response to an instruction from the control device 200.

  Further, in the file system restoration process, each NAS 400 restores the state of the storage area based on the snapshot in accordance with an instruction from the control device 200. Further, each proxy server 300 re-executes the operation on the NAS 400 stored in the operation log in accordance with an instruction from the control device 200.

  FIG. 3 is a diagram showing snapshot acquisition timing in the embodiment of the present invention.

  In FIG. 3, the time Ta indicates the time when the control device 200 confirms the start of operation log acquisition in all proxy servers 300. The time Tb indicates the time when the control apparatus 200 instructs all NAS 400 to acquire the snapshot (snapshot A) of the snapshot ID (Identifier) “A”. The time Tc indicates the time when the control device 200 confirms the completion of acquisition of the snapshot A in all NAS 400. The time Td indicates the time when the control device 200 instructs all NAS 400 to acquire the snapshot B, which is the second snapshot. The time Te indicates the time when the control device 200 confirms the completion of acquisition of the snapshot B in all NAS 400.

  In FIG. 3, the snapshot acquisition times of the same snapshot ID in each NAS 400 are different. This is because NAS 400 (400a, 400b,...) Is distributed on the network, so even if a snapshot acquisition instruction is sent to all NAS 400 at the same time, the instruction does not arrive at the same time depending on the network conditions. Because.

  In the embodiment of the present invention, in each NAS 400, for example, after the storage area is restored based on the snapshot A, the operation from the time Tb at which acquisition of the snapshot A is instructed to the time Tc at which acquisition is confirmed is performed. Is re-executed. Thereby, as will be described later, in all NAS 400, the state of the storage area when the operation up to time Tc is executed can be restored. Therefore, for example, even when the same file is stored in a plurality of NASs 400 as in mirroring, the file states are the same among the NASs 400 (the file state can be matched).

  Further, the state of the storage area when the operation from the time Tc to the time Te is executed is executed, the operation from the time Tc to the restoration time is performed in addition to the operation from the time Tc to the re-execution. It can be restored by re-execution. As a result, from the operation log from time Tb to time Te and the snapshot A of each NAS 400, the state of the storage area when the operation up to the restoration time is executed for any restoration time from time Tc to time Te. Can be restored.

  Therefore, the operation log acquired by each proxy server 300 and the snapshot acquired by each NAS 400 can be regarded as equivalent to the backup data of the entire file system acquired by securing all the static points of the NAS 400 at the same time. it can.

  Similarly, the state of the storage area when the operation up to an arbitrary restoration time after the time Te is executed can be restored from the operation log and the snapshot B after the time Te to the time Te. Therefore, backup data at an arbitrary point in the entire file system can be acquired by periodically acquiring snapshots. In addition, by periodically and repeatedly acquiring a snapshot, even if the snapshot cannot be acquired at a certain timing due to a temporary failure of the proxy server 300 or the NAS 400, the next snapshot can be used as backup data. . Thereby, it is possible to stably acquire backup data of the entire file system.

  Furthermore, when the restoration time is limited to the time Te or later, the snapshot A and the operation log from the time Ta to the time Td are unnecessary and can be deleted.

  Here, based on the snapshot A and the operation from the time Tb to the time Tc, it will be described that the state of the storage area when the operation up to the time Tc is executed in the NAS 400 can be restored.

  4 and 5 are diagrams showing operation log patterns and file states at the time of restoration according to the snapshot acquisition period in the embodiment of the present invention. 4 and 5, “operation log” for each operation pattern indicates a log of a series of operations acquired for a write operation of a certain file during the snapshot acquisition period. The snapshot acquisition period is a period from the time when the control apparatus 200 instructs the NAS 400 to acquire a snapshot (time Tb) until the completion of acquisition of the snapshot is confirmed (time Tc). “File status (on snapshot)” indicates a case that can be considered as a file status in the snapshot acquired when the operation log is acquired. “File state (on storage area)” indicates a change in the state of the file when the operation log operation is re-executed after the file is restored on the storage area based on the snapshot of each case. .

