JP2007018407A - Data replication system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a distributed data replication system that can efficiently and easily decide a replication destination storage device and can decide an optimum replication destination. <P>SOLUTION: The distributed data replication system, which stores a replica of data stored in an area in a plurality of storage devices 140 used by a plurality of computers 101 to 103 in an area in a different storage device, is configured to have dedicated areas and replicated areas and replicate data stored in the dedicated areas into the replicated areas of different storage devices. The computers 101 to 103 each have a policy holding part 160 for holding a redundancy level specifying the total number of replicated areas in different storage devices, and a configuration information holding part 170 for specifying the location of the dedicated areas, the location of the replicated areas corresponding to the dedicated areas and the location of unused replicated areas. A management server 200 for managing the data replication has a processing request control part 300 for deciding a replicated area in a different storage device in response to a replicated area decision request from each computer or storage device. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a data replication system, and more particularly to a method for managing a plurality of storage areas connected to a network.

  In preparation for a failure of a storage device such as a hard disk that stores a large amount of data, the data is made redundant by RAID (Redundant Array of Independent Disks) technology, or a backup storage device such as RAID A system that prepares a magnetic tape and stores data for backup in the storage device has been put into practical use.

  As a backup-type storage device, RAID requires a large-scale device, so that there is a problem that it becomes an expensive system. Tape backup also has the problem that it takes time to recover the data.

  For example, Patent Document 1 (Japanese Patent Application Laid-Open No. 2004-46874) discloses a method of replicating data in a plurality of storage areas connected to a network, and a data distribution system. There is a possibility that the above problems can be solved.

JP 2004-46874 A

However, the method disclosed in Document 1 has the following problems.
First, the method of determining the copy destination is inefficient or the setting of management information is complicated. As a specific method for determining the connection destination, a device ID (IP address in the literature) is specified at random, and a device ID that provides a service fortunately is used. A method of setting a destination list and selecting a connection destination by some method has been proposed. The former method is inefficient because the device ID that provides the service is not determined at one time. In the latter method, since all connection destination information needs to be set in advance for all apparatuses, management is complicated.

  Secondly, the copy data storage destination cannot be selected according to the policy of each storage area. For example, when remote storage is carried out in preparation for a disaster, the replication destination needs to be separated from the original data storage location. In the proposed method, such a condition is added to determine the replication destination. I can't.

  Accordingly, an object of the present invention is to provide a distributed data replication system that can efficiently and easily determine a storage device of a replication destination and can determine an optimal replication destination according to the purpose.

The present invention is a distributed data replication system for storing data stored in a certain area of a plurality of storage devices used by a plurality of computers as a replica in a certain area of another storage device, and used by the computer The storage device includes a dedicated area and a replication area, and is configured to replicate data stored in the dedicated area to a replication area of another storage device.
The computer defines a policy holding unit that holds redundancy that defines the total number of replication areas in other storage devices, the location of the dedicated area, the location of the replication area corresponding to the dedicated area, and the location of the unused replication area Having a configuration information holding unit;
The management server that manages data replication includes a processing request control unit that determines a replication area in another storage device in response to a replication area determination request from each computer or storage device. In the system.

  According to the present invention, it is possible to automatically select a storage device as a data duplication destination without intervention of a user operation.

Example 1
FIG. 1 illustrates a distributed data replication system according to one embodiment.
The distributed data replication system is configured by connecting a plurality of computers 101 to 103 and a management server 200 to each other via a network 400. The computers 101 to 103 (the reference numeral 101 is used as a representative unless otherwise specified) includes a processor 110, a memory 120, a storage device control mechanism 130, a communication unit 150, a policy holding unit 160, and a configuration information holding unit 170. The storage device 140 is coupled to the bus 180 and connected to the storage device control mechanism 130. The processor 110 is configured by a CPU, for example, and the memory 120 is configured by a RAM, for example.

  The storage device control mechanism 130 is a mechanism for controlling the storage device 140 in response to a storage device processing request from the processor 110, and is configured by logic such as TTL or CMOS. The storage device 140 can be configured by secondary storage such as a hard disk. The communication unit 150 is a connection network for connecting the computers 101 to 103 and the management server 200 to perform mutual communication, and is configured by logic such as TTL or CMOS, for example.

  The configuration information holding unit 170 is a part that holds the location of an area in which the computer 101 mainly reads and writes, the location of a duplication area of the area, and the like, and is configured by, for example, a RAM or an electrically rewritable ROM. The policy holding unit 160 holds information that becomes a requirement when determining a replication area corresponding to an area in which the computer 101 mainly reads and writes. In the first embodiment, the policy holding unit 160 holds the redundancy. Redundancy is information that designates the total number of areas that the computer 101 mainly reads and writes and the duplication areas. In the first embodiment, the policy holding unit is provided in each computer, but may be provided in the management server if it can be referred to and edited from each computer. The policy holding unit 160 includes, for example, a RAM or an electrically rewritable ROM.

