JPWO2004027625A1 - Storage control device, storage control program, and storage control method - Google Patents

Abstract

A storage control device that is distributed over a network together with a cache device and a disk device to form a storage system, accepts access requests from clients, and causes the cache control device and real disk control device to execute processing according to the access request. Thus, neither the data write processing unit for controlling the data requested by the client to write to the cache memory of the two cache control devices nor the data requested by the client for the two cache control devices. A data read processing unit that reads data from the disk control device and controls the read data to be stored in only one cache control device.

Description

  The present invention is used in a storage system in which a cache device and a disk device are distributed on a network, receives a request for access from a client, and causes the cache device and the disk device to execute processing according to the access request. The present invention relates to a control program and a storage control method, and in particular, a storage control device, a storage control program, and a storage control capable of preventing the entire storage system from being stopped due to a failure of an individual device and constructing a reliable storage system It is about the method.

In recent years, with the rapid increase in the amount of processing data due to the use of multimedia data and the like, storage systems that integrate and manage large-scale data using a large-capacity disk connected to a network are rapidly spreading. In this storage system, the disk access speed is slower than the network data transfer speed, so the disk access performance is the bottleneck of the entire system. For this reason, the storage system uses a cache memory that can be accessed at high speed, and reduces the number of accesses to the disk, thereby improving the access performance of the disk.
Normally, when a cache memory is used, the cache hit rate, which indicates the rate at which requested data is present in the cache memory, greatly affects the system performance. Therefore, it is necessary to use as large a cache memory as possible. However, in the conventional storage device, since the entire storage system control, disk control, cache memory control, etc. are realized as one device, it is not easy to expand the capacity of the cache memory. When the capacity of the disk is expanded by the increase, the relative capacity of the cache memory is reduced, and the cache hit rate is reduced.
For this reason, in order to facilitate the expansion of the cache memory capacity, the entire storage system control, disk control, and cache memory control are implemented as separate devices, and these devices are connected to the network. Distributed storage systems that have been distributed in a distributed manner have come to be used. In this distributed storage system, since only the cache memory and its control are independent cache devices, the capacity of the cache memory can be easily expanded by adding a cache device.
However, this distributed storage system has a problem in that each device fails separately because each device is arranged independently. Further, since each device is independently connected to the network, there is a problem that a failure occurs in the individual network connecting each device, and the individual device cannot be used.
That is, in the conventional integrated storage apparatus, a failure occurs in the entire storage apparatus. Therefore, it is sufficient to perform RAID (Redundant Array of Independent Disks) control considering the failure of the entire storage apparatus. In the system, it is necessary to consider a failure for each individual device arranged in a distributed manner.
The present invention has been made to solve the above-described problems caused by the prior art, and prevents the entire storage system from being stopped due to a failure of an individual device, thereby constructing a highly reliable storage system. An object of the present invention is to provide a storage control device, a storage control program, and a storage control method.

In order to solve the above-described problems and achieve the object, the present invention forms a storage system in a distributed arrangement on a network together with a cache device and a disk device, receives an access request from a client, and receives the cache device and the disk device. A storage control device that executes processing according to an access request, wherein the data requested by the client to be written is multiplexed and written to a predetermined number of cache devices among the cache devices arranged in the network. The write control means for controlling and the data requested to be read by the client when none of the cache devices arranged in the network has the cache device, the data is read from the disk device, and the read data A certain Characterized by comprising a reading control means for controlling to transmit to said client causes only the storage shoe device.
In addition, the present invention is used in a storage system in which a cache device and a disk device are distributed on a network, and receives storage access requests from a client and causes the cache devices and disk devices to execute processing according to the access request. A write control procedure for controlling to multiplex and write the data requested by the client to be written to a predetermined number of cache devices arranged in the network; When none of the cache devices arranged in the network has the requested data, the data is read from the disk device, and the read data is stored only in a certain cache device and And executes a read control step of controlling to transmit to the client, with the computer.
In addition, the present invention is used in a storage system in which a cache device and a disk device are distributed on a network, and receives storage access requests from a client and causes the cache devices and disk devices to execute processing according to the access request. A write control step for controlling to multiplex and write data requested by the client for writing to a predetermined number of cache devices arranged in the network; When none of the cache devices arranged in the network has the requested data, the data is read from the disk device, the read data is stored only in a certain cache device, and the cluster is stored. Characterized in that it includes a reading control step of controlling to transmit the Ant, the.
According to this invention, control is performed so that the data requested by the client to be written is multiplexed and written in a predetermined number of cache devices arranged in the network, and the data requested by the client is read to the network. When there is no cache device among the arranged cache devices, the data is read from the disk device, and the read data is stored only in a certain cache device and controlled to be transmitted to the client. While using the cache memory efficiently, it is possible to prevent the entire storage system from being stopped due to a failure of one of the cache devices and to construct a highly reliable storage system.

  FIG. 1 is a block diagram showing the system configuration of the storage system according to the first embodiment, and FIG. 2 is a functional block diagram showing the configuration of the storage control device shown in FIG. FIG. 4 is a diagram showing an example of the location information table, FIG. 4 is a functional block diagram showing the configuration of the cache control device shown in FIG. 1, and FIG. 5 is a message transmitted by the storage control device. FIG. 6 is a diagram illustrating an example of a message transmitted by the cache control device, and FIG. 7 is a diagram illustrating an example of a message transmitted by the real disk control device. FIG. 8 is a flowchart showing a processing procedure of the data write processing unit shown in FIG. 2, and FIG. 9 explains a method in which the client determines completion of data writing to the storage system. FIG. 10 is a flowchart showing a data write processing procedure to the cache memory according to the first embodiment, and FIG. 11 is a process of the data read processing unit shown in FIG. FIG. 12 is a flowchart showing a processing procedure of the Fetch processing unit shown in FIG. 2, and FIG. 13 is a flowchart showing a processing procedure of the Flush processing unit shown in FIG. FIG. 14 is an explanatory diagram for explaining the concept of duplication of the storage control apparatus according to the second embodiment, and FIG. 15 shows the system configuration of the storage system according to the second embodiment. FIG. 16 is a diagram showing an example of the location information update message, and FIG. 17 shows the processing procedure of the location information table update unit shown in FIG. FIG. 18 is an explanatory diagram for explaining a modification of duplication of the storage control device according to the second embodiment, and FIG. 19 is a diagram of the storage system according to the third embodiment. FIG. 20 is a block diagram illustrating a system configuration, and FIG. 20 is a diagram illustrating an example of a computer system that executes a storage control program and a cache control program.

Exemplary embodiments of a storage control device, a storage control program, and a storage control method according to the present invention will be explained below in detail with reference to the accompanying drawings.
Embodiment 1 FIG.
First, the system configuration of the storage system according to the first embodiment will be described. FIG. 1 is a block diagram showing a system configuration of the storage system according to the first embodiment. As shown in the figure, this storage system is configured with a storage control device 110, a real disk control device 120, and duplicated cache control devices 130 and 140 distributed on a network 40, and a client via the network 40. 10 to 30 are used. Although only three clients 10 to 30 are shown here for convenience of explanation, this storage system is used by an arbitrary number of clients via the network 40. Although the case where the client 10 uses this storage system will be described here, the other clients 20 and 30 can also use this storage system.
The storage control device 110 is a device that accepts an access request from the client 10 via the network 40 and causes the real disk control device 120 and the cache control devices 130 and 140 to execute processing according to the access request. The storage control device 110 manages data stored in the cache memory included in the cache control devices 130 and 140.
The real disk control device 120 has a disk 121 as a large-capacity storage medium, and reads data from the disk 121 and writes data to the disk 121 based on instructions from the storage control device 110. Although only one real disk control device 120 is shown here for convenience of explanation, in this storage system, a plurality of real disk control devices are arranged on the network 40 according to the scale of data to be stored. Can do. A plurality of disks can be connected to one real disk control device.
The cache control devices 130 and 140 have a cache memory as a high-speed access recording medium, and read data from the cache memory and write data to the cache memory based on instructions from the storage control device 110.
Thus, in this storage system, the cache control device is duplicated and the same data is stored in the cache memory of the cache control devices 130 and 140. However, if not all data is duplicated, but the data stored in the disk 121 matches the data stored in the cache memory, that is, if the data is in a coherent state, the data is either If the data stored only in the cache memory and the data stored in the disk 121 and the data stored in the cache memory are different, that is, if the data is in a dirty state, the data is stored in both cache memories.
For example, in FIG. 1, the data M is data in which the data M ′ on the disk 121 has been rewritten and is in a dirty state, so it is stored in two cache memories of the cache control device 130 and the cache control device 140. Has been. The data N is stored only in the cache memory of the cache control device 130 because the same data exists on the disk 121.
In this way, in this storage system, the cache control device is duplicated and the same data is stored in two cache memories. Therefore, even if one of the cache control devices 130 or 140 fails, the other data The storage system can be operated using the cache control device, and a highly reliable storage system can be realized. Further, in this storage system, when the data stored in the disk 121 and the cache memory are the same, the data is stored only in one cache memory, so that the cache memory can be used effectively. The cache hit rate can be improved.
Next, the configuration of the storage control device 110 shown in FIG. 1 will be described. FIG. 2 is a functional block diagram showing the configuration of the storage control device 110 shown in FIG. As shown in the figure, the storage control device 110 includes a location information table 111, a control unit 112, a network interface unit 113, a data read processing unit 114, a data write processing unit 115, and a fetch processing unit 116. , And a flush processing unit 117.
The location information table 111 includes a logical location designated by the client 10 for data that the client 10 requests access to the storage system, a physical location of the disk 121 where the data is actually stored, a physical location of the cache memory, and the data. (Ie, whether the data is coherent or dirty) is stored in association with each other.
FIG. 3 is a diagram illustrating an example of the position information table 111. In the figure, data whose logical positions are 0 to 7 of the storage with the device number 0 are the 0 to 7 of the disk 121 with the device number 0 and 0 to 0 of the cache memory of the cache control device 130 with the device number 0. 7 indicates a coherent state. Further, the data whose logical positions are 8 to 15 of the storage of the device number 1 are the 8 to 15 of the disk 121 of the device number 0, the cache memories 8 to 15 of the cache control device 130 of the device number 0, and It is stored in 0 to 7 of the cache memory of the cache control device 140 having the device number 1, indicating that it is in a dirty state.
In this way, the location information table 111 stores the physical position of the data stored in the cache memory included in the two cache control devices 130 and 140 in correspondence with the logical position of the data, so that the storage control device 110 can perform processing corresponding to the duplication of the cache control device.
The control unit 112 is a processing unit that controls the entire storage control device 110, and the network interface unit 113 communicates with the client 10, the real disk control device 120, and the cache control devices 130 and 140 using a predetermined communication protocol. It is a processing unit that performs.
The data read processing unit 114 is a processing unit that reads data from the storage system in response to a data read request from the client 10, obtains a physical position from the logical position of the data using the position information table 111, and stores the data If it is stored in the cache memory, it instructs the cache control device having the cache memory storing the data to transfer data to the client 10. On the other hand, if the data is not stored in the cache memory, the Fetch processing unit 116 is requested to transfer data from the real disk control device 120 to the cache control device. Then, after the data is transferred from the real disk control device 120 to the cache control device, the cache control device to which the data has been transferred is instructed to transfer data to the client 10.
