JP2010072974A - Program, device and method for update control - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To update a disk node system within a short time period. <P>SOLUTION: By an update target selection means 1b, a disk node 3o originally having no effective connection is selected as an update target disk node, and software is updated. On completion of the update processing of the disk node 3o, a disk node 3n is determined as a next selection disk node, by the update target selection means 1b. Then, based on an instruction by a connection switchover instruction means 1d, the effective connection of a storage device 2n is switched over from the disk node 3n to the disk node 3o. Thereafter, the software of the disk node 3n is updated. Thereafter, similarly, disk nodes are selected successively, and software is updated. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an update control program for controlling software update processing, an update control device, and an update control method, and more particularly to an update control program for controlling software update processing of a disk node connected to a storage device, an update control device, and It relates to the update control method.

  In a computer system composed of a plurality of nodes, it is desired that the versions of software executed on each node be prepared in order to ensure reliability. If the software can be updated without stopping the service, an update program can be transmitted to each node, and the update process can be executed at each node.

On the other hand, in the case of OS (Operating System) upgrade (system update), it is necessary to stop the node processing and perform the update operation. When performing a system update, if the service can be stopped, the service by all the disk nodes can be stopped and the system update of all the nodes can be performed. When the service is continuously updated, the update operation of each node is performed in a predetermined order in order to avoid a decrease in the processing capacity of the entire system as much as possible.
Japanese translation of PCT publication No. 2003-501881

  Meanwhile, there is a multi-node storage system in which data is distributedly stored in a plurality of storage apparatuses as one of computer systems composed of a plurality of nodes. In a multi-node storage system, a plurality of disk nodes are connected via a network. Each disk node has a large-scale storage device. The user communicates with the disk node via an access node connected to the network, and accesses an arbitrary storage device.

  Also in such a multi-node storage system, a system update is performed as necessary. The multi-node storage system manages an enormous amount of data and provides services to a large number of users. Therefore, it is often necessary to update the system without stopping the service. In a multi-node storage system, when updating a system without stopping service provision, it is necessary to move data in the storage device managed by the disk node to be updated to another storage device in advance.

  However, when a large amount of data is stored in the storage device, it takes a long time to move the data. In particular, with the recent increase in capacity of storage devices, there are storage devices having a terabyte order storage capacity. When all data in such a large-capacity storage device is moved to another storage device, it may take several hours to move the data.

  Moreover, in order to avoid a decrease in the processing capacity of the entire system as much as possible, the system update operation is performed one by one or several. Therefore, it takes a very long time to complete the update operation for all the disk nodes constituting the multi-node storage system.

  The present invention has been made in view of these points, and an object thereof is to provide an update control program, an update control apparatus, and an update control method capable of updating a disk node system in a short time.

  In order to solve the above-described problem, an update control apparatus that controls software update processing of a plurality of disk nodes is provided. The plurality of disk nodes is more than the plurality of storage apparatuses. Each storage device is connected to at least two disk nodes, and only one connection is defined as an effective connection relationship used for data input / output in the storage device. One or two storage devices are connected to the disk node. The update control apparatus includes update target selection means, update instruction means, connection switching instruction means, and connection information update means.

  The update target selection means refers to connection information storage means for storing connection relations between a plurality of storage devices and a plurality of disk nodes and valid connection relations, and disk nodes that do not have a valid connection relation are subject to update. When the disk node is selected and the update of the update target disk node software is completed, the disk node that has not been updated among the disk nodes connected to the same storage device as the update target disk node is determined as the next selected disk node. After the connection relationship of the next selected disk node is invalidated, the next selected disk node is selected as the update target disk node. The update instruction unit instructs the selected update target disk node to update the software every time the update target disk node is selected by the update target selection unit. When the next selected disk node is determined, the connection switching instruction means selects the next selection instruction to switch the valid connection of the storage device that is effectively connected to the next selected disk node from the next selected disk node to the update target disk node. Output to the disk node and the update target disk node. The connection information update means updates information on valid connections in the connection information storage means according to the connection relation validated or invalidated according to the instruction from the connection switching instruction means.

  An update control program for realizing the same function as that of the update control apparatus by a computer is provided. Furthermore, an update control method is provided in which the processing executed by the update control device is performed by a computer.

