JP6676203B2 - Management device, storage system, storage management method, and program - Google Patents

Description

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

  2. Description of the Related Art In recent years, a distributed storage system such as an object storage system (hereinafter, referred to as an object storage) has been known (for example, see Patent Document 1). In such a storage system, for example, data to be stored is distributed and stored in a predetermined number of replicas (replicas), thereby ensuring data reliability. Here, the number of replicas refers to the number of replicas (replicas) including original data. In a storage system, as the number of replicas increases, the reliability improves, but the unit price of storage capacity tends to increase.

Japanese Patent No. 5396848

  However, in the storage system described above, the reliability guaranteed by the set number of replicas is at a minimum guaranteed level, and the reliability is improved as much as possible to reduce the risk of data loss. It is desirable to make it. Further, in the storage system as described above, even if there is a surplus capacity of the storage, it is not reflected on the number of replicas, and there is a possibility that the surplus capacity of the storage cannot be effectively used.

  The present invention has been made to solve the above problems, and an object of the present invention is to provide a management apparatus, a storage system, and a storage system that can dynamically improve the reliability of stored data while effectively utilizing the surplus capacity of storage. It is to provide a storage management method and a program.

  In order to solve the above problem, one embodiment of the present invention is a replica number indicating the number of copies of an object that is a set of data to be stored, and the object is distributed based on the set number of replicas. A storage unit for detecting the free space of the storage system, the free space of the storage system, and the importance of each container that stores the object as a lower hierarchy. The change information of the number of replicas including a decimal value set correspondingly, wherein the lower the importance, the lower the number of replicas from the state where the free space of the storage system is large, and the number of replicas is set to decrease. A change information storage unit for storing change information of the number of replicas, and a free space of the storage system detected by the detection unit; The change information of the number of replicas corresponding to the importance of each container storing the object is read from the change information storage unit, and the lower limit of the predetermined number of replicas is determined according to the read change information of the number of replicas. A changing unit configured to change a setting of the number of replicas including the decimal value as an average value for each container so as not to fall below a value.

  According to another aspect of the present invention, in the management device described above, the management device further includes a replica number storage unit that stores an upper limit value and a lower limit value of the replica number, and the change unit stores the container and the importance in association with each other. The importance of each container is obtained from the importance storage unit, and based on the obtained importance of each container and the free space, the upper limit value and the lower limit value stored in the replica number storage unit are stored. The setting of the number of replicas is changed for each container.

Further, according to one aspect of the present invention, in the management device described above, the replica count storage unit stores an upper limit value and a lower limit value of the replica count for each importance, and among the upper limit value and the lower limit value of the replica count. Is set based on the importance of the container.
Further, according to one aspect of the present invention, in the above management device, the detecting unit periodically detects the free space of the storage system, and the changing unit detects the free space of the storage system 1 detected by the detecting unit. The setting of the number of replicas is periodically changed according to

  Further, according to an aspect of the present invention, the management device and the object stored in the container are controlled such that an average number of replicas in the container becomes the number of replicas set by the management device. And a server device for adjusting the number of replicas.

  Further, according to one aspect of the present invention, in the above storage system, the storage system further includes a plurality of storage devices in which a copy of the stored object is deleted or added based on the number of replicas changed by the management device, When deleting or adding a copy of the object in the storage device, the server device gives priority to the object having a larger size among the objects stored under the container.

  Further, one aspect of the present invention is a replica number indicating the number of copies of an object that is a set of data to be stored, and the object is distributed and stored based on the set number of replicas. And the replica number change information including a decimal value set corresponding to the importance of each container that stores the object under the lower layer, and the lower the importance is, A storage management method for a storage system, comprising: a change information storage unit that stores change information of the number of replicas set so that the number of replicas is reduced from a state where the free space of the storage system is large, wherein the detection unit includes: A detecting step of detecting a free space of the storage system; and a changing unit detecting the storage capacity detected in the detecting step. The change information of the number of replicas corresponding to the free space of the storage system and the importance of each container storing the object is read from the change information storage unit, and predetermined according to the read change information of the number of replicas. Changing the setting of the number of replicas including the decimal value as an average value for each container so that the number of replicas does not fall below the lower limit of the number of replicas.

  Further, one aspect of the present invention is a replica number indicating the number of copies of an object that is a set of data to be stored, and the object is distributed and stored based on the set number of replicas. And the replica number change information including a decimal value set corresponding to the importance of each container that stores the object under the lower layer, and the lower the importance is, Detecting the free space of the storage system in a computer of the storage system including a change information storage unit that stores change information of the number of replicas set so that the number of replicas is reduced from a state where the free space of the storage system is large. Detecting the free space of the storage system detected in the detecting step, The change information of the number of replicas corresponding to the degree of importance of each container storing the object is read from the change information storage unit, and according to the read change information of the number of replicas, a predetermined lower limit of the number of replicas is determined. To change the setting of the number of replicas including the decimal value as an average value for each container so as not to fall below the average value for each container.

  According to the present invention, it is possible to dynamically improve the reliability of stored data while effectively utilizing the surplus capacity of a storage.

FIG. 2 is a block diagram illustrating an example of a storage system according to the first embodiment. FIG. 3 is a diagram illustrating an example of a ring file storage unit according to the embodiment. FIG. 4 is a diagram illustrating a data example of ring information of a ring file according to the embodiment. FIG. 6 is a diagram illustrating a data example of device list information of a ring file according to the embodiment. FIG. 4 is a diagram illustrating an example of data of account information according to the embodiment. It is a figure showing the example of data of container list information in this embodiment. It is a figure showing the example of data of container information in this embodiment. FIG. 6 is a diagram illustrating a data example of object list information according to the present embodiment. FIG. 6 is a diagram illustrating an example of data in a replica number storage unit according to the embodiment. It is a figure showing an example of change information on the number of replicas in this embodiment. FIG. 2 is a diagram illustrating an example of a logical hierarchy structure of a storage according to the present embodiment. FIG. 7 is a diagram illustrating an example of a process of changing the number of replicas according to the embodiment. 9 is a flowchart illustrating an example of an object adding process according to the embodiment. It is a flow chart which shows an example of change processing of the number of replicas in this embodiment. 5 is a flowchart illustrating an example of a process of adjusting the number of replicas according to the embodiment. FIG. 9 is a block diagram illustrating an example of a storage system according to a second embodiment. FIG. 14 is a diagram illustrating an example of data in a replica number storage unit according to the second embodiment.