The data write processing unit 115 is a processing unit that writes data to the storage system in response to a data write request from the client 10, and is always provided to both cache control devices 130 and 140 based on the location information table 111. To write data to the cache memory. That is, when the data storage position in the cache memory is registered in the position information table 111, the data write instruction is issued to the storage position, and the data storage position in the cache memory is registered in the position information table 111. If not, a new cache memory is allocated and data writing is instructed to the allocated cache memory position.
As described above, since the data write processing unit 115 writes the same data in both cache memories in response to the data write request from the client 10, even when one of the cache control devices fails, The operation stop of the storage system can be prevented, and the reliability of the storage system can be improved.
When the data requested to be read from the client 10 is not stored in the cache memory, the fetch processing unit 116 selects a cache control device that newly stores the data and selects the real disk control device 120. A processing unit for instructing transfer of data to the cache control device.
The flush processing unit 117 is a processing unit that writes dirty data out of data stored in the cache memory to the disk 121. When data is written to the cache memory, an empty area is insufficient, or the cache It is used to create an empty area in the cache memory when the used area of the memory exceeds a predetermined ratio. When creating an empty area in the cache memory, selection of which data is deleted from the cache memory is determined using a method such as LRU (Least Recently Used), and the data determined to be deleted from the cache memory This flush processing unit 117 is used when the state is dirty.
Next, the configuration of the cache control devices 130 and 140 shown in FIG. 1 will be described. Since the cache control devices 130 and 140 have the same configuration, the cache control device 130 will be described as an example here. FIG. 4 is a functional block diagram showing the configuration of the cache control device 130 shown in FIG. As shown in the figure, the cache control device 130 includes a cache memory 131, a control unit 132, a network interface unit 133, a cache read processing unit 134, a cache write processing unit 135, a fetch correspondence unit 136, And a flush corresponding part 137.
The cache memory 131 is a storage unit that stores a part of data stored in the disk 121 and data stored in the disk 121 in order to speed up access to the storage system. It is characterized by short access time and small capacity. The data stored in the cache memory 131 and its storage location are stored in the location information table 111 of the storage control device 110.
The control unit 132 is a processing unit that controls the entire cache control device 130, and the network interface unit 133 communicates with the client 10, the storage control device 110, and the real disk control device 120 using a predetermined communication protocol. It is a processing unit.
The cache read processing unit 134 is a processing unit that reads data from the cache memory 131 based on a data read instruction from the storage control device 110 and transfers the read data to the client 10.
The cache write processing unit 135 is a processing unit that reads data from the buffer of the client 10 based on a data write instruction from the storage control device 110 and writes the read data to the cache memory 131. In addition, when the writing of data to the cache memory 131 is completed, the cache writing processing unit 135 notifies the client 10 of the completion of data writing. This completion notification corresponds to the duplication of the cache control device, and details thereof will be described later.
The fetch correspondence unit 136 is a processing unit that writes data to the cache memory 131 that is transferred from the real disk control device 120 to the cache control device 130 based on an instruction from the fetch processing unit 116 of the storage control device 110.
The flush corresponding unit 137 is a processing unit that reads data from the cache memory 131 based on an instruction from the flush processing unit 117 of the storage control device 110 and transfers the data to the real disk control device 120.
Next, a message that the storage control device 110 transmits to instruct the cache control devices 130 and 140 and the real disk control device 120 to perform processing will be described. FIG. 5 is a diagram showing an example of a message transmitted by the storage control device 110. 2A shows a data access instruction message to the cache memory transmitted from the storage control device 110 to the cache control devices 130 and 140. FIG. 2B shows the storage control device 110 informing the real disk control device 120. A data transfer instruction message to be transmitted is shown.
As shown in FIG. 6A, the data access instruction message 510 to the cache memory transmitted from the storage controller 110 to the cache controllers 130 and 140 has an internal header 511, a response header 512, and a data location 513. .
The internal header 511 includes an address of the client 10 or the real disk control device 120 to which the cache control devices 130 and 140 perform data transmission / reception, a data read instruction (“read”) from the cache memory, or data to the cache memory A command type indicating whether the instruction is a write instruction (“write”) is included.
That is, when the data access instruction message 510 is an instruction to write data to the cache memory, the address is the address of the client 10 that has made a data write request, and the command type is “write”. When the data access instruction message 510 is an instruction to read data from the cache memory, the address is the address of the client 10 that has made a data read request, and the command type is “read”. In addition, when the data access instruction message 510 is an instruction to transfer data from the cache memory to the disk 121, the address is the address of the real disk controller 120 as the data transfer destination, and the command type is “read”.
The response header 512 is a header format for transmitting the result of execution of the processing instructed by the storage control device 110 by the cache control devices 130 and 140 to the client 10 or the storage control device 110. The response header 512 includes the response header 512. A command ID for identifying a request from 10 and an execution result indicating whether the execution has succeeded or failed are included. The data included in the response header 512 of the data access instruction message 510 instructing data writing to the cache memory will be described later.
The data location 513 includes the physical location of the disk 121 storing data and the cache memory. However, if the data is not stored in the cache memory, only the physical location of the disk 121 is included.
As shown in FIG. 5B, the data transfer instruction message 520 transmitted from the storage controller 110 to the real disk controller 120 also has an internal header 521, a response header 522, and a data location 523.
The internal header 521 includes the address of the cache control device 130 or 140 that is the data transfer destination and “read” as the command type. The response header 522 is a header format for the real disk control device 120 to transmit the data transfer result (success or failure) to the storage control device 110. The data location 523 includes a physical location that specifies the location of the data on the disk 121 that stores the data to be transferred, and a data storage location in the cache memory that is the data write destination.
Next, messages that the cache control devices 130 and 140 transmit to the client 10, the storage control device 110, and the real disk control device 120 in response to instructions from the storage control device 110 will be described. FIG. 6 is a diagram illustrating an example of a message transmitted by the cache control devices 130 and 140. FIG. 11A shows a response message 610 for transmitting the result of executing the data read process from the cache memory or the data write process to the cache memory based on the request of the client 10 to the client 10. b) shows a data transfer message 620 for transferring the data in the cache memory to the real disk control device 120 based on the flush processing, and FIG. 10C shows the case where the data transfer to the real disk control device 120 is completed. Shows a data transfer completion message to be transmitted to the storage controller 110.
As shown in FIG. 5A, the response message 610 transmitted from the cache control devices 130 and 140 to the client 10 includes a response header 611 and response data 612. The response header 611 includes a command ID for identifying a request from the client 10 and an execution result of the instructed process. When the request from the client 10 is data read, the data read from the cache memory becomes the response data 612. On the other hand, when the request from the client is data writing, there is no response data 612.
As shown in FIG. 5B, the data transfer message 620 transmitted from the cache control devices 130 and 140 to the real disk control device 120 includes a response header 621, a data location 622, and response data 623. The response header 621 includes a command ID for identifying that the data transfer message 620 is data transfer based on the flush process, and the data position 622 specifies the write position on the disk 121 of the data to be transmitted. The response data 623 is flush data that is read from the cache memory and stored in the disk 121.
Further, as shown in FIG. 5C, the data transfer completion message 630 transmitted from the cache control apparatuses 130 and 140 to the storage control apparatus 110 has a response header 631. The response header 631 includes the command ID included in the instruction message from the storage controller 110 corresponding to this response and the execution result of the instructed process.
Next, a message transmitted from the real disk control device 120 to the cache control devices 130 and 140 and the storage control device 110 in response to an instruction from the storage control device 110 will be described. FIG. 7 is a diagram showing an example of a message transmitted by the real disk control device 120. FIG. 11A shows a data transfer message 710 in which the real disk control device 120 transfers the data stored in the disk 121 to the cache control device 130 or 140 based on the Fetch process, and FIG. A data transfer completion message 720 transmitted from the real disk control device 120 to the storage control device 110 when the data transfer to the control device 130 or 140 is completed is shown.
As shown in FIG. 5A, the data transfer message 710 transmitted from the real disk control device 120 to the cache control device 130 or 140 has a response header 711, a data location 712, and response data 713. The response header 711 includes a command ID for identifying that the data transfer message 710 is data transfer based on the Fetch process, and the data position 712 specifies the write position of the data to be transferred in the cache memory. The response data 713 is fetch data that is read from the disk 121 and written to the cache memory.
Further, as shown in FIG. 5B, the data transfer completion message 720 transmitted to the storage control device 110 has a response header 721. The response header 721 includes the command ID included in the instruction message from the storage controller 110 corresponding to this response and the execution result of the instructed process.
Next, the processing of the data write processing unit 115 of the storage control device 110 shown in FIG. 2 will be described. FIG. 8 is a flowchart showing a processing procedure of the data write processing unit 115 of the storage control apparatus 110 shown in FIG.
As shown in the figure, the data write processing unit 115 searches the position information table 111 using the logical position of the write data designated by the client 10 (step S801), and there is data in the cache memory, and the state is dirty. Whether there is data in the cache memory and the state is coherent, or there is no data in the cache memory (step S802).
If there is data in both cache memories and the data state is dirty, a data access instruction message 510 is transmitted with a command type of “write” to both cache control devices 130 and 140. Then, an instruction to write data designated by the client 10 to the cache memory is issued (step S803).
If there is data in one of the cache memories and the state of the data is coherent, an area for storing the data is secured in the cache memory without data, and both the cache control devices 130 and 140 are provided. On the other hand, the data access instruction message 510 is transmitted with the command type as “write” (step S806). Then, the location information table 111 is updated for the newly secured cache memory area (step S807).
If there is no data in both cache memories, an area for storing data in both cache memories is secured (step S805), and the command type for both cache control devices 130 and 140 is set to “ The data access instruction message 510 is transmitted as “write” (step S806). Then, the location information table 111 is updated for the newly secured cache memory area (step S807).
As described above, since the data write processing unit 115 always writes the data requested by the client 10 to both cache memories, even if one of the cache control devices fails, the data is lost. Can be prevented.
Next, a method in which the client 10 determines completion of data writing to the storage system will be described. In this storage system, two cache control devices 130 and 140 are used, and it is necessary to write data to the cache memory included in the two cache control devices 130 and 140. Therefore, the completion of data writing is the time when both cache control devices 130 and 140 have completed writing data to the cache memory. FIG. 9 is an explanatory diagram for explaining a method in which the client 10 determines completion of data writing to the storage system.
As shown in the figure, when the storage controller 110 receives a data write request from the client 10 (command ID is “command 1”), the storage controller 110 instructs both cache controllers 130 and 140 to access data. A message 510 is transmitted. In this data access instruction message 510, the internal header 511 includes the address of the client 10 that is the data write request source and “write” that is the command type, and the response header 512 includes two responses to “command 1”. The data position 513 includes data position P and data position Q which are data write positions in the cache memory.
The cache controllers 130 and 140 having received the data write message 510 read the data from the buffer of the client 10, the cache controller 130 is in the data position P of the cache memory, and the cache controller 140 is in the data position Q of the cache memory. Write data. When the data writing is completed, a response message 610 created based on the response header 512 is transmitted to the client 10. This response message 610 includes data (two responses) specifying that there are two responses to “command 1”.
The client 10 that has received the response message 610 from the cache control device 130 or 140 confirms that there are two responses to “command 1”, waits for a response message 610 from another cache control device, When the response message 610 from the cache control device is received, it is determined that the data writing to the storage system is completed.