  In the embodiment of the present invention, as described above, file creation, update, and deletion are performed in file units as file write operations to the NAS 400. In addition, there is no dependency relationship between files, and overwriting to the same file does not occur.

  In this case, the operation log for a certain file in the snapshot acquisition period can be classified as operation patterns 1 to 6 in FIGS.

  Also, as described above, the time at which a snapshot is acquired by each NAS 400 is different for the same snapshot acquisition instruction. Therefore, for each operation pattern, there are a plurality of cases as shown in “file status (on snapshot)” in FIGS. 4 and 5 as the file status in the snapshot acquired by each NAS 400. To do.

  In any case, the operation log operation is re-executed after restoring the file on the storage area based on the snapshot. As a result, the state of the file on the storage area changes as shown in “file state (on storage area)” in FIGS.

  For example, consider a case where a file update is performed in the order of data A1 and A2 during the snapshot acquisition period as in operation pattern 4. Here, as in Case 1, if the file has not been updated when the snapshot is acquired, the file status is updated with the data A2 by re-executing the operation log operation. . Also, as in Case 2, even when the file is updated with the data A1 when the snapshot is acquired, the file status is updated with the data A2 by re-executing the operation log. become. Furthermore, even when the file has already been updated with the data A2 when the snapshot is acquired as in the case 2, the file status is updated with the data A2 by performing the operation of the operation log again. It becomes a state.

  Here, as in the case 2 of the operation pattern 1 and the cases 2 and 3 of the operation pattern 2, when the file is generated when the file has already been generated, the file is generated with the data instructed in the file generation. Assume that it is updated. Also, as in case 3 of the operation pattern 5, it is assumed that when a file is updated when the file has already been deleted, the file is generated with the data instructed by the file update. Further, as in the case 2 of the operation pattern 6, when a file is deleted when the file has already been deleted, it is assumed that the state in which the file has been deleted is retained.

  As described above, in any case of any operation pattern, the file state in the storage area after the re-execution of the operation log operation is restored to the state in which the operation up to the time Tc has been executed in the operation log. The

  Next, details of the configuration of the embodiment of the present invention will be described.

  Referring to FIG. 2, the proxy server 300 includes an operation execution unit 310, an operation information storage unit 320, an operation information acquisition unit 330, an operation re-execution unit 340, and an operation information deletion unit 350.

  The operation execution unit 310 writes (creates, updates, deletes) and reads a file to and from the NAS 400 in response to a request from the client device 500.

  The operation information storage unit 320 stores operation information 321. The operation information 321 represents a log of a write operation to the NAS 400 performed by the operation execution unit 310.

  6 and 7 are diagrams showing examples of the operation information 321 in the embodiment of the present invention.

  The operation information 321 stores the contents of the operation (type of write operation (file creation / update / delete), file name, file data created / updated) for each time when the write operation is performed. In FIG. 6, for example, file A of file data A0 is generated at time “2014/3/1 01:30”, and file A is updated with file data A1 at time “2014/3/1 02:30”. It has been shown.

  The operation information acquisition unit 330 acquires an operation log for the NAS 400 in accordance with an instruction from the control device 200 and stores the operation log in the operation information 321.

  The operation re-execution unit 340 re-executes the operation stored in the operation information 321 in response to an instruction from the control device 200.

  The operation information deletion unit 350 deletes the operation log stored in the operation information 321 in response to an instruction from the control device 200.

  The NAS 400 includes a file storage unit 410, an operation execution unit 420, a snapshot acquisition unit 430, a snapshot storage unit 440, and a snapshot restoration unit 450.

  The file storage unit 410 stores a file on a predetermined storage area of the storage medium of the NAS 400.

  The operation execution unit 420 writes (creates / updates / deletes) and reads a file to and from the file storage unit 410 in response to a request from the proxy server 300.

  The snapshot acquisition unit 430 acquires a snapshot of the state of the storage area of the file storage unit 410 in response to an instruction from the control device 200.

  The snapshot storage unit 440 stores the snapshot acquired by the snapshot acquisition unit 430.

  The snapshot restoration unit 450 restores the state of the storage area of the file storage unit 410 based on the snapshot.