The management server 200 receives a replication area determination request to be described later from each of the computers 101 to 103, determines a replication area, and notifies the location of the determined replication area to the computer that issued the replication area determination request.
In the management server 200, the processor 210, the memory 220, the storage device control mechanism 230, the communication unit 250, the configuration information holding unit 270, and the processing request control unit 300 are coupled to the internal bus 280, and the storage device 240 is connected to the storage device control mechanism 230. Connected and configured. The processor 210 is, for example, a CPU, and the memory 220 is, for example, a RAM.

  The storage device control mechanism 230 is a mechanism for controlling the storage device 240 in response to a storage device processing request from the processor 210, and is configured by logic such as TTL or CMOS. The storage device 240 is configured by secondary storage such as a hard disk. The communication unit 250 is a connection network for connecting the computers 101 to 103 and the management server 200 to perform mutual communication, and is configured by logic such as TTL or CMOS.

  The configuration information holding unit 270 holds configuration information of an area for storing main data, configuration information of an area for storing a copy thereof, and configuration information of an unused area in a storage device connected to a computer constituting the system. To do. The configuration information holding unit is configured by, for example, a RAM or an electrically rewritable ROM. The processing request control unit 300 receives a replication area determination request issued from the computers 101 to 103, serializes it, and processes it. The processing request control unit is configured by logic such as TTL or CMOS.

FIG. 3 shows an example of area allocation in each storage device 140 according to an embodiment.
In the illustrated example, two storage devices 1411 and 1421 are connected to the computer 101 ((1) (2)), one storage device 1412 is connected to the computer 102 (3), and the computer 103 is connected to the computer 103. One storage device 1413 is connected (4), and one storage device 241 is connected to the management server 200 (5). Each of the computers 101 to 103 performs reading and writing using the hatched portion of the storage device connected to the computer as a main storage area. This area is called a dedicated area. In FIG. 3, for example, the computer 102 indicates that the hatched portion of the storage device 1402 is handled as the dedicated area C.

  In a distributed data replication system, a copy of a dedicated area can be stored in another storage device in preparation for a failure of a dedicated area in its own computer, and this area is called a replication area. In FIG. 3, the white portion indicates a duplication area. For example, an area indicated by “C” in the storage device 1401 indicates a duplication area corresponding to a dedicated area of the storage device 1402. It is possible to use not only one duplication area but also a plurality of duplication areas. For example, as the duplication area corresponding to the dedicated area of the storage device 1421, the area indicated by “B” in the storage device 1402 and the storage device 1403 is used. The number of such replication areas is defined by the policy holding unit in the computer to which the corresponding dedicated area is connected.

  The replication area corresponding to the dedicated storage area is sequentially allocated at the timing when the dedicated area is used. By assigning at such timing, it is possible to select the best replication area at the time of use. For example, the dedicated area of the storage device 1401 is unused. As a replication area corresponding to this area, an area marked “(unused)” in the storage device 1421 or 1403 can be selected. The management server 200 assigns the replication area. In addition, a copy area determination request issued by each computer to the management server triggers allocation.

  The capacity of each dedicated area is a value obtained by dividing the capacity of the storage device in which the dedicated area exists by the redundancy. In this way, all dedicated areas can have duplicate areas. However, in this case, the replication area may have to be set in the same storage device as the dedicated area. In this case, if the storage device fails, the replication area cannot be accessed in addition to the dedicated area. In order to avoid such a situation, the management server 200 may include a storage device 241 and reserve a replication area in the storage device 241. The maximum value of the capacity of the replication area that needs to be secured in the storage device 241 in the management server is equal to the capacity of the maximum capacity storage device connected to this system. For this reason, it is desirable to place a restriction so that the capacity of the storage device connected to each computer is equal to or less than the capacity of the storage device 241 connected to the management server.

  In this embodiment, each storage device is divided into units called segments (Seg), and a dedicated storage area or a replication area is assigned to each unit. As another method, it is also possible to allocate a dedicated area or a replication area in units of LBA.