In this way, the storage controller 110 uses the response header 512 to specify that there are two response messages 610 for the data write request, and the client 10 counts the number of response messages 610 from the cache controller, Since it is determined that the writing of data to the storage system is completed only when two response messages 610 are received, the same data is written to the cache memories of the two cache control devices 130 and 140. Can be ensured.
Next, a procedure for writing data into the cache memory according to the first embodiment will be described. FIG. 10 is a flowchart showing a procedure for writing data to the cache memory according to the first embodiment. Here, it is assumed that the cache control device 130 completes writing of data to the cache memory before the cache control device 140.
As shown in the figure, when the cache control device 130 receives the data access instruction message 510 transmitted from the storage control device 110 (step S1001), the cache control device 130 reads data from the buffer of the client 10 (step S1002), and reads the read data. Write to cache memory. A data write position in the cache memory is designated by a data position 513 included in the data access instruction message 510. When the writing of data to the cache memory is completed, a response message 610 is transmitted to the client 10 (step S1003).
On the other hand, the other cache control devices 140 similarly receive the data access instruction message 510 transmitted by the storage control device 110 (step S1005), read the data from the buffer of the client 10 (step S1006), and cache the read data. Write to memory. When the writing of data to the cache memory is completed, a response message 610 is transmitted to the client 10 (step S1007).
The client 10 first receives the response message 610 transmitted by the cache control device 130 (step S1004), recognizes that it is necessary to decode the response header 611 and receive two response messages 610, and then The response message 610 is waited for. When the response message 610 from the cache control device 140 is received (step S1008), it is determined that the data write processing is completed in both the cache control devices 130 and 140.
In this example, when the cache control devices 130 and 140 complete the writing of data to the cache memory, the response message 610 is transmitted to the client 10. However, the cache control devices 130 and 140 write to the cache memory. Can be notified to the storage controller 110 instead of the client 10 when the data write is completed. In this case, since the storage control device 110 issues a data write instruction to the two cache control devices 130 and 140, the client control unit 110 receives the data write completion notification from the cache control devices 130 and 140 that issued the instruction, and then 10 can be notified of the completion of data writing.
Next, the processing procedure of the data read processing unit 114 of the storage control device 110 shown in FIG. 2 will be described. FIG. 11 is a flowchart showing a processing procedure of the data read processing unit 114 of the storage control device 110 shown in FIG.
As shown in the figure, the data read processing unit 114 uses the position information table 111 to check whether or not the data requested by the client 10 to be read exists in the 0th cache control device 130 (step S1101). . If there is data in the 0th cache control device 130, a data access instruction message 510 is transmitted to the 0th cache control device 130 with the command type “read” (step S1102), and the cache memory Is instructed to transmit data stored in the client 10 to the client 10.
On the other hand, if the data requested by the client 10 to be read is not in the number 0 cache control device 130, it is checked whether or not the data is in the number 1 cache control device 140 (step S1103). Further, even when all the data requested by the client 10 to be read out are not in the 0th cache control device 130, it is checked whether or not the remaining data is in the 1st cache control device 140 (step S1103). . When there is data to be requested from the first cache control device 140, the data access instruction message 510 is transmitted to the first cache control device 140 with the command type “read” (step S1104). An instruction to transmit the data stored in the memory to the client 10 is given.
On the other hand, if the data requested by the client 10 is not in the first cache control device 140, it is checked whether or not the data is in the disk 121 (step S1105). Further, even if the 0th and 1st cache control devices 140 do not have all the data that the client 10 requested to read, it is checked whether or not the remaining data exists in the disk 121 (step S1105). If the data to be sought is on the disk 121, the data is read from the disk 121 using the Fetch process and written to the cache memory (step S1106). Then, a data access instruction message 510 is transmitted with the command type “read” to the cache control apparatus having the cache memory in which the data has been written (step S1107), and the data stored in the cache memory is transmitted to the client 10. Instruct.
As described above, the read processing unit 114 checks the storage location of the data requested to be read by the client 10 using the location information table 111, and the data is in one of the two cache control devices 130 and 140. In such a case, by issuing an instruction to transmit the data to the client 10, data reading corresponding to the duplication of the cache is performed.
Next, the processing procedure of the fetch processing unit 116 of the storage control device 110 shown in FIG. 2 will be described. Note that the process of the fetch processing unit 116 is a process corresponding to the fetch process (step S1106) shown in FIG. FIG. 12 is a flowchart showing a processing procedure of the fetch processing unit 116 of the storage control apparatus 110 shown in FIG.
As shown in the figure, the fetch processing unit 116 secures a cache memory area for storing data using the position information table 111 (step S1201), and uses the data transfer instruction message 520 to secure the secured cache memory area. The real disk controller 120 is instructed to transfer data to the cache controller having the area (step S1202). Then, it waits for the completion of data transfer from the real disk control device 120 to the cache control device (step S1203), and receives a data transfer completion message 720 from the real disk control device 120 (step S1204). The position information table 111 is updated for (Step S1205).
Next, the processing procedure of the flash processing unit 117 of the storage control apparatus 110 shown in FIG. 2 will be described. FIG. 13 is a flowchart showing the processing procedure of the flash processing unit 117 of the storage control apparatus 110 shown in FIG.
As shown in the figure, the flush processing unit 117 transmits a message instructing data transfer to the real disk control device 120 to the cache control device storing the data to be transferred to the disk 121. That is, the address of the internal header 511 is used as the address of the real disk control device 120, the command type is “read”, the data access instruction message 510 is transmitted, and data transfer to the real disk control device 120 is instructed (step S1301). .
Then, it waits for the completion of data transfer from the cache control device to the real disk control device 120 (step S1302). When the data transfer completion message 630 is received from the cache control device (step S1303), the cache storing the transferred data is stored. The memory area is released, and the position information table 111 is updated for the released cache memory area (step S1304).
As described above, in the first embodiment, two cache control devices 130 and 140 are provided in the storage system, and the same data is stored in both cache control devices 130 and 140. Even when the cache control device fails, the storage system can be operated using another cache control device, and a highly reliable storage system can be realized.
When the data stored in the disk 121 and the data stored in the cache control devices 130 and 140 are the same (coherent), the cache memory of one of the cache control devices is released and stored in the disk 121. Since the same data is stored in both the cache control devices 130 and 140 only when the stored data and the data stored in the cache control device are different (dirty), the cache memory can be used efficiently, The cache hit rate can be improved.
Embodiment 2. FIG.
In the first embodiment, the case where the reliability of the storage system is improved by duplicating the cache control apparatus has been described. However, in addition to the duplication of the cache control apparatus, the storage control apparatus is also duplexed. Thus, the reliability of the storage system can be further improved. Therefore, in the second embodiment, a storage system in which the storage control device is duplicated in addition to the duplication of the cache control device will be described.
First, the concept of duplication of the storage control device according to the second embodiment will be described. FIG. 14 is an explanatory diagram for explaining the concept of duplication of the storage control apparatus according to the second embodiment. As shown in the figure, this storage system has two storage control devices, a main storage control device and a backup storage control device.
In normal operation, the main storage control device of these two storage control devices functions as a storage control device of the storage system. However, both the main storage control device and the backup storage control device have a location information table, and the main storage control device sends the update information to the backup storage control device every time the location information table is updated. . Then, the backup storage control device that has received this update information updates its own location information table and always keeps the location information table up-to-date.
As described above, in the storage system according to the second embodiment, the main storage control device is always kept in the same state by keeping the position information table of the main storage control device and the position information table of the backup storage control device. Even if a failure occurs, the storage system can be operated using the backup storage control device, and a highly reliable storage system can be realized.
Next, the system configuration of the storage system according to the second embodiment will be described. FIG. 15 is a block diagram showing a system configuration of the storage system according to the second embodiment. Here, for convenience of explanation, functional units that play the same functions as the respective units shown in FIG. 1 are given the same reference numerals, and detailed descriptions thereof are omitted.
As shown in the figure, in this storage system, a main storage control device 1510, a backup storage control device 1520, a real disk control device 120, and two cache control devices 130 and 140 are distributed on a network 40. Be placed.
In the normal operation, the main storage controller 1510 receives an access request from the client 10 via the network 40, and performs processing according to the access request to the real disk controller 120 and the cache controllers 130 and 140. It is a device to be executed. The main storage control device 1510 includes a location information table 1511 and an update information transmission unit 1512.
The location information table 1511 includes the logical location specified by the client 10 for the data that the client 10 requests to access the storage system, the physical location of the disk 121 where the data is actually stored, the physical location of the cache memory, and the data. In other words, the table is stored in association with the distinction of whether the data is coherent or dirty.
The update information transmission unit 1512 is a processing unit that transmits update information to the backup storage control device 1520 each time the location information table 1511 is updated. The location information table 1511 is updated when the data stored in the cache memory by the flush process becomes unnecessary, when a new cache memory is allocated by the fetch process, and when a new cache memory is allocated by the data write process. This is done in some cases.
The backup storage control device 1520 is a device that takes over the processing when the main storage control device 1510 fails, and includes a location information table 1521 and a location information table update unit 1522.
The position information table 1521 is a table that stores the same data as the position information table 1511 included in the main storage control device 1510. The location information table 1521 stores the same data as the location information table 1511, so that when the main storage control device 1510 fails, the backup storage control device 1520 can take over and execute the processing. Become.
The location information table update unit 1522 is a processing unit that receives the update information of the location information table 1511 transmitted by the update information transmission unit 1512 of the main storage control device 1510 and updates the location information table 1521 of the backup storage control device 1520. is there. The location information table update unit 1522 updates the location information table 1521 so that the location information table 1521 can always be kept up-to-date, and when the main storage control device 1510 fails, the processing is performed in the backup system. The storage control device 1520 can take over and execute.
Next, a location information update message that the update information transmission unit 1512 of the main storage control device 1510 transmits to the backup storage control device 1520 as update information will be described. FIG. 16 is a diagram showing an example of the location information update message. As shown in the figure, this location information update message 1610 has a type 1611, logical location information 1612, physical location information 1613, and a status 1614.
The type 1611 indicates the type of update of the position information table, and is either a new allocation (alloc) of the cache memory, a partial release of the redundant cache memory (free), or a full release of the cache memory (free). .
The logical position information 1612, the physical position information 1613, and the state 1614 are information corresponding to the logical position, the physical position, and the state of the position information table, respectively. In other words, the logical position information 1612 is information regarding a logical position that is a data storage position used by the client 10 to access the storage system, and the physical position information 1613 is a data storage position in the disk 121 and the cache memory for storing data. The state 1614 is information indicating whether the data stored in the disk 121 is the same as the data stored in the cache memory.
Next, the processing procedure of the location information table update unit 1522 of the backup storage control device 1520 shown in FIG. 15 will be described. FIG. 17 is a flowchart showing a processing procedure of the location information table update unit 1522 of the backup storage control device 1520 shown in FIG.
As shown in the figure, the location information table update unit 1522 receives a location information update message 1610 transmitted from the update information transmission unit 1512 of the main storage control device 1510 to the backup storage control device 1520 via the network 40. (Step S1701).