  The control device 200 includes a backup control unit 210, a restoration control unit 220, a control information storage unit 230, and a control information acquisition unit 240.

  The backup control unit 210 instructs each proxy server 300 to acquire an operation log in the backup data generation process. Further, the backup control unit 210 instructs each NAS 400 to acquire a snapshot.

  In the file system restoration process, the restoration control unit 220 instructs each NAS 400 to restore the storage area state based on the snapshot. In addition, the restoration control unit 220 instructs each proxy server 300 to re-execute the operation stored in the operation information 321.

  The control information storage unit 230 stores control information 231. The control information 231 represents logs related to various instructions for the proxy server 300 and the NAS 400 performed by the backup control unit 210 and the restoration control unit 220.

  FIG. 8 is a diagram showing an example of the control information 231 in the embodiment of the present invention.

  The control information acquisition unit 240 acquires operation log acquisition start and snapshot acquisition start / completion logs for the proxy server 300 and the NAS 400 and stores them in the operation information 321.

  The operation information 321 stores the contents of the event for each operation log acquisition start and snapshot acquisition start / completion time. In FIG. 8, for example, acquisition of snapshot A is started at time “2014/3/2 00:00”, and acquisition of snapshot A is completed at time “2014/3/2 00:10”. It is shown.

  Note that the control device 200, the proxy server 300, and the NAS 400 may each be a computer that includes a CPU (Central Processing Unit) and a storage medium that stores a program and that operates by control based on the program. In this case, the CPU of the control device 200 executes a computer program for realizing the functions of the backup control unit 210, the restoration control unit 220, and the control information acquisition unit 240. The CPU of the proxy server 300 executes a computer program for realizing the functions of the operation execution unit 310, the operation information acquisition unit 330, the operation re-execution unit 340, and the operation information deletion unit 350. The CPU of the NAS 400 executes computer programs for realizing the functions of the operation execution unit 420 and the snapshot acquisition unit 430. Further, the storage medium of the control device 200 stores information of the control information storage unit 230. The storage medium of the proxy server 300 stores information of the operation information storage unit 320. The storage medium of the NAS 400 stores information of the file storage unit 410 and the snapshot storage unit 440. In addition, the file storage unit 410 and the snapshot storage unit 440 may be configured as individual storage media or a single storage medium.

  Next, details of the operation of the embodiment of the present invention will be described.

  Here, a case where the storage system 100 includes two proxy servers 300a and 300b and four four NAS 400a to 400d will be described as an example. Further, it is assumed that the proxy server 300a executes an operation related to the files A and B, and the proxy server 300b executes an operation related to the file C. Further, it is assumed that each of the files A, B, and C is written in a distributed manner on the four NAS 400a-d.

  First, backup data generation processing in the embodiment of the present invention will be described.

  FIG. 9 is a flowchart showing a backup data generation process according to the embodiment of this invention.

  The backup control unit 210 of the control device 200 instructs the start of operation log acquisition by transmitting an operation log acquisition start request to all proxy servers 300 (step A101).

  For example, the backup control unit 210 transmits an operation log acquisition start request to the proxy servers 300a and 300b.

  When receiving the operation log acquisition start request, the operation information acquisition unit 330 of the proxy server 300 starts acquiring operation logs for the NAS 400 (step B101).

  The operation information acquisition unit 330 notifies the control device 200 of the start of operation log acquisition by transmitting an operation log acquisition start notification (step B102).

  For example, the operation information acquisition unit 330 of the proxy servers 300 a and 300 b transmits an operation log acquisition start notification to the control device 200. Further, the operation information acquisition unit 330 of the proxy server 300a acquires the operation log of the write operation of the files A and B as shown in FIG. 6 and stores it in the operation information 321. Further, the operation information acquisition unit 330 of the proxy server 300b acquires an operation log of the write operation of the file C and stores it in the operation information 321 as shown in FIG.

  When receiving the operation log acquisition start notification from all the proxy servers 300, the backup control unit 210 of the control device 200 records “operation log acquisition start” in the control information 231 (step A102).