  The update control apparatus can update the software of each disk node without moving data in the storage apparatus even during system operation, and can perform update processing in a short time.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram showing an outline of the embodiment. The update control device 1 controls software update processing of a plurality of disk nodes 3a, 3b, 3c,..., 3m, 3n, 3o. The plurality of disk nodes 3a, 3b, 3c,..., 3m, 3n, and 3o are more than the plurality of storage apparatuses 2a, 2b,. In the example of FIG. 1, there are N storage devices 2a, 2b,..., 2m, 2n (N is a natural number), and the disk nodes 3a, 3b, 3c,. N + 1 units.

  The storage devices 2a, 2b,..., 2m, 2n are each connected to at least two disk nodes. Of the connections, only one connection is defined as an effective connection relationship used for data input / output in the storage apparatus. In the example of FIG. 1, the connection relationship between the disk nodes 3a, 3b, 3c,..., 3m, 3n, 3o and the storage devices 2a, 2b,. The invalid connection relation is indicated by a broken line.

  One or two storage devices are connected to the disk nodes 3a, 3b, 3c,..., 3m, 3n, 3o. In the example of FIG. 1, one storage device is connected to each of the disk nodes 3a and 3o at the left and right ends in the figure. Two storage devices are connected to the other disk nodes 3b, 3c,..., 3m, 3n.

  The update control device 1 includes a connection information storage unit 1a, an update target selection unit 1b, an update instruction unit 1c, a connection switching instruction unit 1d, and a connection information update unit 1e.

  The connection information storage unit 1a is connected to the plurality of storage devices 2a, 2b,..., 2m, 2n and the plurality of disk nodes 3a, 3b, 3c,. Stores connection relations.

  The update target selection means 1b refers to the connection information storage means for storing the connection relation between the plurality of storage apparatuses 2a, 2b,..., 2m, 2n and the plurality of disk nodes, and the effective connection relation. A disk node having no connection relationship is selected as an update target disk node. In addition, when the update of the update target disk node is completed, the update target selection unit 1b selects a disk node that has not yet been processed among the disk nodes connected to the same storage device as the update target disk node as the next selected disk. Determine as a node. Then, after the connection relationship of the next selected disk node is invalidated, the update target selecting unit 1b selects the next selected disk node as the update target disk node.

  Each time the update target selection unit 1b selects an update target disk node, the update instruction unit 1c instructs the selected update target disk node to update the software.

  When the next selected disk node is determined, the connection switching instruction unit 1d issues an instruction to switch the effective connection of the storage apparatus that is effectively connected to the next selected disk node from the next selected disk node to the update target disk node. Output to the selected disk node and the update target disk node.

  The connection information update unit 1e updates information on valid connections in the connection information storage unit 1a in accordance with the connection relation validated or invalidated according to the instruction from the connection switching instruction unit 1d.

  According to such an update control device 1, the update target selection unit 1b selects the disk node 3o that does not have a valid connection first as the update target disk node. Then, the software of the disk node 3o is updated based on the instruction from the update instruction unit 1c.

  When the software update of the disk node 3o is completed, the update target selection unit 1b determines that the disk node 3n that is connected to the same storage device 2n as the disk node 3o and has not been updated is the next selected disk node. Then, the effective connection of the storage device 2n is switched from the disk node 3n to the disk node 3o based on the instruction from the connection switching instruction unit 1d. As a result, the disk node 3n has no effective connection. Therefore, the update target selection unit 1b selects the disk node 3n as the update target disk node, and the software of the disk node 3n is updated. Thereafter, similarly, the disk nodes are sequentially selected from the right to the left in the arrangement of the disk nodes in FIG. 1, and the software is updated.

  By executing the update process in this way, the software of the disk node can be updated without moving the data in the storage apparatus. In addition, the storage device cannot be accessed only during connection switching, and the software can be updated without substantially stopping the service.

  Incidentally, there is a multi-node storage system as a system for managing a plurality of storage devices with a plurality of disk nodes. In a multi-node storage system, reliability can be enhanced by storing data in duplicate in different storage devices. Therefore, taking the case where the function shown in FIG. 1 is applied to a multi-storage system with data duplication as an example, details of this embodiment will be described.

  The software update may be performed when an old version of software is upgraded to the latest version or when an unstable latest version of software is downgraded to an older version. In the present embodiment, the process of changing the software version including upgrade and downgrade is called update. In order to make the explanation easy to understand, the version before the update is called the old version, and the version of the software after the update is called the new version.