Then, the type 1611 included in the location information update message 1610 is checked (step S1702). If the type 1611 is “free”, the information corresponding to the released cache memory area is deleted from the location information table 1521. (Step S1703) If the type 1611 is “free”, the information in the location information table 1521 corresponding to the partially released cache memory area is updated (Step S1704). Also, if the type 1611 of the location information update message 1610 is “alloc”, if the location information table 1521 has information corresponding to the newly allocated cache memory area, the information is updated, and the new If there is no information corresponding to the cache memory area allocated to the location information table 1521, new correspondence information is created and added to the location information table 1521 (step S1705).
As described above, in the second embodiment, the storage control device is duplicated using the main storage control device 1510 and the backup storage control device 1520, and the location information tables 1511 and 1521 are stored in both storage control devices 1510 and 1520. The update information transmission unit 1512 of the main storage control device 1510 transmits a location information update message 1610 to the backup storage control device 1520 each time the location information table 1511 is updated, and the backup storage control device 1520 Since the location information table update unit 1522 immediately updates the location information table 1521, even when the main storage control device 1510 fails, the backup storage control device 1520 does not Taking over can be operated storage system, it is possible to realize a highly reliable storage system.
In the second embodiment, the update information transmitting unit 1512 of the main storage control device 1510 transmits a location information update message 1610 to the backup storage control device 1520, and the backup storage control device 1520 receives the location information update message 1610. However, the backup storage control device 1520 saves the location information update message 1610 as an update log without immediately updating the location information table 1521, and the main system The location information table 1521 can be updated when the storage control device 1510 fails and takes over the processing.
Therefore, a backup storage control device that stores the location information update message 1610 as an update log and updates the location information table 1521 when the main storage control device 1510 fails and takes over the processing will be described.
FIG. 18 is an explanatory diagram for explaining the concept of a modification example of duplication of the storage control apparatus according to the second embodiment. As shown in the figure, in this redundant modification of the storage control device, the main storage control device sends update information to the backup storage control device every time the location information table is updated. Further, the backup storage control apparatus that has received the update information saves the update information as an update log instead of immediately updating the position information table. Then, when the main storage control device fails and takes over the processing, the backup storage control device updates the location information table using the saved update log.
Thus, in this modification, instead of immediately updating the location information table when the backup storage control device receives the update information, the update information is stored as an update log, so the backup storage control device And the backup storage control apparatus can be used effectively for other processes.
Embodiment 3 FIG.
In the first and second embodiments described above, the case where two cache control devices are used has been described. However, the number of cache control devices is sequentially increased in order to avoid a decrease in the cache hit rate due to the increase in the number of disks 121 or the like. It is necessary to increase it. In the third embodiment, a storage system using n cache control devices will be described.
FIG. 19 is a block diagram showing a system configuration of the storage system according to the third embodiment. Here, for convenience of explanation, functional units that play the same functions as the respective units shown in FIG. 1 are given the same reference numerals, and detailed descriptions thereof are omitted.
As shown in the figure, in this storage system, a storage control device 1910, real disk control devices 120 and 1920, and cache control devices 130, 140, and 1930 are distributed on the network 40. Here, for convenience of explanation, two real disk control devices 120 and 1920 and three cache control devices 130, 140, and 1930 are shown, but this storage system includes m real disk control devices and n. Cache controller.
The storage control device 1910 performs control so that two of the n cache control devices store data in a dirty state. That is, in the storage control device 1910, even when one of the n cache control devices fails, the data of the failed cache control device always exists in another cache control device or disk. By controlling the cache control device in this way, an efficient use of cache memory and a highly reliable storage system are realized.
The real disk control device 1920 and the cache control device 1930 are devices newly added to the storage system in order to handle larger-scale data. However, the cache control device 1930 need not store the data M in the dirty state in FIG. 19, for example. The reason is that two cache control devices 130 and 140 store data M, and three or more cache control devices do not need to store the same data.
As described above, in the third embodiment, n cache control devices are provided in the storage system, and for dirty data that is not stored in the disk, two cache control devices among the n cache control devices are used. Since the same data is stored in the control device, the cache memory is used efficiently, and even if one of the n cache control devices fails, the remaining cache control devices The storage system can be operated by using it, and a highly reliable storage system can be realized.
In the third embodiment, the case where the same data is stored in two of the n cache control devices has been described. However, the present invention is not limited to this, and the reliability is further improved. In order to obtain a high storage system, the present invention can be similarly applied to the case where the number of cache control devices storing the same data is increased to three or more. Therefore, the number of cache control devices that store the same data can be determined based on the operating environment of the system, the hit rate of the cache memory, the data access performance, the degree of demand for system reliability, and the like.
Furthermore, the present invention can also be applied to a more flexible storage system by changing the number of cache control devices that store the same data as needed due to changes in the operating environment of the storage system. it can.
In the third embodiment, the case where the same data is stored in any two cache control devices has been described. However, the present invention is not limited to this and is the same for each cache control device. The present invention can be similarly applied to a storage system that defines a cache control device for storing the data and performs backup of each cache control device.
In the first, second, and third embodiments, the storage control device and the cache control device have been described. However, the storage control program and the cache control having the same function can be realized by realizing the configuration of these devices by software. You can get a program. Thus, a computer system that executes these storage control program and cache control program will be described.
FIG. 20 is a diagram illustrating an example of a computer system 2000 that executes a storage control program and a cache control program. As shown in the figure, the computer system 2000 includes an MPU (Micro Processing Unit) 2001, a ROM (Read Only Memory) 2002, a RAM (Random Access Memory) 2003, a NIC (Network Interface Card) 2004, and an input. An output device 2005 and a PC interface 2006 are included.
The MPU 2001 is an arithmetic device that executes a program. Here, the MPU 2001 executes a storage control program and a cache control program stored in the ROM 2002.
The ROM 2002 is a memory that stores a read-only program and data, and stores a storage control program and a cache control program prior to execution of the program.
A RAM 2003 is a memory in which data can be written, and stores temporary data and the like created by the MPU 2001 in accordance with execution of a program. The RAM 2003 stores a position information table.
The NIC 2004 is a device that communicates with a client or the like using a predetermined communication protocol, and executes processing of the network interface unit 113 shown in FIG. 2 and the network interface unit 133 shown in FIG.
The input / output device 2005 is a device such as a keyboard, a mouse, and a display. The keyboard and mouse are used for inputting instructions and data to the computer system 2000, and the display is used for displaying an execution state of the computer.
A PC interface 2006 is an interface with a personal computer used for program development and the like, and is used for loading of a program developed by the personal computer, exchange of debug information with the personal computer, and the like.
As described above, according to the present invention, control is performed so that the data requested by the client to be written is multiplexed and written to a predetermined number of cache devices arranged in the network, and the client performs reading. Controls that the requested data is read from the disk device and stored only in a certain cache device and sent to the client when none of the cache devices on the network has the requested data As a result, it is possible to efficiently use the cache memory, prevent the entire storage system from being stopped due to a failure of one of the cache devices, and build a highly reliable storage system. .

  As described above, the storage control device, the storage control program, the storage control method, and the storage system according to the present invention need to provide services without stopping the operation of the system even when some device fails. Suitable for highly reliable storage systems.

[Document Name] Description [Claims]
1. A storage control that is distributed and arranged on a network together with a cache device and a disk device to form a storage system, receives an access request from a client, and causes the cache device and the disk device to execute processing according to the access request A device,
Write control means for controlling the client to write the data requested to be written to a predetermined number of cache devices arranged in the network.
When none of the cache devices arranged in the network has the data requested by the client, the data is read from the disk device, and the read data is stored only in a certain cache device. And read control means for controlling to transmit to the client,
A storage control device comprising:
2. A storage control program used in a storage system in which a cache device and a disk device are distributed on a network, accepting an access request from a client, and causing the cache device and the disk device to execute processing according to the access request Because
A write control procedure for controlling to multiplex and write the data requested by the client to be written to a predetermined number of cache devices arranged in the network;
When none of the cache devices arranged in the network has the data requested by the client, the data is read from the disk device, and the read data is stored only in a certain cache device. And a read control procedure for controlling to transmit to the client,
Is executed by a computer.
3. When writing data that has been multiplexed and written to the predetermined number of cache devices by the write control procedure to the disk device, only a certain cache device of the predetermined number of cache devices is the data storage device. 3. The storage control program according to claim 2, wherein the computer further executes a flash control procedure in which data is continuously stored, and other cache devices are controlled to release the data storage area.
Wherein said write control procedure, characterized in that the client is controlled to write duplicated data requested to write to two cache apparatus of the arrangement cache device to the network The storage control program according to claim 2 or claim 3 .
Wherein said write control procedure, scope second preceding claims, wherein the client is controlled to write and multiplexes all the cache apparatus the requested data is arranged in the network to write Or the storage control program according to item 3 .
Wherein said write control procedure, the claims the for a given number of cache apparatus, characterized in that an instruction to transmit the predetermined number to the client along with the write completion notice of the data The storage control program according to item 2 or item 3 .
7. A storage area of data for which the client requests access is stored by storing a correspondence between a logical position designated by the client, a storage position in a cache device, and a storage location in a disk device. Each time an update is performed on the association, the computer further executes an update information transmission procedure for transmitting information related to the update to a storage control program executed by another computer arranged on the network. The storage control program according to claim 2 or claim 3 .
8. A storage control method used in a storage system in which a cache device and a disk device are distributed on a network, accepting an access request from a client, and causing the cache device and the disk device to execute processing according to the access request Because
A write control step of controlling the client to write the data requested to be written to a predetermined number of cache devices arranged in the network.
When none of the cache devices arranged in the network has the data requested by the client, the data is read from the disk device, and the read data is stored only in a certain cache device. And a read control step for controlling to transmit to the client;
A storage control method comprising:
DETAILED DESCRIPTION OF THE INVENTION
[0001]
【Technical field】
The present invention is used in a storage system in which a cache device and a disk device are distributed on a network, receives a request for access from a client, and causes the cache device and the disk device to execute processing according to the access request. The present invention relates to a control program and a storage control method, and in particular, a storage control device, a storage control program, and a storage control capable of preventing the entire storage system from being stopped due to a failure of an individual device and constructing a highly reliable storage system. It is about the method.
[0002]
[Background]
In recent years, with the rapid increase in the amount of processing data due to the use of multimedia data and the like, storage systems that integrate and manage large-scale data using a large-capacity disk connected to a network are rapidly spreading. In this storage system, the disk access speed is slower than the network data transfer speed, so the disk access performance is the bottleneck of the entire system. For this reason, the storage system uses a cache memory that can be accessed at high speed, and reduces the number of accesses to the disk, thereby improving the access performance of the disk.
[0003]
Normally, when a cache memory is used, the cache hit rate, which indicates the rate at which requested data is present in the cache memory, greatly affects the system performance. Therefore, it is necessary to use as large a cache memory as possible. However, in the conventional storage device, since the entire storage system control, disk control, cache memory control, etc. are realized as one device, it is not easy to expand the capacity of the cache memory. When the capacity of the disk is expanded by the increase, the relative capacity of the cache memory is reduced, and the cache hit rate is reduced.
[0004]
For this reason, in order to facilitate the expansion of the cache memory capacity, the entire storage system control, disk control, and cache memory control are implemented as separate devices, and these devices are connected to the network. Distributed storage systems that have been distributed in a distributed manner have come to be used. In this distributed storage system, since only the cache memory and its control are independent cache devices, the capacity of the cache memory can be easily expanded by adding a cache device.