  Hereinafter, a storage system and a management device according to an embodiment of the present invention will be described with reference to the drawings.

[First Embodiment]
FIG. 1 is a block diagram illustrating an example of a storage system 1 according to the first embodiment.
As shown in FIG. 1, the storage system 1 includes a proxy node 20, a plurality of storage nodes (31, 32, 33,...), And a management device 10. The proxy node 20, the storage nodes (31, 32, 33,...), And the management device 10 are connected to each other via a network NT. The storage system 1 can be connected to the client terminal 2 via the network NT.

The storage system 1 is, for example, an object storage system, and stores data in a distributed manner based on the set number of replicas. Here, the number of replicas indicates the number of data to be copied and stored, and is the number of replicas including original data.
In this embodiment, a group of data is referred to as an “object”. The storage system 1 stores one object in a distributed manner among a plurality of storage nodes (31, 32, 33,...) Based on the set number of replicas.

  Each of the storage nodes 31, 32, 33,... Includes a ring file storage unit (311, 321, 331), which will be described later, and has the same configuration for storing objects. The storage node 31, the storage node 32, the storage node 33,... Indicate an arbitrary storage node included in the storage system 1, or will be described as the storage node 30 unless otherwise specified.

  The client terminal 2 is a computer device that transmits a write request or a read request of an object to a storage provided in the storage system 1. Is a server device provided to the user. The client terminal 2 requests the proxy node 20 to write, read, or delete an object.

The proxy node 20 is, for example, a server device, and distributes processing to each storage node 30 in response to a request from the client terminal 2. For example, when receiving a request to write an object from the client terminal 2, the proxy node 20 selects a plurality of storage nodes 30 specified by a ring file described later from all the storage nodes 30, and selects the selected storage nodes 30. To send a write request for the same object. In the example illustrated in FIG. 1, an example in which the storage system 1 includes one proxy node 20 will be described. However, the storage system 1 may include a plurality of proxy nodes 20.
The proxy node 20 includes a storage unit 210.

The storage unit 210 stores information used for various processes of the proxy node 20. The storage unit 210 includes, for example, a ring file storage unit 211.
The ring file storage unit 211 stores a ring file. Here, the ring file is, for example, table information that associates data (for example, an object) with a physical location (a server device or a disk) where the data is stored. Details of the ring file will be described later. The ring file storage unit 211 stores a plurality of ring files corresponding to the importance of data, for example, as shown in FIG.

FIG. 2 is a diagram illustrating an example of the ring file storage unit 211 according to the present embodiment.
In the example shown in this figure, the ring file storage unit 211 has a plurality of ring files (“ring file A”, “ring file B”,...). Each ring file stored in the ring file storage unit 211 corresponds to the degree of importance of data, and in the present embodiment, the ring file storage unit 211 stores a ring file for each degree of importance.

  The ring file includes, for example, ring information which is a table for specifying a storage destination (device) of the object, and device list information. The ring information includes a head part (prefix) of the hash value calculated from the object name, a storage location (including a device ID and a storage location candidate) of the object corresponding to the head part, the number of replicas, and the importance. The associated information. Here, it is assumed that the hash value is calculated so that storage destinations are evenly allocated.

FIG. 3 is a diagram illustrating a data example of ring information of a ring file according to the present embodiment.
As shown in FIG. 3, the ring information is, for example, information in which a “prefix”, a device ID corresponding to the “prefix”, “importance”, and “the number of replicas” are associated with each other. Further, the ring information includes “first device ID”, “second device ID”, “third device ID”,... Corresponding to “prefix”. Here, the “prefix” is the head of the hash value calculated from the object name described above. Further, “first device ID”, “second device ID”, “third device ID”,... Indicate identification information for identifying a device that stores an object.

  In the example illustrated in FIG. 3, for example, an object whose “prefix” is “00” has “1” as the storage destination, “3” as the first device ID, “0” as the “second device ID”, This indicates that the “third device ID” is stored in the order of “2”,. The ring information also indicates that the “importance” is “50” and the set “number of replicas” is “3.0”. In the ring information, in consideration of the fact that the device is not operating, etc., more devices than the set “number of replicas” are specified for each “prefix”.

FIG. 4 is a diagram illustrating a data example of device list information of a ring file according to the present embodiment.
As shown in FIG. 4, the device list information includes a plurality of pieces of information in which “device ID”, “IP address”, “port”, and “device name” are associated. Here, "device ID" indicates identification information for identifying a device (server device, disk). “IP address” and “port” indicate the IP address and port number of the device. “Device name” indicates the name of the device.

In the example shown in FIG. 4, the “IP address” of the device corresponding to “0” of the “device ID” is “xxxx”, the “port” is “8000”, and the “device name” is “ZZ”. ".
In this way, the ring file sets the “number of replicas” and specifies the “IP address”, “port”, and “device name” of the device that is the storage destination from the name of the object.

Returning to the description of FIG. 1, the storage node 30 (an example of a storage device) is, for example, a server device including a storage (storage device), and performs processing in response to a request received from the proxy node 20, for example. For example, when receiving a request to write or read an object from the proxy node 20, the storage node 30 writes or reads the object and transmits the object to the proxy node 20.
The storage nodes 30 include a storage node 31 (account server) for managing accounts, a storage node 32 (container server) for managing containers, and the like, in addition to a normal storage node 33 for storing objects. Here, the container indicates, for example, a container for storing an object, and corresponds to a directory or a folder in a file system. The account indicates, for example, a container that stores a container. As described above, the upper layer structure of the object is the container, the upper layer of the container is the account, and the object, the container, and the account indicate the data layer structure.