4 to 6 show examples of information held in the configuration information holding unit 270 in the management server 200. FIG.
FIG. 4 is an example of the dedicated area management table 271. The dedicated area management table 271 holds the location of all the dedicated areas in the system and the reference destination to the information related to the corresponding replication area. The dedicated area management table 271 holds the storage device start position (denoted as Seg-ID), area length (denoted as Seg length), storage apparatus-ID, chassis-ID, and redundant information storage position. To do. Here, for example, an IP address can be used as the case-ID. For the storage device-ID, for example, a logical device number defined by SCSI or a drive letter defined by an OS such as Windows (registered trademark) can be used. The redundant information storage position is information indicating the storage position of the replication area information corresponding to the dedicated area. In this example, entry # of a used replication area management table 272 described later is set. As shown in entries # 1 to # 3 of the dedicated area management table 271, in the case of a dedicated storage area for which a replication area has been determined, some value is set in the redundant information storage position. Further, as shown in entry # 0, when the replication area is an undetermined dedicated storage area, all redundant information storage positions are “none”. Registration in the dedicated area management table is performed when the system is started up.

FIG. 5 shows an example of the used replication area management table 272.
This used replication area management table 272 holds the location of all used replication areas in the system and the reference destination of the information if there is a subsequent replication area. The used replication area management table 272 holds the storage device start position (denoted as Seg-ID), area length (denoted as Seg length), storage apparatus-ID, case-ID, and subsequent entries. . The subsequent entry is an area that indicates the storage location of segment information that follows the segment when a replication area corresponding to a dedicated area with a Seg length of 2 or more is secured non-continuously. An entry # of the area management table 272 is set. If there is no subsequent entry, this value is “none”. Registration in the used replication area management table is performed when the management server allocates a replication area.

FIG. 6 shows an example of the unused replication area management table 273.
This unused duplication area management table 273 holds the locations of all used duplication areas in the system. In the unused replication area management table 273, the storage system start position (indicated as Seg-ID), area length (indicated as Seg length), storage apparatus-ID, and enclosure of the unused replication area in the system. Holds Body-ID. Registration in the unused replication area management table is performed when the system is started up. The unused replication area management table is corrected when the management server allocates a replication area.

9, 10 and 12 show examples of information held in the configuration information holding unit 170 in the computers 101 to 103 (the example shown shows the case of the computer 101).
FIG. 9 shows an example of the dedicated area management table 171. The dedicated area management table 171 includes a start position (denoted as Seg-ID), an area length (denoted as Seg length), a storage apparatus-ID, and a chassis-ID of all the dedicated areas in the computer 101. And the redundant information storage position is held. The redundant information storage position is information indicating the storage position of the replication area information corresponding to the dedicated area. In this example, the entry # of the used replication area management table 172 shown below is set. As shown in entry # 2 of the dedicated area management table 171, in the case of a dedicated storage area for which a replication area has been determined, some value is set in the redundant information storage position. Further, as shown in entry # 0, when the replication area is an undetermined dedicated storage area, all redundant information storage positions are “none”. Registration in the dedicated area management table is performed when each computer participates in the system.

FIG. 10 shows an example of the copy destination management table 172.
The replication destination management table 172 includes a storage area start position (denoted as Seg-ID), an area length (denoted as Seg length), and a replication area corresponding to the dedicated area managed by the dedicated area management table 171. The storage device-ID, the housing-ID, and the subsequent entry are held. The subsequent entry is an area that indicates the storage location of segment information that follows the segment when a replication area corresponding to a dedicated area having a Seg length of 2 or more is secured non-continuously. An entry # of the area management table 172 is set. If there is no subsequent entry, this value is “none”.

FIG. 12 shows an example of contents managed by the policy holding unit 160 in the computer 101.
The policy holding unit holds a policy that serves as a determination criterion for determining a replication area corresponding to a dedicated area in a storage area connected to the computer 101. In the first embodiment, the redundancy, the storage device-ID, and the chassis-ID for determining the number of replication areas are essential, but the number of errors, distance, and performance can be omitted.

Next, with reference to the flowchart of FIG. 14, the replication area determination processing operation in the first embodiment will be described.
First, a copy area determination request is generated with reference to the policy holding unit 160 in the computer 101 (1000). In the present embodiment, the replication area determination request includes the request source chassis-ID, storage device-ID, and redundancy. The redundancy can be obtained by referring to the redundancy of the dedicated area held in the policy holding unit 160 in the computer.
Next, a replication area determination request is issued to the management server 200 (1010). In the management server 200, the processing request control unit 300 receives a replication area determination request (1020). Then, a replication area corresponding to the dedicated area is selected (1030).

The duplicate area selection process (1030) will be described in more detail with reference to FIG.
In FIG. 15, first, a storage device different from the request source storage device is selected with reference to the unused replication area management table 273 stored in the configuration information holding unit 270 (1100). The selection method is, for example, the largest entry # among the storage devices different from the request source storage device.