[0005]
However, this distributed storage system has a problem in that each device fails separately because each device is arranged independently. Further, since each device is independently connected to the network, there is a problem that a failure occurs in the individual network connecting each device, and the individual device cannot be used.
[0006]
That is, in the conventional integrated storage apparatus, a failure occurs in the entire storage apparatus. Therefore, it is sufficient to perform Redundant Array of Independent Disks (RAID) control considering the failure of the entire storage apparatus. In the system, it is necessary to consider a failure for each individual device arranged in a distributed manner.
[0007]
The present invention has been made to solve the above-described problems caused by the prior art, and prevents the entire storage system from being stopped due to a failure of an individual device, thereby constructing a highly reliable storage system. An object of the present invention is to provide a storage control device, a storage control program, and a storage control method.
[0008]
DISCLOSURE OF THE INVENTION
In order to solve the above-described problems and achieve the object, the present invention forms a storage system in a distributed arrangement on a network together with a cache device and a disk device, receives an access request from a client, and receives the cache device and the disk device. A storage control device that executes processing according to an access request, wherein the data requested by the client to be written is multiplexed and written to a predetermined number of cache devices among the cache devices arranged in the network. The write control means for controlling and the data requested to be read by the client when none of the cache devices arranged in the network has the cache device, the data is read from the disk device, and the read data A certain Characterized by comprising a reading control means for controlling to transmit to said client causes only the storage shoe device.
[0009]
In addition, the present invention is used in a storage system in which a cache device and a disk device are distributed on a network, and receives storage access requests from a client and causes the cache devices and disk devices to execute processing according to the access request. A write control procedure for controlling to multiplex and write the data requested by the client to be written to a predetermined number of cache devices arranged in the network; When none of the cache devices arranged in the network has the requested data, the data is read from the disk device, and the read data is stored only in a certain cache device and And executes a read control step of controlling to transmit to the client, with the computer.
[0010]
In addition, the present invention is used in a storage system in which a cache device and a disk device are distributed on a network, and receives storage access requests from a client and causes the cache devices and disk devices to execute processing according to the access request. A write control step for controlling to multiplex and write data requested by the client for writing to a predetermined number of cache devices arranged in the network; When none of the cache devices arranged in the network has the requested data, the data is read from the disk device, the read data is stored only in a certain cache device, and the cluster is stored. Characterized in that it includes a reading control step of controlling to transmit the Ant, the.
[0011]
According to this invention, control is performed so that the data requested by the client to be written is multiplexed and written in a predetermined number of cache devices arranged in the network, and the data requested by the client is read to the network. When there is no cache device among the arranged cache devices, the data is read from the disk device, and the read data is stored only in a certain cache device and controlled to be transmitted to the client. While using the cache memory efficiently, it is possible to prevent the entire storage system from being stopped due to a failure of one of the cache devices and to construct a highly reliable storage system.
[0012]
BEST MODE FOR CARRYING OUT THE INVENTION
Exemplary embodiments of a storage control device, a storage control program, and a storage control method according to the present invention will be explained below in detail with reference to the accompanying drawings.
[0013]
Embodiment 1 FIG.
First, the system configuration of the storage system according to the first embodiment will be described. FIG. 1 is a block diagram showing a system configuration of the storage system according to the first embodiment. As shown in the figure, this storage system is configured with a storage control device 110, a real disk control device 120, and duplicated cache control devices 130 and 140 distributed on a network 40, and a client via the network 40. 10 to 30 are used. Although only three clients 10 to 30 are shown here for convenience of explanation, this storage system is used by an arbitrary number of clients via the network 40. Although the case where the client 10 uses this storage system will be described here, the other clients 20 and 30 can also use this storage system.
[0014]
The storage control device 110 is a device that accepts an access request from the client 10 via the network 40 and causes the real disk control device 120 and the cache control devices 130 and 140 to execute processing according to the access request. The storage control device 110 manages data stored in the cache memory included in the cache control devices 130 and 140.
[0015]
The real disk control device 120 has a disk 121 as a large-capacity storage medium, and reads data from the disk 121 and writes data to the disk 121 based on instructions from the storage control device 110. Although only one real disk control device 120 is shown here for convenience of explanation, in this storage system, a plurality of real disk control devices are arranged on the network 40 according to the scale of data to be stored. Can do. A plurality of disks can be connected to one real disk control device.
[0016]
The cache control devices 130 and 140 have a cache memory as a high-speed access recording medium, and read data from the cache memory and write data to the cache memory based on instructions from the storage control device 110.
[0017]
Thus, in this storage system, the cache control device is duplicated and the same data is stored in the cache memory of the cache control devices 130 and 140. However, if not all data is duplicated but the data stored in the disk 121 matches the data stored in the cache memory, that is, if the data is in a coherent state, the data is either When the data stored in the cache memory only and the data stored in the disk 121 are different from the data stored in the cache memory, that is, when the data is in a dirty state, the data is stored in both cache memories.
[0018]
For example, in FIG. 1, the data M is data in which the data M ′ on the disk 121 is rewritten and is in a dirty state, and therefore is stored in two cache memories of the cache control device 130 and the cache control device 140. Has been. The data N is stored only in the cache memory of the cache control device 130 because the same data exists on the disk 121.
[0019]
In this way, in this storage system, the cache control device is duplicated and the same data is stored in two cache memories. Therefore, even if one of the cache control devices 130 or 140 fails, the other data The storage system can be operated using the cache control device, and a highly reliable storage system can be realized. Further, in this storage system, when the data stored in the disk 121 and the cache memory are the same, the data is stored only in one cache memory, so that the cache memory can be used effectively. The cache hit rate can be improved.
[0020]
Next, the configuration of the storage control device 110 shown in FIG. 1 will be described. FIG. 2 is a functional block diagram showing the configuration of the storage control device 110 shown in FIG. As shown in the figure, the storage control device 110 includes a location information table 111, a control unit 112, a network interface unit 113, a data read processing unit 114, a data write processing unit 115, and a fetch processing unit 116. , And a flush processing unit 117.
[0021]
The location information table 111 includes a logical location designated by the client 10 for data that the client 10 requests access to the storage system, a physical location of the disk 121 where the data is actually stored, a physical location of the cache memory, and the data. (Ie, whether the data is coherent or dirty) is stored in association with each other.
[0022]
FIG. 3 is a diagram illustrating an example of the position information table 111. In the figure, data whose logical positions are 0 to 7 of the storage with the device number 0 are the 0 to 7 of the disk 121 with the device number 0 and 0 to 0 of the cache memory of the cache control device 130 with the device number 0. 7 indicates a coherent state. Further, the data whose logical positions are 8 to 15 of the storage of the device number 1 are the 8 to 15 of the disk 121 of the device number 0, the cache memories 8 to 15 of the cache control device 130 of the device number 0, and It is stored in 0-7 of the cache memory of the cache control device 140 having the device number 1 and indicates that it is in a dirty state.
[0023]
In this way, the location information table 111 stores the physical position of the data stored in the cache memory included in the two cache control devices 130 and 140 in correspondence with the logical position of the data, so that the storage control device 110 can perform processing corresponding to the duplication of the cache control device.
[0024]
The control unit 112 is a processing unit that controls the entire storage control device 110, and the network interface unit 113 communicates with the client 10, the real disk control device 120, and the cache control devices 130 and 140 using a predetermined communication protocol. It is a processing unit that performs.
[0025]
The data read processing unit 114 is a processing unit that reads data from the storage system in response to a data read request from the client 10, obtains a physical position from the logical position of the data using the position information table 111, and stores the data If it is stored in the cache memory, it instructs the cache control device having the cache memory storing the data to transfer data to the client 10. On the other hand, if the data is not stored in the cache memory, the Fetch processing unit 116 is requested to transfer data from the real disk control device 120 to the cache control device. Then, after the data is transferred from the real disk control device 120 to the cache control device, the cache control device to which the data has been transferred is instructed to transfer data to the client 10.
[0026]
The data write processing unit 115 is a processing unit that writes data to the storage system in response to a data write request from the client 10, and is always provided to both cache control devices 130 and 140 based on the location information table 111. To write data to the cache memory. That is, when the data storage position in the cache memory is registered in the position information table 111, the data write instruction is issued to the storage position, and the data storage position in the cache memory is registered in the position information table 111. If not, a new cache memory is allocated and data writing is instructed to the allocated cache memory position.
[0027]
As described above, since the data write processing unit 115 writes the same data in both cache memories in response to the data write request from the client 10, even when one of the cache control devices fails, The operation stop of the storage system can be prevented, and the reliability of the storage system can be improved.
[0028]
When the data requested to be read from the client 10 is not stored in the cache memory, the fetch processing unit 116 selects a cache control device that newly stores the data and selects the real disk control device 120. A processing unit for instructing transfer of data to the cache control device.
[0029]
The flush processing unit 117 is a processing unit that writes dirty data among data stored in the cache memory to the disk 121. When data is written to the cache memory, an empty area is insufficient, or a cache It is used to create an empty area in the cache memory when the used area of the memory exceeds a predetermined ratio. When creating an empty area in the cache memory, selection of which data is deleted from the cache memory is determined using a technique such as LRU (Least Recently Used), and the data determined to be deleted from the cache memory When the state is dirty, the flush processing unit 117 is used.
[0030]
Next, the configuration of the cache control devices 130 and 140 shown in FIG. 1 will be described. Since the cache control devices 130 and 140 have the same configuration, the cache control device 130 will be described as an example here. FIG. 4 is a functional block diagram showing the configuration of the cache control device 130 shown in FIG. As shown in the figure, the cache control device 130 includes a cache memory 131, a control unit 132, a network interface unit 133, a cache read processing unit 134, a cache write processing unit 135, a fetch correspondence unit 136, And a flush corresponding part 137.
[0031]
The cache memory 131 is a storage unit that stores a part of data stored in the disk 121 and data stored in the disk 121 in order to speed up access to the storage system. It is characterized by short access time and small capacity. The data stored in the cache memory 131 and its storage location are stored in the location information table 111 of the storage control device 110.
[0032]
The control unit 132 is a processing unit that controls the entire cache control device 130, and the network interface unit 133 communicates with the client 10, the storage control device 110, and the real disk control device 120 using a predetermined communication protocol. It is a processing unit.
[0033]
The cache read processing unit 134 is a processing unit that reads data from the cache memory 131 based on a data read instruction from the storage control device 110 and transfers the read data to the client 10.
[0034]
The cache write processing unit 135 is a processing unit that reads data from the buffer of the client 10 based on a data write instruction from the storage control device 110 and writes the read data to the cache memory 131. In addition, when the writing of data to the cache memory 131 is completed, the cache writing processing unit 135 notifies the client 10 of the completion of data writing. This completion notification corresponds to the duplication of the cache control device, and details thereof will be described later.
[0035]
The fetch correspondence unit 136 is a processing unit that writes data to the cache memory 131 that is transferred from the real disk control device 120 to the cache control device 130 based on an instruction from the fetch processing unit 116 of the storage control device 110.
[0036]
The flush corresponding unit 137 is a processing unit that reads data from the cache memory 131 based on an instruction from the flush processing unit 117 of the storage control device 110 and transfers the data to the real disk control device 120.