The storage node 31 (account server) stores objects and manages accounts, for example. The storage node 31 includes, for example, a storage unit 310.
The storage unit 310 stores information used for various processes of the storage node 31. The storage unit 310 includes, for example, a ring file storage unit 311 and an account information storage unit 312.

The ring file storage unit 311 is the same as the ring file storage unit 211 described above, and a description thereof will be omitted.
The account information storage unit 312 stores account management information for managing accounts. The account information storage unit 312 stores, for example, account information and container list information as account management information.

FIG. 5 is a diagram illustrating a data example of account information according to the present embodiment.
As illustrated in FIG. 5, the account information storage unit 312 stores, for example, account information in association with “account name”, “number of containers”, “hash value of name”, and “account ID”. Here, the “account name” indicates the name of the account, and the “hash value of the name” indicates the hash value of the “account name”. The “number of containers” indicates the number of containers existing under the account, and the “account ID” indicates identification information for identifying the account.
In the example shown in FIG. 5, the account whose “account name” is “ABC” has “3” containers under the account, the “hash value of the name” is “XXXXXXXXXXXXXX”, and the “account ID” is “YYYYYYYYY”. ".

FIG. 6 is a diagram illustrating a data example of the container list information in the present embodiment.
As shown in FIG. 6, the account information storage unit 312 stores, for example, “NO.”, “Container name”, “creation time”, “deletion time”, “name hash value”, And container list information having a plurality of pieces of information associated with the “ID”. Here, “NO.” Indicates the line number of the container list information, and “container name” indicates the name of the container. “Creation time” and “deletion time” indicate the creation time and deletion time of the container, and “name hash value” indicates the hash value of the “container name”. “Container ID” indicates identification information for identifying a container.

  In the example shown in FIG. 6, the container whose “NO.” Is “1” has a “container name” of “testcont1” and a “creation time” of “2014-10-01 10: 00: 00: 00” Is shown. The “hash value of the name” is “YYYYYYYYYYYY”. This indicates that the “container ID” is “ZZZZZ”.

Returning to the description of FIG. 1 again, the storage node 32 (container server) stores objects and manages containers, for example. The storage node 32 includes, for example, a storage unit 320.
The storage unit 320 stores information used for various processes of the storage node 32. The storage unit 320 includes, for example, a ring file storage unit 321 and a container information storage unit 322.

The ring file storage unit 321 is the same as the ring file storage unit 211 described above, and a description thereof will be omitted.
The container information storage unit 322 (an example of the importance storage unit) stores container management information for managing containers. The container information storage unit 322 stores, for example, container information and object list information as container management information.

FIG. 7 is a diagram illustrating an example of data of container information according to the present embodiment.
As illustrated in FIG. 7, the container information storage unit 322 stores, for example, “account name”, “container name”, “creation time”, “deletion time”, “object number”, and “byte number”. , “Name hash value”, “container ID”, and “importance” are stored in association with each other as container information. Here, “account name” indicates the name of a higher-level account, and “container name” indicates the name of the container. The “creation time”, “deletion time”, “name hash value”, and “container ID” are the same as the container list information described above. The “number of objects” and the “number of bytes” indicate the number of objects and the number of bytes of data under the container, and the “importance” indicates the importance of the container.
As described above, the container information storage unit 322 stores a container (hierarchical structure) for storing data in association with importance.

  In the example illustrated in FIG. 7, “account name” indicates container information in which “container name” under “ABC” is “testcont1”, and “object number” and “byte number” are “14” and “byte”. 8096 "bytes. Also, it indicates that the “importance” of the container is “50”.

FIG. 8 is a diagram illustrating a data example of the object list information in the present embodiment.
As shown in FIG. 8, the container information storage unit 322 stores, for example, “NO.”, “Object name”, “creation time”, “byte number”, “content type”, and “hash value”. And object list information having a plurality of pieces of information associated with the “deletion flag”. Here, “NO.” Indicates the line number of the object list information, and “Object name” indicates the name of the object. “Creation time” indicates the creation time of the object, and “number of bytes” and “content type” indicate the number of bytes of the object and the type of the object (for example, image, moving image, text, and the like). The “hash value” indicates a hash value of the “object name”, and the “deleted flag” indicates whether the object has been deleted.

In the example shown in FIG. 8, the object whose “NO.” Is “1” has the “object name” of “obj1” and the “creation time” of “2014-10-02 10: 00: 00: 00”. Is shown. Also, this indicates that the “number of bytes” is “3072” bytes and the “content type” is “image / jpeg” (jpeg image). The “hash value” is “AAAAAAAAAAAAA”. This indicates that the “deleted flag” is “NO”.
In the above-described example, an example in which one account and one container are used has been described, but the storage system 1 may include a plurality of accounts and containers.

Returning to the description of FIG. 1 again, the storage node 33 (30) stores, for example, an object and includes, for example, a storage unit 330.
The storage unit 330 stores information used for various processes of the storage node 33. The storage unit 330 includes, for example, a ring file storage unit 331.
The ring file storage unit 331 is the same as the above-described ring file storage unit 211, and a description thereof will not be repeated.

The management device 10 is a computer device that manages the number of replicas of the storage system 1. The management device 10 includes, for example, a management storage unit 110 and a management control unit 120.
The management storage unit 110 stores information used for various processes of the management device 10. The management storage unit 110 includes, for example, a replica number storage unit 111 and a change information storage unit 112.

The number-of-replicas storage unit 111 stores, for example, an upper limit value and a lower limit value of the number of replicas, as shown in FIG. In this embodiment, the upper limit value and the lower limit value of the number of replicas in the entire storage system 1 are predetermined, and the replica number storage unit 111 stores the upper limit value and the lower limit value of the number of replicas in the storage system 1 as a whole. Is stored.
FIG. 9 is a diagram illustrating an example of data in the replica number storage unit 111 according to the present embodiment.
As shown in this figure, the replica number storage unit 111 stores an “upper limit” and a “lower limit” in association with each other. Here, “upper limit” indicates the upper limit of the number of replicas in the entire storage system 1, and “lower limit” indicates the lower limit of the number of replicas in the entire storage system 1.
In the example shown in FIG. 9, the “upper limit” is “5” and the “lower limit” is “2”.