  Next, the case-ID, the storage device-ID, and the Seg-ID corresponding to the selected replication area are registered in the used replication area management table 272 as information on the selected replication area. Also, 1 is registered as the Seg length of this entry. However, when registering continuously in the already registered replication area, the chassis-ID, storage device-ID, and Seg-ID are not registered, and the Seg length is set to 1 for the registered entry. What is necessary is just to register the added value. Whether to register continuously in the already registered replication area is set in the redundant information storage location of the entry that stores the dedicated area information in the dedicated area management table managed in the management server Is referred to, the subsequent entry of the same entry as this value is referred to in the replication destination management table, the entries in the replication destination management table are sequentially referenced until this subsequent entry becomes “none”, and finally The obtained case-ID and storage device-ID match the case-ID and storage device-ID corresponding to the selected replication area, and the Seg-ID value + Seg length in the entry is the Seg of the selected replication area -Perform by checking whether the ID matches. Further, in step 1110, 1 is added to the Seg-ID of the entry in order to exclude the information of the replication area selected from the unused replication area management table 273. However, when the Seg length of the entry is 1, the subsequent entries are copied to one younger entry, and the last entry is filled with “none” (1110).

  Next, it is determined whether or not the number of selected replication areas has reached the redundancy value −1 in the replication area determination request. If it has been reached, the replication area selection processing is terminated. The processing after 1100 is repeated (1120).

  Returning to FIG. 14, when the number of selected replication areas reaches the redundancy value −1 in the replication area determination request, the used replication area management table 272 is referred to and a response message to the replication area determination request is generated. (1040). The response message to the replication area determination request includes a list of the chassis-ID, storage device-ID, and Seg-ID indicating the determined replication area. Next, a response message to the replication area determination request is sent to the computer 101 that issued the request (1050).

  The computer 101 that issued the request receives a response message to the replication area determination request (1060). Then, the chassis-ID, storage device-ID, and Seg-ID returned in the response message to the replication area determination request are registered in the replication destination management table 172 managed by the configuration information holding unit 170 in the computer 101 ( 1070). Also, 1 is registered as the Seg-ID of this entry. However, when registering continuously in the already registered replication area, the chassis-ID, storage device-ID, and Seg-ID are not registered, and the Seg length is set to 1 for the registered entry. What is necessary is just to register the added value. Whether to register continuously in the already registered replication area is set in the redundant information storage position of the entry storing the dedicated area information in the dedicated area management table 171 managed in the computer. Is referred to, the subsequent entry of the same entry as this value is referred to in the replication destination management table 172, and this entry in the replication destination management table is sequentially referred to until the subsequent entry becomes “none”. The case-ID and storage device-ID obtained in the above are matched with the response message, and the Seg-ID value + Seg length in the entry matches the Seg-ID in the response message.

  Further, in step 1070, the redundant information storage position in the dedicated area management table 171 and the subsequent entry in the replication destination management table 172 are corrected. Specifically, when the redundant information storage position in the dedicated area management table 171 is “none”, this value sets the entry # registered in the replication destination management table 172. When the redundant information storage position in the dedicated area management table 171 is not “none”, the value set as the redundant information storage position of the entry storing the dedicated area information in the dedicated area management table 171 managed in the computer , The subsequent entry of the same entry as this value is referred to in the replication destination management table 172, and the entries in the replication destination management table are sequentially referred to until the subsequent entry becomes “none”. The value of the subsequent entry is corrected to the entry # registered in the replication destination management table 172. This completes the replication area determination process.

According to this embodiment, the configuration information of the dedicated area connected to the computer is set in the configuration information holding unit 170 and the configuration information holding unit 270 of the management server 200 in advance, and the configuration information of the unused replication area is set as the configuration information of the management server. By setting in the holding unit 270 and setting the redundancy of this area in the policy holding unit 160 of the computer, the user can set the duplication destination without setting the duplication destination computer-ID or storage device-ID. It becomes possible to decide.
Example 2 (related to claim 5)
A distributed data replication system according to another embodiment will be described with reference to FIG.
This distributed data replication system is configured by connecting computers 101 to 103 and a management server 200 to each other via a network 400. In the computers 101 to 103, the processor 110, the memory 120, the storage device control mechanism 130, the communication unit 150, the policy holding unit 160, and the configuration information holding unit 170 are coupled to the internal bus 180, and the storage device 140 is connected to the storage device control mechanism 130. Connected and configured. 101 to 103, 110 to 150, 170, 180 and 400 are the same as those in the first embodiment.

  The policy holding unit 160 holds the redundancy as in the first embodiment. In the second embodiment, information for selecting an appropriate duplication area among a plurality of duplication areas is held. The policy holding unit can be configured by, for example, a RAM or an electrically rewritable ROM.