[0037]
Next, a message that the storage control device 110 transmits to instruct the cache control devices 130 and 140 and the real disk control device 120 to perform processing will be described. FIG. 5 is a diagram showing an example of a message transmitted by the storage control device 110. 2A shows a data access instruction message to the cache memory transmitted from the storage control device 110 to the cache control devices 130 and 140. FIG. 2B shows the storage control device 110 informing the real disk control device 120. A data transfer instruction message to be transmitted is shown.
[0038]
As shown in FIG. 6A, the data access instruction message 510 to the cache memory transmitted from the storage controller 110 to the cache controllers 130 and 140 has an internal header 511, a response header 512, and a data location 513. .
[0039]
The internal header 511 includes an address of the client 10 or the real disk control device 120 to which the cache control devices 130 and 140 perform data transmission / reception, a data read instruction (“read”) from the cache memory, or data to the cache memory A command type indicating whether the instruction is a write instruction (“write”) is included.
[0040]
That is, when the data access instruction message 510 is an instruction to write data to the cache memory, the address is the address of the client 10 that has made a data write request, and the command type is “write”. When the data access instruction message 510 is an instruction to read data from the cache memory, the address is the address of the client 10 that has made a data read request, and the command type is “read”. In addition, when the data access instruction message 510 is an instruction to transfer data from the cache memory to the disk 121, the address is the address of the real disk controller 120 as the data transfer destination, and the command type is “read”.
[0041]
The response header 512 is a header format for transmitting the result of execution of the processing instructed by the storage control device 110 by the cache control devices 130 and 140 to the client 10 or the storage control device 110. The response header 512 includes the response header 512. A command ID for identifying a request from 10 and an execution result indicating whether the execution has succeeded or failed are included. The data included in the response header 512 of the data access instruction message 510 instructing data writing to the cache memory will be described later.
[0042]
The data location 513 includes the physical location of the disk 121 storing data and the cache memory. However, if the data is not stored in the cache memory, only the physical location of the disk 121 is included.
[0043]
As shown in FIG. 5B, the data transfer instruction message 520 transmitted from the storage controller 110 to the real disk controller 120 also has an internal header 521, a response header 522, and a data location 523.
[0044]
The internal header 521 includes the address of the cache control device 130 or 140 that is the data transfer destination and “read” as the command type. The response header 522 is a header format for the real disk control device 120 to transmit the data transfer result (success or failure) to the storage control device 110. The data location 523 includes a physical location that specifies the location of the data on the disk 121 that stores the data to be transferred, and a data storage location in the cache memory that is the data write destination.
[0045]
Next, messages that the cache control devices 130 and 140 transmit to the client 10, the storage control device 110, and the real disk control device 120 in response to instructions from the storage control device 110 will be described. FIG. 6 is a diagram illustrating an example of a message transmitted by the cache control devices 130 and 140. FIG. 11A shows a response message 610 for transmitting the result of executing the data read process from the cache memory or the data write process to the cache memory based on the request of the client 10 to the client 10. b) shows a data transfer message 620 for transferring the data in the cache memory to the real disk control device 120 based on the flush processing, and FIG. 10C shows the case where the data transfer to the real disk control device 120 is completed. Shows a data transfer completion message to be transmitted to the storage controller 110.
[0046]
As shown in FIG. 5A, the response message 610 transmitted from the cache control devices 130 and 140 to the client 10 includes a response header 611 and response data 612. The response header 611 includes a command ID for identifying a request from the client 10 and an execution result of the instructed process. When the request from the client 10 is data read, the data read from the cache memory becomes the response data 612. On the other hand, when the request from the client is data writing, there is no response data 612.
[0047]
As shown in FIG. 5B, the data transfer message 620 transmitted from the cache control devices 130 and 140 to the real disk control device 120 includes a response header 621, a data location 622, and response data 623. The response header 621 includes a command ID for identifying that the data transfer message 620 is data transfer based on the flush process, and the data position 622 specifies the write position on the disk 121 of the data to be transmitted. The response data 623 is flush data that is read from the cache memory and stored in the disk 121.
[0048]
Further, as shown in FIG. 5C, the data transfer completion message 630 transmitted from the cache control apparatuses 130 and 140 to the storage control apparatus 110 has a response header 631. The response header 631 includes the command ID included in the instruction message from the storage controller 110 corresponding to this response and the execution result of the instructed process.
[0049]
Next, a message transmitted from the real disk control device 120 to the cache control devices 130 and 140 and the storage control device 110 in response to an instruction from the storage control device 110 will be described. FIG. 7 is a diagram showing an example of a message transmitted by the real disk control device 120. FIG. 11A shows a data transfer message 710 in which the real disk control device 120 transfers the data stored in the disk 121 to the cache control device 130 or 140 based on the Fetch process, and FIG. A data transfer completion message 720 transmitted from the real disk control device 120 to the storage control device 110 when the data transfer to the control device 130 or 140 is completed is shown.
[0050]
As shown in FIG. 5A, the data transfer message 710 transmitted from the real disk control device 120 to the cache control device 130 or 140 has a response header 711, a data location 712, and response data 713. The response header 711 includes a command ID for identifying that the data transfer message 710 is data transfer based on the Fetch process, and the data position 712 specifies the write position of the data to be transferred in the cache memory. The response data 713 is fetch data that is read from the disk 121 and written to the cache memory.
[0051]
Further, as shown in FIG. 5B, the data transfer completion message 720 transmitted to the storage control device 110 has a response header 721. The response header 721 includes the command ID included in the instruction message from the storage controller 110 corresponding to this response and the execution result of the instructed process.
[0052]
Next, the processing of the data write processing unit 115 of the storage control device 110 shown in FIG. 2 will be described. FIG. 8 is a flowchart showing a processing procedure of the data write processing unit 115 of the storage control apparatus 110 shown in FIG.
[0053]
As shown in the figure, the data write processing unit 115 searches the position information table 111 using the logical position of the write data designated by the client 10 (step S801), and there is data in the cache memory and its state is dirty. Whether there is data in the cache memory and the state is coherent or there is no data in the cache memory (step S802).
[0054]
If both cache memories have data and the status of the data is dirty, a data access instruction message 510 is transmitted to both cache control devices 130 and 140 with the command type “write”. Then, an instruction to write data designated by the client 10 to the cache memory is issued (step S803).
[0055]
If there is data in one of the cache memories and the state of the data is coherent, an area for storing the data is secured in the cache memory having no data, and both cache control devices 130 and 140 are provided. On the other hand, the data access instruction message 510 is transmitted with the command type as “write” (step S806). Then, the location information table 111 is updated for the newly secured cache memory area (step S807).
[0056]
If there is no data in both cache memories, an area for storing data in both cache memories is secured (step S805), and the command type for both cache control devices 130 and 140 is set to “ The data access instruction message 510 is transmitted as “write” (step S806). Then, the location information table 111 is updated for the newly secured cache memory area (step S807).
[0057]
As described above, since the data write processing unit 115 always writes the data requested by the client 10 to both cache memories, even if one of the cache control devices fails, the data is lost. Can be prevented.
[0058]
Next, a method in which the client 10 determines completion of data writing to the storage system will be described. In this storage system, two cache control devices 130 and 140 are used, and it is necessary to write data to the cache memory included in the two cache control devices 130 and 140. Therefore, the completion of data writing is the time when both cache control devices 130 and 140 have completed writing data to the cache memory. FIG. 9 is an explanatory diagram for explaining a method in which the client 10 determines completion of data writing to the storage system.
[0059]
As shown in the figure, when the storage controller 110 receives a data write request from the client 10 (command ID is “command 1”), the storage controller 110 instructs both cache controllers 130 and 140 to access data. A message 510 is transmitted. In this data access instruction message 510, the internal header 511 includes the address of the client 10 that is the data write request source and “write” that is the command type, and the response header 512 includes two responses to “command 1”. The data position 513 includes data position P and data position Q which are data write positions in the cache memory.
[0060]
The cache controllers 130 and 140 having received the data write message 510 read the data from the buffer of the client 10, the cache controller 130 is in the data position P of the cache memory, and the cache controller 140 is in the data position Q of the cache memory. Write data. When the data writing is completed, a response message 610 created based on the response header 512 is transmitted to the client 10. This response message 610 includes data (two responses) specifying that there are two responses to “command 1”.
[0061]
The client 10 that has received the response message 610 from the cache control device 130 or 140 confirms that there are two responses to “command 1”, waits for a response message 610 from another cache control device, When the response message 610 from the cache control device is received, it is determined that the data writing to the storage system is completed.
[0062]
In this way, the storage controller 110 uses the response header 512 to specify that there are two response messages 610 for the data write request, and the client 10 counts the number of response messages 610 from the cache controller, Since it is determined that the writing of data to the storage system is completed only when two response messages 610 are received, the same data is written to the cache memories of the two cache control devices 130 and 140. Can be ensured.
[0063]
Next, a procedure for writing data into the cache memory according to the first embodiment will be described. FIG. 10 is a flowchart showing a procedure for writing data to the cache memory according to the first embodiment. Here, it is assumed that the cache control device 130 completes writing of data to the cache memory before the cache control device 140.
[0064]
As shown in the figure, when the cache control device 130 receives the data access instruction message 510 transmitted from the storage control device 110 (step S1001), the cache control device 130 reads data from the buffer of the client 10 (step S1002), and reads the read data. Write to cache memory. A data write position in the cache memory is designated by a data position 513 included in the data access instruction message 510. When the writing of data to the cache memory is completed, a response message 610 is transmitted to the client 10 (step S1003).
[0065]
On the other hand, the other cache control devices 140 similarly receive the data access instruction message 510 transmitted by the storage control device 110 (step S1005), read the data from the buffer of the client 10 (step S1006), and cache the read data. Write to memory. When the writing of data to the cache memory is completed, a response message 610 is transmitted to the client 10 (step S1007).
[0066]
The client 10 first receives the response message 610 transmitted from the cache control device 130 (step S1004), recognizes that it is necessary to decode the response header 611 and receive two response messages 610, and then The response message 610 is waited for. When the response message 610 from the cache control device 140 is received (step S1008), it is determined that the data write processing is completed in both the cache control devices 130 and 140.
[0067]
In this example, when the cache control devices 130 and 140 complete the writing of data to the cache memory, the response message 610 is transmitted to the client 10. However, the cache control devices 130 and 140 write to the cache memory. Can be notified to the storage controller 110 instead of the client 10 when the data write is completed. In this case, since the storage control device 110 issues a data write instruction to the two cache control devices 130 and 140, the client control unit 110 receives the data write completion notification from the cache control devices 130 and 140 that issued the instruction, and then 10 can be notified of the completion of data writing.
[0068]
Next, the processing procedure of the data read processing unit 114 of the storage control device 110 shown in FIG. 2 will be described. FIG. 11 is a flowchart showing a processing procedure of the data read processing unit 114 of the storage control device 110 shown in FIG.
[0069]
As shown in the figure, the data read processing unit 114 uses the position information table 111 to check whether or not the data requested by the client 10 to be read exists in the 0th cache control device 130 (step S1101). . If there is data in the 0th cache control device 130, a data access instruction message 510 is transmitted to the 0th cache control device 130 with the command type “read” (step S1102), and the cache memory Is instructed to transmit data stored in the client 10 to the client 10.