  For example, when the reception of the operation log acquisition start notification from the proxy servers 300a and 300b can be confirmed at the time “2014/3/1 / 00:00”, the backup control unit 210 performs control information as shown in FIG. “Operation log acquisition start” is recorded in 231.

  Next, the backup control unit 210 instructs a snapshot acquisition by transmitting a snapshot acquisition request to all the NAS 400 at a predetermined snapshot acquisition time (step A103). Here, the backup control unit 210 specifies the snapshot ID in the snapshot acquisition request. The snapshot ID is an identifier for uniquely identifying a snapshot acquisition request. The snapshot ID is also an identifier for uniquely identifying the snapshot in each NAS 400.

  For example, at time “2014/3/2/00: 00”, the backup control unit 210 designates the snapshot ID “A” to the NAS 400a to 400d and transmits a snapshot acquisition request.

  The backup control unit 210 records “start of snapshot acquisition (snapshot A)” in the control information 231 (step A104).

  Upon receiving the snapshot acquisition request, the snapshot acquisition unit 430 of the NAS 400 acquires a snapshot of the storage area state of the file storage unit 410 (step C101). The snapshot acquisition unit 430 stores the acquired snapshot in the snapshot storage unit 440 in association with the snapshot ID.

  When the snapshot acquisition is completed, the snapshot acquisition unit 430 notifies the control apparatus 200 of the completion of snapshot acquisition by transmitting a snapshot acquisition completion notification (step C102).

  FIG. 11 is a diagram illustrating an example of a snapshot and a state of a storage area according to the embodiment of the present invention.

  For example, each of the NASs 400a to 400d acquires the snapshot A as shown in FIG. The contents of the snapshot of each NAS 400 differ depending on the timing when the snapshot is actually executed. Each of the NASs 400 a to 400 d transmits a snapshot acquisition completion notification to the control device 200.

  The backup control unit 210 of the control device 200 confirms whether or not snapshot acquisition completion notifications have been received from all NAS 400 (step A105).

  In step A105, for example, when the snapshot acquisition completion notification cannot be sent from all the NASs 400 by a predetermined number of confirmations (step A105 / N), the backup control unit 210 terminates the processing as backup data generation failure (step A110). ).

  In step A105, when the snapshot acquisition completion notification is received from all NAS 400 (step A105 / Y), the backup control unit 210 records “snapshot acquisition completion” in the control information 231 (step A106).

  For example, when the reception of the snapshot acquisition completion notification from the NASs 400a to 400d can be confirmed at the time “2014/3/2/00: 10”, the backup control unit 210 stores the control information 231 as shown in FIG. “Snapshot acquisition completion (snapshot A)” is recorded.

  Further, when there is an unnecessary operation log (step A107 / Y), the backup control unit 210 instructs the proxy server 300 to delete the operation log by transmitting an operation log deletion request (step A108). An unnecessary operation log is, for example, an operation log other than an operation log that is necessary when a specific time or later is used as a restoration time. The specific time is specified, for example, one day before or one month before the current time. The backup control unit 210 refers to the control information 231 and extracts a snapshot related to “snapshot acquisition completion” at the nearest time before the specific time. Then, the backup control unit 210 determines an operation log before the “start of snapshot acquisition” time for the snapshot as an unnecessary operation log. The backup control unit 210 specifies an unnecessary operation log period to be deleted in the operation log deletion request.

  For example, if the current time is “2014/3/5 / 00:10” when used from 2 days before the current time, the time after “2014/3/3 / 00:10” is used as the recovery time. It is done. In this case, based on the control information 231 in FIG. 8, the backup control unit 210 operates at a time prior to the time “2014/3/3/00: 00” of “snapshot acquisition start” related to the snapshot B. It is determined that the log is unnecessary. The backup control unit 210 instructs the proxy servers 300a and 300b to delete the operation log at the time before the time “2014/3/3/00: 00”.

  The operation information deletion unit 350 of the proxy server 300 deletes the operation log for the specified period stored in the operation information 321 (step B103).