  In response to a replication area determination request issued from the computers 101 to 103, the management server 200 determines a replication area corresponding to the dedicated area and notifies the request issuing computer of the replication area storage destination. The management server 200 includes a processor 210, a memory 220, a storage device control mechanism 230, a communication unit 250, a configuration information holding unit 270, a processing request control unit 300, an error information determination unit 310, a distance information determination unit 320, a performance determination unit 300, The comprehensive determination unit 340 is coupled to the internal bus 280, and the storage device 240 is connected to the storage device control mechanism 230. Compared to the first embodiment, it is different in that it includes an error information determination unit 310, a distance information determination unit 320, a performance determination unit 330, and an overall determination unit 340.

  Similar to the first embodiment, the configuration information holding unit 270 stores the configuration information of the dedicated area for storing the main data and the replication information for storing the replication information of the dedicated area in the storage device connected to the computer constituting the system. It holds area configuration information and unused replication area configuration information. In the second embodiment, error information and position information of a storage device connected to a computer in the system is further held. The management server 200 refers to these pieces of information in the replication area determination process and evaluates each replication area. The configuration information holding unit can be configured by, for example, a RAM or an electrically rewritable ROM.

  The error information determination unit 310 evaluates the copy area candidate sent from the processing request control unit 300 from the viewpoint of the number of error occurrences and notifies the comprehensive determination unit. If error information is not required for evaluation of a replication area candidate in this system, the error information determination unit 310 can be omitted. The error information determination unit 310 is configured by logic such as TTL or CMOS.

  The distance information determination unit 320 evaluates the replication area candidate sent from the processing request control unit 300 from the viewpoint of the physical distance to the selection request source computer and notifies the comprehensive determination unit. If no physical distance is required for evaluation of a replication area candidate in this system, the distance information determination unit 320 can be omitted. The distance information determination unit 320 is configured by logic such as TTL or CMOS.

The performance determination unit 330 evaluates the copy area candidate sent from the processing request control unit 300 from the viewpoint of access performance from the selection request source computer and notifies the comprehensive determination unit. In this system, when access performance is not necessary for evaluation of a replication area candidate, the performance determination unit 330 can be omitted. The performance determination unit 330 is configured by logic such as TTL or CMOS.
The overall determination unit 340 receives the results determined by the error information determination unit, the distance information determination unit, and the performance determination unit, and outputs the addition result of multiplication of the respective weight information as a comprehensive determination. The comprehensive determination unit 340 is configured by logic such as TTL or CMOS.

  The state of each storage device shown in FIG. 3 is the same as in the first embodiment. FIGS. 4, 5, and 6 show examples of information held in the configuration information holding unit 270 in the management server 200, which are the same as those in the first embodiment.

FIG. 7 shows an example of the error information table 274.
The error information table 274 holds the number of occurrences of errors in the storage devices in the system (denoted as the number of errors), presence / absence of blockage (denoted as blockage), storage device-ID, and chassis-ID. As the number of error occurrences, a cumulative total when the computer accesses the storage device and an error occurs is held. SMART management information may be used as the number of error occurrences. The presence / absence of blockage is stored as a flag indicating that the storage device has been disconnected from the system when no response is received even though the storage device is accessed a plurality of times. In the system, when error information is not required for evaluation of a replication area candidate, the error information table 274 can be omitted.

  FIG. 8 shows an example of the position information table 275 that holds the installation position information (latitude and longitude in this example) of the computers in the system and the device-ID. By managing the position information of each computer in the position information table 275, the distance between the computers can be calculated. If physical distance information is not required for evaluation of a replication area candidate in this system, the position information table 274 can be omitted.

  9, 10, 11, and 12 show examples of information held in the configuration information holding unit 170 in the computers 101 to 103. Each figure shows an example in the case of the computer 101. 9 and 10, the second embodiment is the same as the first embodiment.

FIG. 11 shows an example of the error information table 174.
The error information table 174 holds the number of error occurrences of the storage device in the computer 101 (denoted as the number of errors), presence / absence of blockage (denoted as blockage), storage device-ID, and device-ID. As the number of error occurrences, a cumulative total when the computer accesses the storage device and an error occurs is counted. SMART management information may be used for determining whether there is an error. The presence / absence of blockage is stored as a flag indicating that the storage device has been disconnected from the system when no response is received even though the storage device is accessed a plurality of times.

FIG. 12 shows an example of the policy holding unit 177.
The policy holding unit 177 holds a policy serving as a determination criterion for determining a replication area corresponding to a dedicated area in the computer 101. In the second embodiment, the redundancy, the storage device-ID, and the housing-ID are the same as those in the first embodiment. In the second embodiment, reference values for the number of errors, distance, and performance are held as evaluation criteria for determining a replication area, and weights for the respective evaluations are held.

  The number of errors indicates that a storage device whose number of errors managed in the error information table 274 is less than a reference value is preferentially selected. In the system, when the number of errors is not necessary as an evaluation criterion for the replication area, the evaluation criterion value for the number of errors can be omitted.