  FIG. 2 is a diagram illustrating a configuration example of the multi-node storage system according to the present embodiment. In the present embodiment, a plurality of disk nodes 100, 200, 300, 400, a control node 500, a management node 600, and an access node 700 are connected via the network 10. In addition, the number of storage apparatuses 810, 820, and 830, which is one less than the disk nodes 100, 200, 300, and 400, is provided. Each disk node 100, 200, 300, 400 can connect two storage devices. Each storage device 810, 820, 830 can be connected to two disk nodes. In the example of FIG. 2, the storage apparatus 810 is connected to the disk node 100 and the disk node 200. The storage device 820 is connected to the disk node 200 and the disk node 300. The storage device 830 is connected to the disk node 300 and the disk node 400.

  Each storage device 810, 820, 830 is equipped with a plurality of hard disk devices (HDDs). Each storage device 810, 820, 830 is a RAID system using a built-in HDD. In this embodiment, each storage device 810, 820, 830 provides a RAID 5 disk management service.

  The disk nodes 100, 200, 300, and 400 are computers having an architecture called IA (Intel Architecture), for example. When two storage apparatuses are connected to the disk nodes 100, 200, 300, and 400, only one connection is valid. Each disk node 100, 200, 300, 400 manages data stored in an effectively connected storage device, and provides the managed data to the access node 700 via the network 10. The disk nodes 100, 200, 300, and 400 manage data having redundancy. That is, the same data is managed by at least two disk nodes.

  The control node 500 manages the disk nodes 100, 200, 300, 400. For example, when the control node 500 receives a new storage device connection notification from the disk nodes 100, 200, 300, and 400, the control node 500 defines a new virtual disk and stores it in the storage device connected via the virtual disk. Makes it possible to access previously stored data.

  In the access node 700, a virtual disk is defined. Then, the access node 700 accesses the corresponding data in the disk nodes 100, 200, 300, and 400 assigned to the virtual disk.

  The management node 600 is a computer used by an administrator to manage the operation of the multi-node storage system. For example, the management node 600 manages version information (version information) of software installed in the disk nodes 100, 200, 300, and 400. Then, in response to a software update instruction from the administrator, the management node 600 sequentially transmits a software update instruction to each of the disk nodes 100, 200, 300, 400. At that time, the management node 600 controls the effective connection relationship between the disk nodes 100, 200, 300, and 400 and the storage apparatuses 810, 820, and 830 so that the service in the multi-node storage system does not stop.

  In the multi-node storage system having such a configuration, the access environment from the access node 700 to the data in the storage devices 810, 820, and 830 can be provided using three disk nodes. As a result, one of the four disk nodes does not have an effective connection relationship with any storage device. Software update processing such as an OS for such unused disk nodes can be executed without affecting the services of the multi-node storage system. Therefore, the validity / invalidity of the connection relationship between the disk nodes 100, 200, 300, and 400 and the storage apparatuses 810, 820, and 830 is controlled, and the disk nodes 100, 200, 300, and 400 are sequentially set to the unused state. Transition. Then, software update processing is sequentially executed for the disk nodes that are not in use. As a result, an update process involving the restart of the OS can be executed for all the disk nodes 100, 200, 300, and 400 without affecting the service provided by the multi-node storage system.

  FIG. 3 is a diagram illustrating a hardware configuration example of the management node used in the present embodiment. The management node 600 is entirely controlled by a CPU (Central Processing Unit) 601. A RAM (Random Access Memory) 602, a hard disk drive (HDD: Hard Disk Drive) 603, a graphic processing device 604, an input interface 605, and a communication interface 606 are connected to the CPU 601 through a bus 607.

  The RAM 602 is used as a main storage device of the management node 600. The RAM 602 temporarily stores at least part of an OS (Operating System) program and application programs to be executed by the CPU 601. The RAM 602 stores various data necessary for processing by the CPU 601. The HDD 603 is used as a secondary storage device of the management node 600. The HDD 603 stores an OS program, application programs, and various data. Note that a semiconductor storage device such as a flash memory can also be used as the secondary storage device.

  A monitor 11 is connected to the graphic processing device 604. The graphic processing device 604 displays an image on the screen of the monitor 11 in accordance with a command from the CPU 601. Examples of the monitor 11 include a display device using a CRT (Cathode Ray Tube) and a liquid crystal display device.

  A keyboard 12 and a mouse 13 are connected to the input interface 605. The input interface 605 transmits a signal sent from the keyboard 12 or the mouse 13 to the CPU 601 via the bus 607. The mouse 13 is an example of a pointing device, and other pointing devices can also be used. Examples of other pointing devices include a touch panel, a tablet, a touch pad, and a trackball.