  Note that the backup control unit 210 may similarly instruct the NAS 400 to delete unnecessary snapshots in addition to unnecessary operation logs.

  The backup control unit 210 of the control device 200 ends the processing as successful backup data generation (step A109).

  Thereafter, the processes from steps A103, B103, and C101 are repeated at regular snapshot acquisition times.

  Next, file system restoration processing in the embodiment of the present invention will be described.

  FIG. 10 is a flowchart showing file system restoration processing according to the embodiment of the present invention.

  The restoration control unit 220 of the control device 200 receives an input of the restoration time of the file system from an administrator or the like (step A201).

  For example, the restoration control unit 220 receives the restoration time “2014/3/2 00:30”.

  Based on the control information 231, the restoration control unit 220 determines a snapshot to be used for restoring a storage area in the file storage unit 410 of the NAS 400 (step A202). Here, the restoration control unit 220 determines a snapshot related to “snapshot acquisition completion” at the nearest time before the restoration time as a snapshot used for restoring the storage area.

  For example, based on the control information 231 in FIG. 8, the restoration control unit 220 relates a snapshot used for restoring the storage area to “snapshot acquisition completion” at time “2014/3/2 00:10”. The snapshot A is determined.

  The restoration control unit 220 of the control device 200 instructs the restoration of the storage area using the snapshot by transmitting a snapshot restoration request to the NAS 400 (step A203). Here, the restoration control unit 220 specifies the snapshot ID extracted in step A203 in the snapshot acquisition request.

  For example, the restoration control unit 220 instructs the NASs 400a to 400d to restore the storage area using the snapshot A.

  Upon receiving the snapshot restoration request, the snapshot restoration unit 450 of the NAS 400 restores the storage area of the file storage unit 410 using the snapshot (step C201). Here, the snapshot restoration unit 450 restores the state of the storage area using the snapshot with the snapshot ID specified in the snapshot restoration request.

  When the restoration of the storage area is completed, the snapshot restoration unit 450 notifies the completion of restoration of the storage area by transmitting a snapshot restoration completion notification to the control device 200 (step C202).

  For example, the snapshot restoration unit 450 of each of the NASs 400a to 400d uses the snapshot A in FIG. 11 to restore the storage area state as in the state a. The state of the storage area of each NAS 400 differs depending on the contents of the snapshot A. The snapshot restoration unit 450 of each of the NAS 400 a to d transmits a snapshot restoration completion notification to the control device 200.

  The restoration control unit 220 of the control device 200 confirms whether or not snapshot restoration completion notifications have been received from all NAS 400 (step A204).

  In step A204, for example, when the snapshot restoration completion notification cannot be received from all the NASs 400 after a predetermined number of confirmations (step A204 / N), the restoration control unit 220 ends the processing as a file system restoration failure (step S204). A209).

  In step A204, when the snapshot restoration completion notification is received from all the NAS 400 (step A204 / Y), the restoration control unit 220 determines the operation re-execution start time based on the operation information 321 (step A205). Here, the restoration control unit 220 determines, from the operation information 321, the time when “snapshot acquisition start” related to the snapshot extracted in step A 203 is recorded as the operation re-execution start time.

  For example, the restoration control unit 220 sets the time “2014/3/2 00:00” of “snapshot acquisition start” related to the snapshot A as the operation re-execution start time based on the control information 231 of FIG. decide.

  The restoration control unit 220 instructs the proxy server 300 to re-execute the operation stored in the operation information 321 by transmitting an operation re-execution request (step A206). Here, the restoration control unit 220 specifies the operation re-execution start time determined in step A203 and the restoration time determined in step A201 in the operation re-execution request.

  For example, the restoration control unit 220 sets the operation re-execution start time “2014/3/2 00:00” and the restoration time “2014/3/2 00:30” to the proxy servers 300a and 300b. Specify and send an operation re-execution request.

  The operation re-execution unit 340 of the proxy server 300 re-executes the operation stored in the operation information 321 (step B201). Here, the operation re-execution unit 340 re-executes the operation from the operation re-execution start time to the restoration time stored in the operation information 321 to the NAS 400.