  The distance is a value for preferentially selecting a storage device whose physical distance between the replication area selection request source computer and the selected computer is within the reference value range or the reference value. When two types of values are set for the reference value, this means that storage devices located at a distance within the range of the reference value are preferentially selected, and when one type of value is set for the reference value, This means that storage devices that are at a distance greater than the reference value are preferentially selected. For example, in the distance reference value set in Example 177 of the policy-holding unit in computer-1, entry # 0 preferentially selects computers at a distance of 0 km or more and less than 100 km, and entry # 1 has a distance of 100 km or more. It shows that the computer in is preferentially elected. If the physical distance is not necessary as an evaluation criterion of the replication area in this system, the evaluation criterion value of the physical distance can be omitted.

  The performance is a value indicating that the storage device whose access time to the replication area selected from the replication area selection requesting computer is lower than the reference value is given priority. For example, in the performance reference value set in Example 177 of the computer-1 policy holding unit, entry # 0 indicates that a replication area whose access time is less than 20 seconds is preferentially selected. In this system, when performance is unnecessary as an evaluation criterion for a replication area, the evaluation criterion value for performance can be omitted.

  Each evaluation criterion includes weight information indicating an evaluation weight. It is possible to specify which evaluation item is prioritized by the weight information.

The replication area determination processing operation in the second embodiment will be described with reference to the flowchart of FIG. First, the policy holding unit 160 in the computer 101 is referenced to generate a replication area determination request (1000). In the case of the second embodiment, the replication area determination request includes the request source chassis-ID, storage device-ID, redundancy, error frequency reference value, distance reference value, performance reference value, and weight information of various evaluation criteria. included.
Steps 1010 to 1020 are the same as those in the first embodiment.

Next, a replication area corresponding to the dedicated area is selected (1030).
Referring to FIG. 16, the replication area selection processing operation in the second embodiment will be described in detail. The unused replication area management table 273 stored in the configuration information holding unit 270 is referenced to select one entry at a time (1200). . However, the replication area in the storage device connected to the management server is excluded.

  Next, the selected entry # and the error region reference value of the replication area determination request are sent to the error information determination unit 310 and a determination request is made (1210). As an example of determination by the error information determination unit 310, if the error count of the selected entry in the unused replication area management table 273 is 0 or more (error count reference value) / 5, it is 5 points (error count reference value). ) / 5 or more (error count reference value) x 2/5 if less than (error count reference value) x 2/5 or more (error count reference value) x 3/5 if less than 3 points, (error count) (Reference value) x 3/5 or more (error count reference value) x 2 points if less than 4/5, (error count reference value) x 4/5 or more (error count reference value) 1 point, (error count) If it is less than the reference value), it can be calculated as 0 points. If the number of errors in this system is not necessary as a criterion for evaluating the replication area, step 1210 can be omitted.

Next, the selected entry # and the distance reference value of the replication area determination request are sent to the distance information determination unit 320 and a determination request is made (1220). As an example of determination by the distance information determination unit 320, the following method may be mentioned. The distance between the replication area request source computer and the computer corresponding to the selected entry # is calculated, and it is determined whether or not the value is within the reference range. Specifically, the position information table 275 is referenced to obtain the latitude and longitude of the two types of computers. The latitude is α [deg], the longitude is β [deg], north latitude 0 [deg] to north latitude 90 [deg], 0 [deg] to 90 [deg], south latitude 0 [deg] to south latitude 90 [deg] − 0 [deg] to -90 [deg], east longitude 0 [deg] to east longitude 180 [deg] from 0 [deg] to 180 [deg], west longitude 0 [deg] to west longitude 180 [deg]-0 [deg] From -180 [deg], each latitude and longitude can be converted into an orthogonal coordinate system using Equation 1. Here, the x axis is regarded as a direction from negative to positive in the direction of 0 [deg] east longitude 0 [deg] from the point of north latitude 0 [deg] east longitude 180 [deg], and the y axis is 0 [deg] west longitude 90 Considering the direction from negative to positive in the direction of 0 [deg] east longitude 90 [deg] north from the point of [deg], the z-axis is 90 [deg] east longitude north of 90 [deg] east longitude 0 [deg] A direction from negative to positive is considered in the direction of 0 [deg]. Also, let the radius of the earth be r [km]. If the orthogonal coordinates of the two points calculated by Equation 1 are (x _A , y _A , z _A ) and (x _B , y _B , z _B ), the linear distance d between the two points can be calculated by Equation 2. From Equation 3, the distance s between two points on the spherical surface can be calculated using d. The calculated distance s is compared with the distance reference value of the replication area determination request, and 5 points can be calculated when the distance s is within the range of the reference value, and 0 points can be calculated otherwise. In this system, step 1220 can be omitted if distance is not necessary as a criterion for evaluating the replication area.