  Returning to the description of FIG. 1 again, the change information storage unit 112 stores the change information of the number of replicas for changing the setting of the number of replicas according to the free space of the storage system 1. The change information storage unit 112 is, for example, table information that specifies the number of replicas corresponding to the free space of the storage system 1 in a range between the “upper limit value” and the “lower limit value” stored in the replica number storage unit 111. is there. Further, the change information storage unit 112 may store different change information according to the importance of each container, for example. The change information storage unit 112 stores, for example, table information that is changed according to the free space of the storage system 1, as shown in FIG.

FIG. 10 is a diagram illustrating an example of change information of the number of replicas in the present embodiment.
The graph shown in FIG. 10 shows the setting of the number of replicas with respect to the free space, the vertical axis shows the number of replicas, and the horizontal axis shows the free space (%).
As shown in this figure, the change information of the number of replicas, between “2” of the “lower limit value” and “5” of the “upper limit value” stored in the replica number storage unit 111, The value of the number of replicas is set to be large. In this figure, a waveform W1 shows an example of change information when the importance is the highest, and a waveform W2 shows an example of change information when the importance is the next highest. The waveform W3 shows an example of the change information when the importance is the lowest.
As described above, in the embodiment, the change information of the number of replicas is set in accordance with the degree of importance, and the lower the degree of importance, the lower the number of replicas from the state of more free space.

  Note that the number of replicas is not limited to a natural number, and may be a real number including a natural number and a decimal number. When the number of replicas is a real number including a decimal, for example, as shown in FIG. 11, the storage system 1 duplicates and stores each object so that the number of replicas can be realized as an average value for each container.

FIG. 11 is a diagram illustrating an example of a logical hierarchical structure of a storage according to the present embodiment.
The example illustrated in FIG. 11 illustrates an example in which, for example, a container C1 exists under the account A1. The container C1 stores “Object 1” (OB1), “Object 2” (OB2), and “Object 3” (OB3). Here, “Object 1” (OB1) and “Object 2” (OB2) each have a replica number of “3” and are stored at three locations. “Object 3” (OB3) has a replica count of “2” and is stored at two locations. In this case, the number of replicas of the entire container C1 is “2.666” according to the following equation (1).

  Number of replicas = (3 + 3 + 2) /3=2.666 (1)

  As described above, in the storage system 1 according to the present embodiment, it is possible to set the number of replicas for each container by using real numbers including decimal numbers.

  Returning to the description of FIG. 1 again, the management control unit 120 is, for example, a processor including a CPU (Central Processing Unit), and controls the management device 10 in an integrated manner. The management control unit 120 includes, for example, a free space detection unit 121 and a ring change unit 122.

  The free space detection unit 121 (an example of a detection unit) detects the free space of the storage system 1. The free space detecting unit 121 periodically detects the free space of the storage system 1 at predetermined time intervals. The free space detecting unit 121 acquires, for example, account management information stored in an account server corresponding to each account and container management information stored in a container server corresponding to each container, and acquires the acquired account management information and container management. Based on the information, the free capacity of the storage system 1 is detected.

  The ring changing unit 122 (an example of a changing unit) sets the number of replicas according to the free space of the storage system 1 detected by the free space detecting unit 121 so as not to fall below a predetermined lower limit of the number of replicas. change. The ring changing unit 122 acquires, for example, the importance of each container from the container information storage unit 322 that stores the container storing the data in association with the importance. The ring changing unit 122 changes the setting of the number of replicas between the upper limit and the lower limit stored in the replica number storage unit 111 for each container based on the acquired importance of each container and the free space. That is, the ring changing unit 122 determines the number of replicas for each container based on the importance and the free space, changes the number of replicas in the ring file to the determined number of replicas, and sets each node (proxy node). 20 and the storage node 30). That is, the ring changing unit 122 updates (replaces) the ring file stored in each node.

  For example, the ring changing unit 122 acquires the upper limit value and the lower limit value stored in the replica number storage unit 111. When the upper limit value is “5” and the lower limit value is “2”, the ring changing unit 122 refers to the change information stored in the change information storage unit 112 and acquires the replica corresponding to the acquired importance for each container. Number (the number of replicas between an upper limit of “5” and a lower limit of “2”). For example, as shown in FIG. 10, the ring changing unit 122 determines the number of replicas including a natural number and a decimal number according to the free space of the storage system 1. The ring changing unit 122 generates a ring file reflecting the determined number of replicas. Then, the ring changing unit 122 distributes the generated ring file to each node (the proxy node 20 and the storage node 30).

  The above-described proxy node 20 periodically checks the number of replicas in the ring file stored in the ring file storage unit 211 and the container management information stored in the container information storage unit 322 to determine the number of replicas in the ring file. Change the number of replicas for each object accordingly. That is, the proxy node 20 periodically deletes or adds copies of each object so that the average number of replicas of each container matches the number of replicas in the ring file. That is, the proxy node 20 adjusts the number of replicas of the data stored under the container such that the average number of replicas in the container becomes the number of replicas set by the management device 10.

Next, the operation of the storage system 1 according to the present embodiment will be described with reference to the drawings.
FIG. 12 is a diagram illustrating an example of a process of changing the number of replicas according to the present embodiment.
FIG. 12 shows an example in which the storage system 1 has five containers “container A” to “container E”. In this example, the importance of “container A” and “container B” is “100”, the importance of “container C” and “container D” is “50”, and the importance of “container C” is It shall be “10”. It is assumed that the greater the value, the higher the importance. Further, in this example, it is assumed that the “upper limit” stored in the replica number storage unit 111 is “3” and the “lower limit” is “2”.