  When the re-execution of the operation up to the restoration time is completed, the operation re-execution unit 340 notifies the completion of the operation re-execution by transmitting an operation re-execution completion notification to the control device 200 (step B202).

  For example, the operation re-execution unit 340 of the proxy server 300a uses the operation re-execution start time “2014/3/2 00:00” to the restoration time “2014/3/2 00: 30 ”for the files A and B (within the dotted lines) is re-executed for the NAS 400a-d. Similarly, the operation re-execution unit 340 of the proxy server 300b uses the operation re-execution start time “2014/3/2 00:00” to the restoration time “2014/3/2 00 based on the operation information 321 of FIG. : 30 ”for the file C (within the dotted line) is re-executed for the NASs 400a to 400d. As a result, the state of each storage area of the NASs 400a to 400d is restored to the same state as in the state b of FIG.

  The operation re-execution unit 340 of the proxy servers 300a and 300b transmits an operation re-execution completion notification to the control device 200.

  The restoration control unit 220 of the control device 200 confirms whether or not an operation re-execution completion notification has been received from all proxy servers 300 (step A207).

  In step A207, for example, when the operation re-execution completion notification cannot be received from all the proxy servers 300 by a predetermined number of confirmations (step A207 / N), the restoration control unit 220 performs processing as a file system restoration failure. The process ends (step A209).

  In step A206, when the operation re-execution completion notification is received from all the proxy servers 300 (step A207 / Y), the restoration control unit 220 ends the process as a file system restoration success (step A208).

  Thus, the operation of the embodiment of the present invention is completed.

  In the above description, the case where the same file is stored in all of the plurality of NAS 400 has been described as an example. However, the same file may be any one or a part of the plurality of NAS 400. May be stored.

  Next, a characteristic configuration of the embodiment of the present invention will be described. FIG. 1 is a block diagram showing a characteristic configuration of an embodiment of the present invention.

  Referring to FIG. 1, the storage system 100 includes an operation information storage unit 320 and an operation re-execution unit 340.

  The operation information storage unit 320 stores operation information 321 indicating operations related to files executed on a plurality of NAS 400 (storage apparatuses).

  The operation re-execution unit 340 performs a plurality of operations related to the file after the state of the storage area for storing the files in the plurality of NAS 400 is restored based on the snapshots acquired by the plurality of NAS 400 before the restoration time. The NAS 400 is re-executed. Here, the operation re-execution unit 340 re-executes the operation related to the file from the time when the snapshot acquisition is instructed to the plurality of NAS 400 included in the operation information 321 to the restoration time.

  According to the embodiment of the present invention, backup data of the entire file system in the distributed file system can be acquired without stopping the service. The reason is that the operation re-execution unit 340 performs a plurality of operations related to the file after the state of the storage area in the plurality of NAS 400 is restored based on the snapshot acquired by the plurality of NAS 400 before the restoration time. This is because the NAS 400 is re-executed. Here, the operation re-execution unit 340 re-executes the operation related to the file from the time when the snapshot acquisition is instructed to the plurality of NAS 400 included in the operation information 321 to the restoration time. Thereby, the state of the plurality of NAS 400 becomes the state when the file operation is performed up to the restoration time. The snapshot can be acquired without making the user recognize the service stop.

  Therefore, the state of the distributed file system in which consistency is ensured between the file servers (NAS 400) at the restoration time is restored without stopping the service.

  Moreover, according to the embodiment of the present invention, the backup data of the file system can be maintained in an appropriate amount. The reason is that the operation information deletion unit 350 deletes, from the operation information 321, an operation log that has become unnecessary due to periodic snapshot acquisition or the like.

  While the present invention has been described with reference to the embodiments, the present invention is not limited to the above embodiments. 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.

  For example, in the embodiment of the present invention, the case where the snapshot is stored in the snapshot storage unit 440 of the NAS 400 and the operation log is stored in the operation information storage unit 320 of the proxy server 300 has been described as an example. However, the present invention is not limited to this, and the snapshot and operation log may be stored in a recording medium outside the storage system 100 such as a tape device. This is because the snapshot and operation log can be copied without stopping the service operation of the system and without being restricted in time.