  Next, the selected entry # and the performance reference value of the replication area determination request are sent to the performance information determination unit 330, and a determination request is made (1230). As an example of determination by the performance information determination unit 330, the selected housing-ID, storage device-ID, and start position in the storage device are calculated, and a request is made for the replication area determination request source computer to access the storage device. The copy area determination request source computer measures the time required for access and sends the result back to the management server. If the time required for access is 0 or more (performance standard value) / 5, 5 points, (performance standard value) / 5 or more (performance standard value) × 2/5, 4 points, (performance standard value) X 2/5 or more (performance standard value) x 3 points if less than 3/5, (performance standard value) x 3/5 or more (performance standard value) x 2 points if less than 4/5, (performance standard value) It can be calculated as 1 point if it is 4/5 or more (performance standard value) and 0 point if it is (performance standard value) or more. In the system, step 1230 can be omitted when the performance is unnecessary as a criterion for evaluating the replication area.

  Next, a wait is inserted until the above three types of determination requests are completed. When the determination request is completed, the process proceeds to step 1250 (1240). Then, the processing request control unit 300 sends an error information determination result, a distance information determination result, a performance information determination result, and various weight information to the comprehensive determination unit 340, and requests a comprehensive determination (1250).

  Next, the overall determination unit 340 performs multiplication of determination result × weighting information for each error number, distance, and performance item, calculates the result of adding each product as an overall point, and the result is processed by the processing request control unit 300. Reply to (1260). Next, it is checked whether all entries in the unused replication area management table have been confirmed (1270). If the result of the check is incomplete, the processing after step 1200 is executed to determine the next entry. If completed, go to Step 1280.

As a result of the above check, if the processing is completed, the processing request control unit selects a replication area (redundancy −1) in descending order of the total score. However, the duplicate area in the same storage area as the dedicated area is excluded. If the necessary number of replication areas cannot be selected as a result of the exclusion, a replication area in the management server is selected (1280). Thus, the duplication area selection flow is completed.
Steps 1060 to 1070 are the same as those in the first embodiment.

As described above, according to the second embodiment, the configuration information of the dedicated area connected to the computer is set in advance in the configuration information holding unit of the computer and the configuration information holding unit of the management server, and the configuration information of the unused replication area is set to the management server. In the configuration information holding unit, the redundancy of this area and the error reference value, distance reference value, performance reference value and weight information of the determination element as the duplication area determination standard suitable for this area are stored in the policy holding part of the computer. If set, a duplication destination suitable for the dedicated area can be determined without having information on the duplication destination computer-ID and storage device-ID.
Example 3 (Related to claim 2)
In the second embodiment, when the evaluation item for selecting a replication area is only the number of errors, the number of errors, the performance, and the configuration used as the criteria for determining the replication area held by the components 310, 330, and 340 and the policy holding unit The error information table held in the information holding unit 270 and the steps 1220 and 1230 in the copy area selection flow can be omitted.
-Example 4 (Related to claim 3)
In the second embodiment, when the evaluation item for selecting the replication area is only the physical distance, the distance, performance, and configuration used as the criteria for determining the replication area held by the components 320, 330, and 340 and the policy holding unit The position information table held in the information holding unit 270 and the steps 1210 and 1230 in the copy area selection flow can be omitted.
-Example 5 (Related to claim 4)
In the second embodiment, when the evaluation item for selecting a replication area is only performance, the number of errors, distance, and configuration information used as the criteria for determining the replication area held by the components 310, 320, and 340 and the policy holding unit Steps 1210 and 1220 in the error information table, the position information table, and the copy area selection flow held by the holding unit 270 can be omitted.

  As described above, according to the present embodiment, selection of a storage device as a data copy destination can be automatically performed without a user's setting operation, and an increase in load on each storage device or computer can be minimized. Can be realized. Further, the optimum storage device can be determined by evaluating the copy destination storage device with respect to evaluation criteria set in advance from the viewpoint of the number of error occurrences, physical distance, and access time.