  FIG. 12A shows a state where the free space of the storage system 1 is large (for example, 80%). In this case, the management device 10 sets the number of replicas of the ring file corresponding to the importance “100” to “3”, the number of replicas of the ring file corresponding to the importance “50” to “3”, and sets the importance to “3”. Sets the number of replicas of the ring file corresponding to “10” to “3”. In this case, the proxy node 20 adjusts by deleting or adding a copy of each object so that the average number of replicas of each container matches the number of replicas in the ring file.

  Next, FIG. 12B shows a state (for example, 50%) in which the free space of the storage system 1 is reduced as compared with FIG. 12A. In this case, the management device 10 sets the number of replicas of the ring file corresponding to the importance “100” to “3”, the number of replicas of the ring file corresponding to the importance “50” to “3”, and sets the importance to “3”. Set the number of replicas of the ring file corresponding to “10” to “2.5”, respectively. In this case, the proxy node 20 deletes and adjusts the copy of the object of “Container E” so that the number of replicas of “Container E” matches “2.5”.

  Next, FIG. 12C shows a state (for example, 10%) in which the free space of the storage system 1 is further reduced from that of FIG. 12B. In this case, the management device 10 sets the number of replicas of the ring file corresponding to the importance “100” to “3”, the number of replicas of the ring file corresponding to the importance “50” to “2.6”, The number of replicas of the ring file corresponding to the importance of “10” is set to “2”. In this case, the proxy node 20 deletes the duplicates of the objects of “Container C” and “Container D” so that the number of replicas of “Container C” and “Container D” match “2.6”. adjust. Also, the proxy node 20 deletes and adjusts the copy of the object of “Container E” so that the number of replicas of “Container E” matches “2”.

  As described above, the management device 10 (ring change unit 122) sets the number of replicas for each container between the upper limit and the lower limit stored in the replica number storage unit 111 based on the free space of the storage system 1. change.

Next, with reference to FIG. 13, the operation of the object addition processing of the storage system 1 in the present embodiment will be described.
FIG. 13 is a flowchart illustrating an example of an object adding process according to the present embodiment.

  In the figure, first, the proxy node 20 determines whether or not there is an object addition request (step S101). That is, the proxy node 20 determines, for example, whether or not there is an object addition request from the client terminal 2 via the network NT. When there is an object addition request (step S101: YES), the proxy node 20 advances the processing to step S102. When there is no object addition request (step S101: NO), the proxy node 20 returns the process to step S101.

  In step S102, the proxy node 20 acquires the management information of the container. That is, the proxy node 20 acquires the container management information (see FIGS. 7 and 8) stored in the container information storage unit 322 of the storage node 33 (container server). Here, the proxy node 20 acquires the importance of the container, for example.

  Next, the proxy node 20 refers to the ring file corresponding to the container (Step S103). That is, the proxy node 20 refers to the ring file corresponding to the importance of the container from the ring file storage unit 211.

  Next, the proxy node 20 determines the arrangement of the object to be added based on the referred ring file (step S104). The proxy node 20 determines a plurality of devices specified by the ring information (see FIG. 3) in the ring file based on the setting of the number of replicas in the ring file. Here, the proxy node 20 specifies a device corresponding to the “device ID” of the ring information from the device list information shown in FIG. The proxy node 20 instructs the device of each storage node 30 about the determined arrangement of the object.

  Next, the storage node 30 arranges the object based on the object arrangement instruction (step S105). That is, the storage node 30 adds the object based on the object placement instruction.

  Next, the proxy node 20 updates the management information of the container (Step S106). That is, the proxy node 20 adds the object to the container management information stored in the container information storage unit 322 of the storage node 33 (container server) to which the object has been added, for example. For example, the proxy node 20 changes the “number of objects” and the “number of bytes” of the container information stored in the container information storage unit 322, and adds a row corresponding to the added object to the object list information. After the processing in step S106, the proxy node 20 returns the processing to step S101.

  In the example shown in FIG. 13 described above, an example in which an object is added has been described. However, the case in which an object is deleted is basically the same as the processing shown in FIG. However, when deleting an object, the proxy node 20 changes the “deletion flag” of the row corresponding to the object to be deleted to “YES” in the object list information stored in the container information storage unit 322.

Next, with reference to FIG. 14, an operation of the change processing of the number of replicas of the storage system 1 in the present embodiment will be described.
FIG. 14 is a flowchart illustrating an example of a process of changing the number of replicas according to the present embodiment. It is assumed that the process illustrated in FIG. 14 is periodically executed at predetermined time intervals (for example, in units of several minutes or several hours).

  In this figure, first, the management device 10 detects the free space of the storage system 1 (Step S201). That is, the free space detection unit 121 of the management device 10 acquires, for example, account management information stored in the account server corresponding to each account and container management information stored in the container server corresponding to each container. The free space detection unit 121 detects the free space of the storage system 1 based on the acquired account management information and container management information.

  Next, the management device 10 determines the number of replicas according to the free space (Step S202). That is, the ring changing unit 122 of the management device 10 determines the number of replicas based on the free space detected by the free space detection unit 121. Specifically, based on the change information stored in the change information storage unit 112 and the free space, the ring change unit 122 is configured to fall within the range of the upper limit value and the lower limit value of the number of replicas stored in the replica number storage unit 111. Next, the number of replicas for each importance is determined. The number of replicas determined here is a real number value including a natural number and a decimal number.

  Next, the ring changing unit 122 generates a ring file (Step S203). That is, the ring changing unit 122 generates a ring file for each importance based on the determined number of replicas for each importance. Specifically, for example, the ring changing unit 122 generates a ring file in which the number of replicas in the current ring file has been changed. In the present embodiment, “importance” is stored in the container information stored in the container information storage unit 322, and the importance is determined for each container. Therefore, the ring changing unit 122 changes the number of replicas for each container by changing the number of replicas of the ring file.