  In the embodiment of the present invention, the case where the file system is restored on the same NAS 400 based on the snapshot and the operation log acquired regarding the NAS 400 has been described as an example. However, the present invention is not limited to this, and a file system at a specific restoration time may be restored on another NAS 400 having the same capacity. As a result, failure analysis using a file system restored on another NAS 400, or utilization of the restored file system as a spare file system becomes possible.

DESCRIPTION OF SYMBOLS 100 Storage system 200 Control apparatus 210 Backup control part 220 Restoration control part 230 Control information storage part 231 Control information 240 Control information acquisition part 300 Proxy server 310 Operation execution part 320 Operation information storage part 321 Operation information 330 Operation information acquisition part 340 Execution unit 350 Operation information deletion unit 400 NAS
410 File storage unit 420 Operation execution unit 430 Snapshot acquisition unit 440 Snapshot storage unit 450 Snapshot restoration unit 500 Client device

Claims (10)

A first device that instructs a plurality of storage devices to acquire a snapshot;
An operation execution means for executing an operation related to a file on the plurality of storage devices;
Storing operation information indicating an operation relating to the file that is executed for the plurality of storage devices, and the operation information storage means,
Wherein the plurality of states of the storage area for storing the file in the storage device, after being restored to acquired by the plurality of storage devices to restore the time before the snapshot to the original, are included in the operation information, the the operation related to the file from the time that instructs acquisition of the snapshot pair to said plurality of storage devices in the first device to said recovery time, re-executed for the plurality of storage devices, operation redo Means,
A second device comprising:
Storage system with The state of the storage area in the plurality of storage devices is restored based on the snapshot acquired by the plurality of storage devices at a time closest to the restoration time before the restoration time.
The storage system according to claim 1. The snapshot acquisition instruction for the plurality of storage devices is periodically performed.
The storage system according to claim 1 or 2. The second device is further acquired by the plurality of storage devices at a time closest to the one time before the one time, which is included in the operation information, when the restoration time is after one time. An operation information deleting unit that deletes an operation related to the file before the time when the acquisition of the performed snapshot is instructed;
The storage system according to any one of claims 1 to 3, further comprising: The same file is stored in one or more of the plurality of storage devices, and the same operation relating to the file is executed on the same file on the one or more storage devices.
The storage system according to any one of claims 1 to 4. The operation related to the file includes generation, update, and deletion of the file.
The storage system according to any one of claims 1 to 5. In a storage system comprising: a first device that instructs a plurality of storage devices to acquire a snapshot; and a second device that executes an operation related to a file with respect to the plurality of storage devices.
The second device stores operation information indicating an operation related to a file executed on the plurality of storage devices,
After the second device restores the state of the storage area for storing the files in the plurality of storage devices based on the snapshots acquired in the plurality of storage devices before the restoration time, included in the operation information is an operation relating to the file from the time when the instruction to acquire the snapshot pair to said plurality of storage devices in the first device to said recovery time, to the plurality of storage devices Re-execute,
Storage method. The state of the storage area in the plurality of storage devices is restored based on the snapshot acquired by the plurality of storage devices at a time closest to the restoration time before the restoration time.
The storage method according to claim 7. In a storage system comprising: a first device that instructs a plurality of storage devices to acquire a snapshot; and a second device that executes a file operation on the plurality of storage devices. A program for a device of
On the computer,
Storing operation information indicating operations related to files executed on the plurality of storage devices;
Wherein the plurality of states of the storage area for storing the file in the storage device, after being restored to acquired by the plurality of storage devices to restore the time before the snapshot to the original, are included in the operation information, the the operation according to the plurality of storage devices in pairs to the file from the time that instructs acquisition of the snapshot to the restoration time in the first device and re-executed for the plurality of storage devices,
A program that executes processing. The state of the storage area in the plurality of storage devices is restored based on the snapshot acquired by the plurality of storage devices at a time closest to the restoration time before the restoration time.
The program according to claim 9.

Images