The block diagram which shows the distributed data replication system in one Example (Example 1), The block diagram which shows the distributed data replication system in another Example (Example 2), The figure which shows the example of area allocation in the memory | storage device 140 by one Example. The figure which shows the example of the exclusive area management table 271 managed with the management server by one Example. The figure which shows the example of the used replication area management table 272 managed by the management server by one Example. The figure which shows the example of the unused replication area management table 273 managed by the management server by one Example. The figure which shows the example of the error information table 274 managed by the management server by one Example. The figure which shows the example of the position information table 275 managed by the management server by one Example. The figure which shows the exclusive area management table 171 managed with the computer 101 in one Example. The figure which shows the example of the replication destination management table 172 managed with the computer 101 in one Example. The figure which shows the example of the error information table 174 managed by the computer 101 in one Example. The figure which shows the example of the content of the policy holding | maintenance part in the computer-1 in one Example. The figure which shows the comprehensive determination content of the replication area determination corresponding to the computer-1, HDD0, and Seg-ID1 in one Example. The flowchart figure which shows the processing operation of replication area decision in one execution example, The flowchart figure which shows the replication area selection processing operation in one execution example, The flowchart figure which shows the replication area selection processing operation in another Example (Example 2).

Explanation of symbols

101-103: Computer 110: Processor 120: Memory 130: Storage device control mechanism 140: Storage device
1401 to 1403: Storage device 150: Communication unit 160: Policy holding unit 170: Configuration information holding unit 171: Dedicated area management table 172: Replication destination management table 174: Error information table 177: Contents of policy holding unit 180: Internal bus 200 : Management server 210: Processor 220: Memory 230: Storage device control mechanism 240: Storage device 250: Communication unit 270: Configuration information holding unit 271: Dedicated area management table 272: Used replication area management table 273: Unused replication area management Table 274: Error information table 275: Position information table 300: Processing request control unit 310: Error information determination unit 320: Distance information determination unit 330: Performance determination unit 340: Comprehensive determination unit 341: Content of determination of replication area determination

Claims (5)

A distributed data replication system for storing data stored in a region of a plurality of storage devices used by a plurality of computers as a replica in a region of another storage device,
The storage device used by the computer includes a dedicated area and a replication area, and is configured to replicate data stored in the dedicated area to a replication area of the other storage device;
The computer includes a policy holding unit for holding redundancy that defines the total number of replication areas in other storage devices, the location of the dedicated area, the location of the replication area corresponding to the dedicated area, and the location of unused replication areas. A configuration information holding unit that defines
The management server that manages data replication has a processing request control unit that determines a replication area in another storage device in response to a replication area determination request from each of the computers or the storage devices. Distributed data replication system. The system according to claim 1, wherein the policy holding unit defines the number of error occurrences in the replication area as a selection criterion for determining a replication area suitable for the dedicated area in addition to the redundancy.
The management server further includes a configuration information holding unit that holds the number of error occurrences of each storage device, and an error that evaluates an unused replication area in terms of the number of error occurrences in order to determine a replication area corresponding to the dedicated area. A distributed data replication system comprising an information determination unit, wherein the processing request control unit determines a replication region suitable for the dedicated region with respect to the number of error occurrences. The system according to claim 1, wherein the policy holding unit defines a physical distance between the replication area and the dedicated area as a selection criterion for determining a replication area suitable for the dedicated area in addition to the redundancy.
The management server further defines the location of the dedicated area, the location of the replication area corresponding to the dedicated area, and the location of the unused replication area in the system, and holds the physical location of each computer or storage device A holding unit, and a distance information determination unit that evaluates an unused replication area with respect to a physical distance to determine a replication area corresponding to the dedicated area, and the processing request control unit has a physical distance in the dedicated area. A distributed data replication system characterized by determining a suitable replication area. The system according to claim 1, wherein the policy holding unit defines an access time to the replication area as a selection criterion for determining a replication area suitable for the dedicated area in addition to the redundancy.
The management server includes a performance determination unit that evaluates an unused replication area with respect to access time in order to determine a replication area corresponding to the dedicated area, and the processing request control unit has a replication suitable for the dedicated area with respect to access time. A distributed data replication system characterized by determining an area. The system according to claim 1, wherein the policy holding unit is configured to select the number of errors of the storage device, the duplication area, and the dedicated area as selection criteria for determining a duplication area suitable for the dedicated area in addition to the redundancy. Specify some of the physical distance and access time to the replication area, and specify some of the error count, physical distance, and access time evaluation weight,
The management server is not used to determine a replication area corresponding to the dedicated information area, and a configuration information holding unit that holds some of the number of error occurrences of each storage device, each computer or physical location of the storage device An error information determination unit that evaluates a replication area in terms of the number of error occurrences, a distance information determination unit that evaluates an unused replication area in terms of physical distance to determine a replication area corresponding to the dedicated area, A performance determination unit that evaluates an unused replication region in terms of access time to determine a replication region corresponding to a dedicated region, and some of the error information determination unit, the distance information determination unit, and the performance determination unit A comprehensive determination unit for comprehensively evaluating the determination results with weights;
The processing request control unit determines a replication region suitable for the dedicated region with respect to the number of error occurrences, physical distance, or access time as a result of determination by the comprehensive determination unit.

Images