  Next, the ring changing unit 122 distributes the generated ring file (Step S204). That is, the ring changing unit 122 transmits the generated ring file for each importance to the proxy node 20 and the storage node 30. As a result, the proxy node 20 and the storage node 30 store the received ring file in the ring file storage unit (211, 311, 321, 331) and update the ring file. As a result, the number of replicas set in the ring file held by each node (proxy node 20 and storage node 30) is updated. After the processing in step S204, the ring changing unit 122 ends the processing.

Next, with reference to FIG. 15, a description will be given of a process of adjusting the number of replicas in which the number of replicas for each container of the storage node 30 is changed based on the updated ring file in the storage system 1.
FIG. 15 is a flowchart illustrating an example of a process of adjusting the number of replicas according to the present embodiment. Note that the process illustrated in FIG. 15 is periodically executed at predetermined time intervals (for example, in units of several minutes).

  In this figure, first, the storage system 1 checks a ring file (step S301). For example, the proxy node 20 checks each ring file stored in the ring file storage unit 211. The proxy node 20 checks, for example, the importance of the ring file and the number of replicas.

  Next, the proxy node 20 checks the container object corresponding to the ring file (Step S302). For example, the proxy node 20 acquires the container management information stored in the container information storage unit 322, and confirms the object in the container by using a ring file whose importance of the container management information matches. The proxy node 20 calculates, for example, the average number of replicas in the current container.

  Next, the proxy node 20 determines whether or not the arrangement of the object needs to be changed (Step S303). That is, the proxy node 20 determines, for each container, whether or not the number of replicas set in the ring file matches the average number of replicas in the current container. When it is necessary to change the arrangement of the objects (step S303: YES), the proxy node 20 advances the processing to step S304. In addition, the proxy node 20 ends the process when it is not necessary to change the arrangement of the objects (Step S303: NO).

  In step S304, the proxy node 20 determines an object to increase or decrease the number of copies (replica). That is, the proxy node 20 determines the object to be deleted or added so that the number of replicas set in the ring file and the average number of replicas in the container match in the container that needs to be changed. When deleting or adding a copy of an object, the proxy node 20 gives priority to an object having a large size among objects in the container. That is, when deleting or adding a copy of an object in the storage node 30, the proxy node 20 gives priority to an object having a larger size among objects stored under the container.

  Next, the proxy node 20 deletes or adds the determined object (Step S305). The proxy node 20 instructs the storage node 30 to delete or add the determined object to delete or add the determined object. The storage node 30 deletes or adds an object based on an instruction from the proxy node 20. The deletion or addition of an object is executed for a device set by a corresponding ring file.

  Next, the proxy node 20 updates the management information of the container (Step S306). That is, the proxy node 20 updates the container management information corresponding to the container from which the object has been deleted or added, out of the container management information stored in the container information storage unit 322. After the processing in step S306, the proxy node 20 ends the processing.

  As described above, in the storage system 1, the proxy node 20 periodically checks the ring file and the container management information so that the number of replicas becomes equal to the number of replicas according to the change in the setting of the number of replicas of the ring file. Adjust each time.

As described above, the management device 10 according to the present embodiment is a management device of the storage system 1 that stores data (eg, objects) in a distributed manner based on the set number of replicas. 121 (detection unit) and a ring change unit 122 (change unit). Here, the number of replicas indicates, for example, the number of times data is copied and stored. The free space detection unit 121 detects the free space of the storage system 1. The ring changing unit 122 changes the setting of the number of replicas according to the free space of the storage system 1 detected by the free space detecting unit 121 so as not to fall below a predetermined lower limit of the number of replicas.
Accordingly, the management device 10 according to the present embodiment increases the number of replicas when the free space is large, and decreases the number of replicas when the free space is small. Can be enhanced. Therefore, the management device 10 according to the present embodiment can dynamically improve the reliability of the stored data while effectively utilizing the surplus capacity of the storage.

In the present embodiment, the management device 10 includes a replica number storage unit 111 that stores an upper limit value and a lower limit value of the number of replicas. The ring changing unit 122 acquires the importance for each container from the container information storage unit 322 (importance storage unit) that stores the containers (hierarchical structure) for storing data and the importance in association with each other. The ring changing unit 122 changes the setting of the number of replicas for each container between the upper limit value and the lower limit value stored in the replica number storage unit 111 based on the acquired importance of each container and the free space.
Thereby, the management device 10 according to the present embodiment appropriately changes the number of replicas between the upper limit value and the lower limit value stored in the replica number storage unit 111 based on the importance of each container and the free space. can do.

In the present embodiment, the ring changing unit 122 generates a ring file in which the number of replicas has been changed, and transmits the generated ring file to the proxy node 20 (proxy device) and the storage node 30 (storage device).
Accordingly, by providing the ring file in which the number of replicas has been changed according to the free space and importance to the proxy node 20 and the storage node 30, the management apparatus 10 according to the present embodiment can use the conventional object storage. The number of replicas can be dynamically changed by a simpler method.

The storage system 1 according to the present embodiment is stored under the container such that the average number of replicas in the management device 10 and the container (hierarchical structure) becomes the number of replicas set by the management device 10. And a proxy node 20 that adjusts the number of replicas of existing data.
Thereby, the storage system 1 according to the present embodiment can set a decimal value for the number of replicas, for example, and can increase the degree of freedom in setting the number of replicas.

Further, the storage system 1 according to the present embodiment includes a plurality of storage nodes 30 (storage devices) from which a copy of stored data is deleted or added based on the number of replicas changed by the management device 10. . When deleting or adding a copy of data in the storage node 30, the proxy node 20 gives priority to data having a large size among data stored under the container (hierarchical structure).
Accordingly, when the number of replicas is changed, for example, the number of replicas is changed from a large-sized object, so that the storage system 1 according to the present embodiment can quickly respond to a change in free space. Therefore, the storage system 1 according to the present embodiment can dynamically improve the reliability of the stored data while effectively utilizing the surplus capacity of the storage.