[0070]
On the other hand, if the data requested by the client 10 to be read is not in the number 0 cache control device 130, it is checked whether or not the data is in the number 1 cache control device 140 (step S1103). Further, even when all the data requested by the client 10 to be read out are not in the 0th cache control device 130, it is checked whether or not the remaining data is in the 1st cache control device 140 (step S1103). . When there is data to be requested from the first cache control device 140, the data access instruction message 510 is transmitted to the first cache control device 140 with the command type “read” (step S1104). An instruction to transmit the data stored in the memory to the client 10 is given.
[0071]
On the other hand, if the data requested by the client 10 is not in the first cache control device 140, it is checked whether or not the data is in the disk 121 (step S1105). Further, even if the 0th and 1st cache control devices 140 do not have all the data that the client 10 requested to read, it is checked whether or not the remaining data exists in the disk 121 (step S1105). If the data to be sought is on the disk 121, the data is read from the disk 121 using the Fetch process and written to the cache memory (step S1106). Then, a data access instruction message 510 is transmitted with the command type “read” to the cache control apparatus having the cache memory in which the data has been written (step S1107), and the data stored in the cache memory is transmitted to the client 10. Instruct.
[0072]
As described above, the read processing unit 114 checks the storage location of the data requested to be read by the client 10 using the location information table 111, and the data is in one of the two cache control devices 130 and 140. In such a case, by issuing an instruction to transmit the data to the client 10, data reading corresponding to the duplication of the cache is performed.
[0073]
Next, the processing procedure of the fetch processing unit 116 of the storage control device 110 shown in FIG. 2 will be described. Note that the process of the fetch processing unit 116 is a process corresponding to the fetch process (step S1106) shown in FIG. FIG. 12 is a flowchart showing a processing procedure of the fetch processing unit 116 of the storage control apparatus 110 shown in FIG.
[0074]
As shown in the figure, the fetch processing unit 116 secures a cache memory area for storing data using the position information table 111 (step S1201), and uses the data transfer instruction message 520 to secure the secured cache memory area. The real disk controller 120 is instructed to transfer data to the cache controller having the area (step S1202). Then, it waits for the completion of data transfer from the real disk control device 120 to the cache control device (step S1203), and receives a data transfer completion message 720 from the real disk control device 120 (step S1204). The position information table 111 is updated for (Step S1205).
[0075]
Next, the processing procedure of the flash processing unit 117 of the storage control apparatus 110 shown in FIG. 2 will be described. FIG. 13 is a flowchart showing the processing procedure of the flash processing unit 117 of the storage control apparatus 110 shown in FIG.
[0076]
As shown in the figure, the flush processing unit 117 transmits a message instructing data transfer to the real disk control device 120 to the cache control device storing the data to be transferred to the disk 121. That is, the address of the internal header 511 is used as the address of the real disk control device 120, the command type is “read”, the data access instruction message 510 is transmitted, and data transfer to the real disk control device 120 is instructed (step S1301). .
[0077]
Then, it waits for the completion of data transfer from the cache control device to the real disk control device 120 (step S1302). When the data transfer completion message 630 is received from the cache control device (step S1303), the cache storing the transferred data is stored. The memory area is released, and the position information table 111 is updated for the released cache memory area (step S1304).
[0078]
As described above, in the first embodiment, two cache control devices 130 and 140 are provided in the storage system, and the same data is stored in both cache control devices 130 and 140. Even when the cache control device fails, the storage system can be operated using another cache control device, and a highly reliable storage system can be realized.
[0079]
If the data stored in the disk 121 and the data stored in the cache control devices 130 and 140 are the same (coherent), the cache memory of one of the cache control devices is released and stored in the disk 121. Since the same data is stored in both cache control devices 130 and 140 only when the stored data and the data stored in the cache control device are different (dirty), the cache memory can be used efficiently, The cache hit rate can be improved.
[0080]
Embodiment 2. FIG.
In the first embodiment, the case where the reliability of the storage system is improved by duplicating the cache control apparatus has been described. However, in addition to the duplication of the cache control apparatus, the storage control apparatus is also duplexed. Thus, the reliability of the storage system can be further improved. Therefore, in the second embodiment, a storage system in which the storage control device is duplicated in addition to the duplication of the cache control device will be described.
[0081]
First, the concept of duplication of the storage control device according to the second embodiment will be described. FIG. 14 is an explanatory diagram for explaining the concept of duplication of the storage control apparatus according to the second embodiment. As shown in the figure, this storage system has two storage control devices, a main storage control device and a backup storage control device.
[0082]
In normal operation, the main storage control device of these two storage control devices functions as a storage control device of the storage system. However, both the main storage control device and the backup storage control device have a location information table, and the main storage control device sends the update information to the backup storage control device every time the location information table is updated. . Then, the backup storage control device that has received this update information updates its own location information table and always keeps the location information table up-to-date.
[0083]
As described above, in the storage system according to the second embodiment, the main storage control device is always kept in the same state by keeping the position information table of the main storage control device and the position information table of the backup storage control device. Even if a failure occurs, the storage system can be operated using the backup storage control device, and a highly reliable storage system can be realized.
[0084]
Next, the system configuration of the storage system according to the second embodiment will be described. FIG. 15 is a block diagram showing a system configuration of the storage system according to the second embodiment. Here, for convenience of explanation, functional units that play the same functions as the respective units shown in FIG. 1 are given the same reference numerals, and detailed descriptions thereof are omitted.
[0085]
As shown in the figure, in this storage system, a main storage control device 1510, a backup storage control device 1520, a real disk control device 120, and two cache control devices 130 and 140 are distributed on a network 40. Be placed.
[0086]
In the normal operation, the main storage controller 1510 receives an access request from the client 10 via the network 40, and performs processing according to the access request to the real disk controller 120 and the cache controllers 130 and 140. It is a device to be executed. The main storage control device 1510 includes a location information table 1511 and an update information transmission unit 1512.
[0087]
The location information table 1511 includes the logical location specified by the client 10 for the data that the client 10 requests to access the storage system, the physical location of the disk 121 where the data is actually stored, the physical location of the cache memory, and the data. This is a table in which the state of data, that is, the distinction of whether the data is coherent or dirty is stored in correspondence.
[0088]
The update information transmission unit 1512 is a processing unit that transmits update information to the backup storage control device 1520 each time the location information table 1511 is updated. The location information table 1511 is updated when the data stored in the cache memory by the flush process becomes unnecessary, when a new cache memory is allocated by the fetch process, and when a new cache memory is allocated by the data write process. This is done in some cases.
[0089]
The backup storage control device 1520 is a device that takes over the processing when the main storage control device 1510 fails, and includes a location information table 1521 and a location information table update unit 1522.
[0090]
The position information table 1521 is a table that stores the same data as the position information table 1511 included in the main storage control device 1510. The location information table 1521 stores the same data as the location information table 1511, so that when the main storage control device 1510 fails, the backup storage control device 1520 can take over and execute the processing. Become.
[0091]
The location information table update unit 1522 is a processing unit that receives the update information of the location information table 1511 transmitted by the update information transmission unit 1512 of the main storage control device 1510 and updates the location information table 1521 of the backup storage control device 1520. is there. The location information table update unit 1522 updates the location information table 1521 so that the location information table 1521 can always be kept up-to-date, and when the main storage control device 1510 fails, the processing is performed in the backup system. The storage control device 1520 can take over and execute.
[0092]
Next, a location information update message that the update information transmission unit 1512 of the main storage control device 1510 transmits to the backup storage control device 1520 as update information will be described. FIG. 16 is a diagram showing an example of the location information update message. As shown in the figure, this location information update message 1610 has a type 1611, logical location information 1612, physical location information 1613, and a status 1614.
[0093]
The type 1611 indicates the type of update of the location information table, and is either new allocation (alloc) of cache memory, partial release of a duplicated cache memory (free1), or full release of a cache memory (free). .
[0094]
The logical position information 1612, the physical position information 1613, and the state 1614 are information corresponding to the logical position, the physical position, and the state of the position information table, respectively. In other words, the logical position information 1612 is information regarding a logical position that is a data storage position used by the client 10 to access the storage system, and the physical position information 1613 is a data storage position in the disk 121 and the cache memory for storing data. The state 1614 is information indicating whether the data stored in the disk 121 is the same as the data stored in the cache memory.
[0095]
Next, the processing procedure of the location information table update unit 1522 of the backup storage control device 1520 shown in FIG. 15 will be described. FIG. 17 is a flowchart showing a processing procedure of the location information table update unit 1522 of the backup storage control device 1520 shown in FIG.
[0096]
As shown in the figure, the location information table update unit 1522 sends a location information update message 1610 transmitted from the update information transmission unit 1512 of the main storage control device 1510 to the backup storage control device 1520 via the network 40. (Step S1701).
[0097]
Then, the type 1611 included in the location information update message 1610 is checked (step S1702). If the type 1611 is “free”, the information corresponding to the released cache memory area is deleted from the location information table 1521. (Step S1703) If the type 1611 is “free1”, the information in the location information table 1521 corresponding to the partially freed cache memory area is updated (Step S1704). If the type 1611 included in the location information update message 1610 is “alloc”, if there is information in the location information table 1521 corresponding to the newly allocated cache memory area, the information is updated. If there is no information corresponding to the cache memory area allocated to the location information table 1521, new correspondence information is created and added to the location information table 1521 (step S1705).
[0098]
As described above, in the second embodiment, the storage control device is duplicated using the main storage control device 1510 and the backup storage control device 1520, and the location information tables 1511 and 1521 are stored in both storage control devices 1510 and 1520. The update information transmission unit 1512 of the main storage control device 1510 transmits a location information update message 1610 to the backup storage control device 1520 each time the location information table 1511 is updated, and the backup storage control device 1520 Since the location information table update unit 1522 immediately updates the location information table 1521, even when the main storage control device 1510 fails, the backup storage control device 1520 does not Taking over can be operated storage system, it is possible to realize a highly reliable storage system.
[0099]
In the second embodiment, the update information transmitting unit 1512 of the main storage control device 1510 transmits a location information update message 1610 to the backup storage control device 1520, and the backup storage control device 1520 receives the location information update message 1610. However, the backup storage control device 1520 saves the location information update message 1610 as an update log without immediately updating the location information table 1521, and the main system The location information table 1521 can be updated when the storage control device 1510 fails and takes over the processing.
[0100]
Therefore, a backup storage control device that stores the location information update message 1610 as an update log and updates the location information table 1521 when the main storage control device 1510 fails and takes over the processing will be described.
[0101]
FIG. 18 is an explanatory diagram for explaining the concept of a modification example of duplication of the storage control apparatus according to the second embodiment. As shown in the figure, in this redundant modification of the storage control device, the main storage control device sends update information to the backup storage control device every time the location information table is updated. Further, the backup storage control apparatus that has received the update information saves the update information as an update log instead of immediately updating the position information table. Then, when the main storage control device fails and takes over the processing, the backup storage control device updates the location information table using the saved update log.
[0102]
Thus, in this modification, instead of immediately updating the location information table when the backup storage control device receives the update information, the update information is stored as an update log, so the backup storage control device And the backup storage control apparatus can be used effectively for other processes.
[0103]
Embodiment 3 FIG.
In the first and second embodiments described above, the case where two cache control devices are used has been described. However, the number of cache control devices is sequentially increased in order to avoid a decrease in the cache hit rate due to the increase in the number of disks 121 or the like. It is necessary to increase it. In the third embodiment, a storage system using n cache control devices will be described.