Further, the storage management method according to the present embodiment is a storage management method for the storage system 1 that stores data in a distributed manner based on the number of replicas indicating the number of data to be copied and stored, based on the set number of replicas. And includes a detecting step and a changing step. In the detection step, the management device 10 detects the free space of the storage system 1. In the change step, the management device 10 changes the setting of the number of replicas according to the free space of the storage system 1 detected in the detection step so as not to fall below a predetermined lower limit of the number of replicas.
Thus, the storage management method according to the present embodiment can dynamically improve the reliability of the stored data while effectively utilizing the surplus capacity of the storage, similarly to the management device 10 and the storage system 1 described above. .

[Second embodiment]
Next, a storage system and a management device according to a second embodiment will be described with reference to the drawings.
This embodiment differs from the first embodiment in that the upper limit and the lower limit of the number of replicas are changed according to the importance.

FIG. 16 is a block diagram illustrating an example of the storage system 1a according to the second embodiment.
As shown in this figure, the storage system 1a includes a proxy node 20, a plurality of storage nodes (31, 32, 33,...), And a management device 10a. The proxy node 20, the storage nodes (31, 32, 33,...), And the management device 10a are connected to each other via a network NT. The storage system 1a can be connected to the client terminal 2 via the network NT.
In this figure, the same components as those shown in FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted.

The management device 10a is a computer device that manages the number of replicas of the storage system 1a. The management device 10a includes, for example, a management storage unit 110a and a management control unit 120a.
The management storage unit 110a stores information used for various processes of the management device 10a. The management storage unit 110a includes, for example, a replica number storage unit 111a and a change information storage unit 112.

The replica number storage unit 111a stores, for example, an upper limit value and a lower limit value of the number of replicas as illustrated in FIG. In this embodiment, the upper limit and the lower limit of the number of replicas are determined in advance for each importance, and the replica number storage unit 111a associates the importance with the upper limit and the lower limit of the number of replicas. To remember.
FIG. 17 is a diagram illustrating an example of data in the replica number storage unit 111a according to the present embodiment.
As shown in the figure, the replica number storage unit 111a stores “importance”, “upper limit”, and “lower limit” in association with each other. The “importance” here is, for example, an importance set for each container.
The example shown in FIG. 17 indicates that when the “importance” is “100”, the “upper limit” is “5” and the “lower limit” is “3”. When the “importance” is “10”, it indicates that the “upper limit” is “3” and the “lower limit” is “2”.
As described above, the replica number storage unit 111a stores the upper limit value and the lower limit value of the replica number for each importance. In the example shown in FIG. 17 described above, an example in which the upper limit and the lower limit of the number of replicas are changed for each importance has been described. However, at least one of the upper limit and the lower limit of the number of replicas is changed for each importance. It may be changed.

  Returning to the description of FIG. 16, the change information storage unit 112 of the present embodiment is the same as that of the first embodiment, except that the upper limit and the lower limit of the number of replicas are set for each importance.

  The management control unit 120a is, for example, a processor including a CPU and the like, and controls the management device 10a as a whole. The management control unit 120a includes, for example, a free space detection unit 121, a ring change unit 122a, and an importance change unit 123.

  The ring changing unit 122a is basically the same as the ring changing unit 122 in the first embodiment, except that the upper limit value and the lower limit value for each importance stored in the replica number storage unit 111a are used. The ring changing unit 122a acquires the importance of each container from the container information storage unit 322, and acquires the upper limit and the lower limit of the number of replicas corresponding to the acquired importance of each container from the replica number storage unit 111a. The ring changing unit 122a changes the setting of the number of replicas between the acquired upper limit and lower limit of the number of replicas for each container.

  The importance changing unit 123 (an example of a changing unit) changes the importance of each container in response to a change request from the client terminal 2, for example. The importance change unit 123 changes the value of “importance” stored in the container information storage unit 322 based on the acquired change request. The importance changing unit 123 may change the importance according to the access frequency to each container, in addition to the change request from the client terminal 2. That is, the importance changing unit 123 detects the access frequency to each container, and executes, for example, a process of increasing the importance of the container having a high access frequency. In the storage system 1a of the present embodiment, by increasing the importance of a container with a high access frequency, the upper limit and the lower limit of the number of replicas corresponding to the container are changed, and the reliability of the stored data is improved.

As described above, in the management device 10a according to the present embodiment, the replica number storage unit 111a stores the upper limit value and the lower limit value of the replica number for each importance. Here, at least one of the upper limit value and the lower limit value of the number of replicas is set based on the importance of the container (hierarchical structure).
Thus, the management device 10a according to the present embodiment can flexibly change the number of replicas for each container according to the degree of importance.

The management device 10a according to the present embodiment includes an importance changing unit 123 that changes the importance according to the frequency of access to each container.
Thereby, the management device 10a according to the present embodiment can dynamically improve the reliability of the stored data while effectively utilizing the surplus capacity of the storage according to the access frequency to the container.

The present invention is not limited to the above embodiments, and can be modified without departing from the spirit of the present invention.
For example, in each of the embodiments described above, an example in which importance is set for each container has been described as an example of a hierarchical structure. However, the present invention is not limited to this, and importance is set for each object and each account. You may make it.

Further, in each of the above embodiments, an example in which a ring file is provided for each importance level has been described. However, the present invention is not limited to this. For example, a ring file may be provided for each container.
Further, in each of the above embodiments, the example in which the proxy node 20 executes the process of adjusting the number of replicas has been described, but another device such as a container server (the storage node 32) may execute the process.
Also, an example in which the proxy node 20 periodically executes the process of adjusting the number of replicas has been described. However, the present invention is not limited to this, and a request for addition or deletion (deletion) of an object is received from the client terminal 2. In this case, the process of adjusting the number of replicas may be executed.

Further, in each of the above embodiments, an example has been described in which the management device 10 (10a) is realized by one device, but may be realized by a plurality of server devices. Further, a part or all of the management storage unit 110 (110a) may be provided outside the management device 10 (10a).
Further, the example in which the management device 10 (10a) includes the change information storage unit 112 has been described. However, instead of including the change information storage unit 112, the management device 10 (10a) calculates the difference between the upper limit value and the lower limit value by another means such as calculation. The number of replicas may be generated.

  In each of the above embodiments, an example has been described in which, when deleting or adding a copy of data (for example, an object), giving priority to data having a large size (for example, an object), the present invention is not limited to this. For example, data to be deleted or added (for example, an object) may be determined randomly, or may be determined according to another rule.

Each configuration of the storage system 1 (1a) has a computer system inside. Then, a program for realizing the function of each configuration provided in the storage system 1 (1a) is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read by a computer system and executed. By doing so, the processing in each configuration of the storage system 1 (1a) described above may be performed. Here, "to make the computer system read and execute the program recorded on the recording medium" includes installing the program in the computer system. The “computer system” here includes an OS and hardware such as peripheral devices.
Further, the “computer system” may include a plurality of computer devices connected via a network including a communication line such as the Internet, a WAN, a LAN, and a dedicated line. The “computer-readable recording medium” refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, and a CD-ROM, and a storage device such as a hard disk built in a computer system. As described above, the recording medium storing the program may be a non-transitory recording medium such as a CD-ROM.

  The recording medium also includes an internal or external recording medium accessible from the distribution server for distributing the program. It should be noted that the program may be divided into a plurality of pieces, downloaded at different timings, and then combined with each other included in the storage system 1 (1a), or the distribution server that distributes each of the divided programs may be different. . Further, the “computer-readable recording medium” holds a program for a certain period of time, such as a volatile memory (RAM) in a computer system serving as a server or a client when the program is transmitted via a network. Shall be included. Further, the program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

  Further, some or all of the above functions may be realized as an integrated circuit such as an LSI (Large Scale Integration). Each of the above-described functions may be individually formed into a processor, or a part or all of the functions may be integrated into a processor. The method of circuit integration is not limited to an LSI, and may be realized by a dedicated circuit or a general-purpose processor. Further, in the case where a technology for forming an integrated circuit that replaces the LSI appears due to the advance of the semiconductor technology, an integrated circuit based on the technology may be used.

1, 1a Storage system 2 Client terminal 10, 10a Management device 20 Proxy node 30, 33 Storage node 31 Storage node (account server)
32 Storage node (container server)
110, 110a Management storage unit 111, 111a Replica number storage unit 112 Change information storage unit 120, 120a Management control unit 121 Free space detection unit 122, 122a Ring change unit 123 Importance change unit 210, 310, 320, 330 Storage unit 211 , 311, 321, 331 Ring file storage 312 Account information storage 322 Container information storage NT Network

Claims (8)

A management apparatus for a storage system that stores the objects in a distributed manner based on the set number of replicas, the number of replicas indicating the number of copies and storage of objects that are a set of data,
A detection unit for detecting the free space of the storage system;
The change information of the number of replicas, including the free space of the storage system and a decimal value set corresponding to the importance of each container that stores the object as a lower tier, wherein the importance is low. A change information storage unit that stores change information of the number of replicas set such that the number of replicas decreases from a state where the free space of the storage system is large,
The change information of the replica number corresponding to the free space of the storage system detected by the detection unit and the importance of each container storing the object is read from the change information storage unit, and the read change of the replica number is performed. A changing unit that changes the setting of the number of replicas including the decimal value as an average value for each container so as not to fall below a predetermined lower limit of the number of replicas according to information. Management device. A replica number storage unit that stores an upper limit value and a lower limit value of the replica number,
The change unit includes:
The importance of each container is acquired from the importance storage unit that stores the container in association with the importance, and based on the acquired importance of each container and the free space, the number of replicas storage unit is The management device according to claim 1, wherein the setting of the number of replicas is changed for each container between the upper limit value and the lower limit value to be stored. The replica number storage unit stores an upper limit value and a lower limit value of the replica number for each importance,
The management device according to claim 2, wherein at least one of an upper limit and a lower limit of the number of replicas is set based on the importance of the container. The detection unit periodically detects the free space of the storage system,
The said change part changes the setting of the said number of replicas periodically according to the free space of the storage system 1 which the said detection part detected. The Claim 1 characterized by the above-mentioned. The management device according to 1. A management device according to any one of claims 1 to 4,
A server device that adjusts the number of replicas of the object stored under the container so that the average number of replicas in the container becomes the number of replicas set by the management device. system. Based on the number of replicas changed by the management device, comprising a plurality of storage devices in which a copy of the stored object is deleted or added,
The server device, when deleting or adding a copy of the object in the storage device, gives priority to an object having a large size among the objects stored under the container. 6. The storage system according to 5. The number of replicas indicating the number of objects to be copied and stored as a set of data, and the objects are distributed and stored based on the set number of replicas. The change information of the number of replicas including a decimal value set corresponding to the importance of each container that stores the object under the subordinate, wherein the lower the importance, the more free space in the storage system. A storage management method for a storage system, comprising: a change information storage unit that stores change information of the number of replicas set to decrease the number of replicas,
A detecting step of detecting a free space of the storage system,
The change unit reads out the change information of the number of replicas corresponding to the free space of the storage system detected in the detection step and the importance of each container storing the object from the change information storage unit, and reads the change information. A changing step of changing the setting of the number of replicas including the decimal value as an average value for each container so as not to fall below a predetermined lower limit of the number of replicas according to the change information of the number of replicas. A storage management method comprising: The number of replicas indicating the number of objects to be copied and stored as a set of data, and the objects are distributed and stored based on the set number of replicas. The change information of the number of replicas including a decimal value set corresponding to the importance of each container that stores the object under the subordinate, wherein the lower the importance, the more free space in the storage system. In a computer of a storage system including a change information storage unit that stores change information of the number of replicas set to decrease the number of replicas,
A detecting step of detecting a free space of the storage system;
The change information of the replica number corresponding to the free space of the storage system detected in the detecting step and the importance of each container storing the object is read from the change information storage unit, and the read replica number is Changing the setting of the number of replicas including the decimal value as an average value for each container so as not to fall below a predetermined lower limit of the number of replicas in accordance with the change information. program.

Images