[0104]
FIG. 19 is a block diagram showing a system configuration of the storage system according to the third embodiment. Here, for convenience of explanation, functional units that play the same functions as the respective units shown in FIG. 1 are given the same reference numerals, and detailed descriptions thereof are omitted.
[0105]
As shown in the figure, in this storage system, a storage control device 1910, real disk control devices 120 and 1920, and cache control devices 130, 140, and 1930 are distributed on the network 40. Here, for convenience of explanation, two real disk control devices 120 and 1920 and three cache control devices 130, 140, and 1930 are shown, but this storage system includes m real disk control devices and n. Cache controller.
[0106]
The storage control device 1910 performs control such that two of the n cache control devices store data in a dirty state. That is, in the storage control device 1910, even when one of the n cache control devices fails, the data of the failed cache control device always exists in another cache control device or disk. By controlling the cache control device in this way, an efficient use of cache memory and a highly reliable storage system are realized.
[0107]
The real disk control device 1920 and the cache control device 1930 are devices newly added to the storage system in order to handle larger-scale data. However, the cache control device 1930 need not store the data M in the dirty state in FIG. 19, for example. The reason is that two cache control devices 130 and 140 store data M, and three or more cache control devices do not need to store the same data.
[0108]
As described above, in the third embodiment, n cache control devices are provided in the storage system, and for the dirty data that is not stored in the disk, two cache control devices among the n cache control devices are used. Since the same data is stored in the control device, the cache memory is used efficiently, and even if one of the n cache control devices fails, the remaining cache control devices The storage system can be operated by using it, and a highly reliable storage system can be realized.
[0109]
In the third embodiment, the case where the same data is stored in two of the n cache control devices has been described. However, the present invention is not limited to this, and the reliability is further improved. In order to obtain a high storage system, the present invention can be similarly applied to the case where the number of cache control devices storing the same data is increased to three or more. Therefore, the number of cache control devices that store the same data can be determined based on the operating environment of the system, the hit rate of the cache memory, the data access performance, the degree of demand for system reliability, and the like.
[0110]
Furthermore, the present invention can also be applied to a more flexible storage system by changing the number of cache control devices that store the same data as needed due to changes in the operating environment of the storage system. it can.
[0111]
In the third embodiment, the case where the same data is stored in any two cache control devices has been described. However, the present invention is not limited to this and is the same for each cache control device. The present invention can be similarly applied to a storage system that defines a cache control device for storing the data and performs backup of each cache control device.
[0112]
In the first, second, and third embodiments, the storage control device and the cache control device have been described. However, the storage control program and the cache control having the same function can be realized by realizing the configuration of these devices by software. You can get a program. Thus, a computer system that executes these storage control program and cache control program will be described.
[0113]
FIG. 20 is a diagram illustrating an example of a computer system 2000 that executes a storage control program and a cache control program. As shown in the figure, the computer system 2000 includes an MPU (Micro Processing Unit) 2001, a ROM (Read Only Memory) 2002, a RAM (Random Access Memory) 2003, a NIC (Network Interface Card) 2004, and an input. An output device 2005 and a PC interface 2006 are included.
[0114]
The MPU 2001 is an arithmetic device that executes a program. Here, the MPU 2001 executes a storage control program and a cache control program stored in the ROM 2002.
[0115]
The ROM 2002 is a memory that stores a read-only program and data, and stores a storage control program and a cache control program prior to execution of the program.
[0116]
A RAM 2003 is a memory in which data can be written, and stores temporary data and the like created by the MPU 2001 in accordance with execution of a program. The RAM 2003 stores a position information table.
[0117]
The NIC 2004 is a device that communicates with a client or the like using a predetermined communication protocol, and executes processing of the network interface unit 113 shown in FIG. 2 and the network interface unit 133 shown in FIG.
[0118]
The input / output device 2005 is a device such as a keyboard, a mouse, and a display. The keyboard and mouse are used for inputting instructions and data to the computer system 2000, and the display is used for displaying an execution state of the computer.
[0119]
A PC interface 2006 is an interface with a personal computer used for program development and the like, and is used for loading of a program developed by the personal computer, exchange of debug information with the personal computer, and the like.
[0120]
As described above, according to the present invention, control is performed so that the data requested by the client to be written is multiplexed and written to a predetermined number of cache devices arranged in the network, and the client performs reading. Controls that the requested data is read from the disk device and stored only in a certain cache device and sent to the client when none of the cache devices on the network has the requested data As a result, it is possible to efficiently use the cache memory, prevent the entire storage system from being stopped due to a failure of one of the cache devices, and build a highly reliable storage system. .
[0121]
[Industrial applicability]
As described above, the storage control device, the storage control program, the storage control method, and the storage system according to the present invention need to provide services without stopping the operation of the system even when some device fails. Suitable for highly reliable storage systems.
[Brief description of the drawings]
[Figure 1]
FIG. 1 is a block diagram showing a system configuration of the storage system according to the first embodiment.
[Figure 2]
FIG. 2 is a functional block diagram showing the configuration of the storage control apparatus shown in FIG.
[Fig. 3]
FIG. 3 is a diagram showing an example of a position information table.
[Fig. 4]
FIG. 4 is a functional block diagram showing the configuration of the cache control device shown in FIG.
[Figure 5]
FIG. 5 is a diagram showing an example of a message transmitted by the storage control apparatus.
[Fig. 6]
FIG. 6 is a diagram showing an example of a message transmitted by the cache control device.
[Fig. 7]
FIG. 7 is a diagram showing an example of a message transmitted by the real disk control device.
[Fig. 8]
FIG. 8 is a flowchart showing a processing procedure of the data write processing unit shown in FIG.
FIG. 9
FIG. 9 is an explanatory diagram for explaining a method in which the client determines completion of data writing to the storage system.
FIG. 10
FIG. 10 is a flowchart showing a data write processing procedure to the cache memory according to the first embodiment.
FIG. 11
FIG. 11 is a flowchart showing a processing procedure of the data read processing unit shown in FIG.
FIG.
FIG. 12 is a flowchart showing a processing procedure of the Fetch processing section shown in FIG.
FIG. 13
FIG. 13 is a flowchart showing a processing procedure of the flush processing unit shown in FIG.
FIG. 14
FIG. 14 is an explanatory diagram for explaining the concept of duplication of the storage control apparatus according to the second embodiment.
FIG. 15
FIG. 15 is a block diagram showing a system configuration of the storage system according to the second embodiment.
FIG. 16
FIG. 16 is a diagram showing an example of the location information update message.
FIG. 17
FIG. 17 is a flowchart showing a processing procedure of the position information table update unit shown in FIG.
FIG. 18
FIG. 18 is an explanatory diagram for explaining a modification of duplication of the storage control apparatus according to the second embodiment.
FIG. 19
FIG. 19 is a block diagram showing a system configuration of the storage system according to the third embodiment.
FIG. 20
FIG. 20 is a diagram illustrating an example of a computer system that executes a storage control program and a cache control program.

Claims (17)

A storage control device that is distributed and arranged on a network together with a cache device and a disk device to form a storage system, accepts an access request from a client, and causes the cache device and the disk device to execute processing according to the access request,
Write control means for controlling the client to write the data requested to be written to a predetermined number of cache devices arranged in the network.
When none of the cache devices arranged in the network has the data requested by the client, the data is read from the disk device, and the read data is stored only in a certain cache device. And read control means for controlling to transmit to the client,
A storage control device comprising: When writing the data multiplexed and written in the predetermined number of cache devices by the write control means to the disk device, only a certain cache device of the predetermined number of cache devices continues to store the data. 2. The storage control apparatus according to claim 1, further comprising a flash control unit for controlling the other cache apparatus to release the data storage area. 2. The write control unit according to claim 1, wherein the write control unit performs control so that data requested by the client to be written is duplicated and written to two of the cache devices arranged in the network. Or the storage control device according to item 2. 3. The write control unit according to claim 1, wherein the write control unit performs control so that the data requested by the client to be written is multiplexed and written in all the cache devices arranged in the network. The storage control device described. 2. The write control unit according to claim 1, wherein the write control unit instructs the predetermined number of cache devices to transmit the predetermined number together with the data write completion notification to the client. Item 3. The storage control device according to item 2. A location information storage means for storing a correspondence between a logical location designated by the client, a storage location in the cache device, and a storage location in the disk device with respect to a storage area of data that the client requests access to;
Update information transmitting means for transmitting information relating to the update to another storage control device arranged on the network each time an update is performed on the association stored in the position information storage means;
The storage control device according to claim 1 or 2, further comprising: A storage control program that is used in a storage system in which a cache device and a disk device are distributed on a network, receives an access request from a client, and causes the cache device and the disk device to execute processing according to the access request,
A write control procedure for controlling to multiplex and write the data requested by the client to be written to a predetermined number of cache devices arranged in the network;
When none of the cache devices arranged in the network has the data requested by the client, the data is read from the disk device, and the read data is stored only in a certain cache device. And a read control procedure for controlling to transmit to the client,
Is executed by a computer. When writing the data multiplexed and written in the predetermined number of cache devices by the write control procedure to the disk device, only a certain cache device of the predetermined number of cache devices continuously stores the data. The storage control program according to claim 7, wherein the other cache device further executes a flash control procedure for controlling the data storage area to be released by the computer. 8. The write control procedure according to claim 7, wherein the data requested by the client to be written is controlled so as to be duplicated and written in two of the cache devices arranged in the network. Item 9. or storage control program according to item 8. 9. The write control procedure according to claim 7, wherein the write control procedure performs control so that the data requested by the client to be written is multiplexed and written to all the cache devices arranged in the network. The storage control program described. 8. The write control procedure according to claim 7, wherein the write control procedure instructs the predetermined number of cache devices to transmit the predetermined number together with the data write completion notification to the client. 9. The storage control program according to item 8. For the storage area of the data requested to be accessed by the client, a correspondence between a logical position designated by the client, a storage position in the cache device, and a storage position in the disk device is stored, and the stored correspondence is stored. The update information transmission procedure for transmitting information related to the update to a storage control program executed by another computer arranged on the network each time an update is performed is further executed by the computer. The storage control program according to item 7 or 8. A storage control method used in a storage system in which a cache device and a disk device are distributed on a network, accepts an access request from a client, and causes the cache device and the disk device to execute processing according to the access request.
A write control step of controlling the client to write the data requested to be written to a predetermined number of cache devices arranged in the network.
When none of the cache devices arranged in the network has the data requested by the client, the data is read from the disk device, and the read data is stored only in a certain cache device. And a read control step for controlling to transmit to the client;
A storage control method comprising: When writing the data multiplexed and written in the predetermined number of cache devices by the write control step to the disk device, only one of the predetermined number of cache devices continues to store the data. 14. The storage control method according to claim 13, further comprising a flush control step of controlling the other cache device to release the data storage area. The write control step controls to write the data requested by the client to be written in duplicate in two of the cache devices arranged in the network. 15. The storage control method according to item 14 or 14. 15. The write control process according to claim 13, wherein the write control step performs control so that the data requested by the client to be written is multiplexed and written in all the cache devices arranged in the network. The storage control method described. 15. The write control step of instructing the predetermined number of cache devices to transmit the predetermined number to the client together with a data write completion notification. The storage control method according to item.

Images