  The communication interface 606 is connected to the network 10. The communication interface 606 transmits and receives data to and from other computers via the network 10.

  With the hardware configuration as described above, the processing functions of the present embodiment can be realized. Although FIG. 3 shows the hardware configuration of the management node 600, the disk nodes 100, 200, 300, 400, the control node 500, and the access node 700 can also be realized with the same hardware configuration. . In addition to the configuration shown in FIG. 3, the disk nodes 100, 200, 300, and 400 have two interfaces for connecting external devices in order to connect two storage devices.

Next, an outline of software update processing in the disk nodes 100, 200, 300, and 400 in the multi-node storage system will be described.
FIG. 4 is a first diagram illustrating state transitions in the software update process. In the example of FIG. 4, the version of the update target software in each disk node 100, 200, 300, 400 is shown. Further, if the line connecting each of the disk nodes 100, 200, 300, 400 and the storage apparatuses 810, 820, 830 is a solid line, it indicates that the connection is valid, and if the line is a broken line, the connection is invalid. It shows that.

  The first state (ST1) shows the connection relationship at the start of the update process. In this state, the software version of each disk node 100, 200, 300, 400 is “V1”. The connection between the disk node 100 and the storage apparatus 810, the connection between the disk node 200 and the storage apparatus 820, and the connection between the disk node 300 and the storage apparatus 830 are valid connections.

In the first state, the disk node 400 is not used. Therefore, first, software update processing of the disk node 400 is performed.
The second state (ST2) shows a state after completion of the update process for the first disk node. The update target software of the disk node 400 has the version “V2” due to the update process.

  When the update process of the disk node 400 is completed, the management of the storage apparatus 830 is transferred from the disk node 300 to the disk node 400. Therefore, the connection between the disk node 300 and the storage device 830 is invalidated, and the connection between the disk node 400 and the storage device 830 is validated.

  The third state (ST3) shows the state after switching the connection to the storage apparatus 830. Since the connection between the disk node 300 and the storage device 830 is invalidated, the disk node 300 is in a non-use state. On the other hand, since the connection between the disk node 400 and the storage apparatus 830 is validated, the data of the storage apparatus 830 is managed by the disk node 400 thereafter.

Since the disk node 300 is not used in the third state, the software update process of the disk node 300 is performed.
The fourth state (ST4) shows the state after completion of the update process for the second disk node. The update target software of the disk node 300 has the version “V2” by the update process.

  When the update process of the disk node 300 is completed, the management of the storage device 820 is transferred from the disk node 200 to the disk node 300. Therefore, the connection between the disk node 200 and the storage device 820 is invalidated, and the connection between the disk node 300 and the storage device 820 is validated.

  FIG. 5 is a second diagram illustrating state transition in the software update process. The fifth state (ST5) shows a state after switching the connection to the storage apparatus 820. Since the connection between the disk node 200 and the storage device 820 is invalidated, the disk node 200 is in a non-use state. On the other hand, since the connection between the disk node 300 and the storage device 820 is validated, the data of the storage device 820 is managed by the disk node 300 thereafter.

Since the disk node 200 is not used in the fifth state, the software update process of the disk node 200 is performed.
The sixth state (ST6) shows a state after the update process for the third disk node is completed. The update target software of the disk node 200 has the version “V2” by the update process.

  When the update process of the disk node 200 is completed, the management of the storage apparatus 810 is transferred from the disk node 100 to the disk node 200. Therefore, the connection between the disk node 100 and the storage apparatus 810 is invalidated, and the connection between the disk node 200 and the storage apparatus 810 is validated.

  The seventh state (ST7) shows the state after switching the connection to the storage apparatus 810. Since the connection between the disk node 100 and the storage device 810 is invalidated, the disk node 100 is not in use. On the other hand, since the connection between the disk node 200 and the storage device 810 is validated, the data of the storage device 810 is managed by the disk node 200 thereafter.

Since the disk node 100 is not used in the seventh state, the software update process of the disk node 100 is performed.
The eighth state (ST8) shows a state after the update process for the fourth disk node is completed. The update target software of the disk node 100 has the version “V2” due to the update process.

  In this way, by sequentially switching the effective connection to the storage apparatus, the software of all the disk nodes 100, 200, 300, 400 can be updated without stopping the service. The management node 600 controls the timing of validation / invalidation of the connection between the disk node and the storage apparatus and the update processing in each of the disk nodes 100, 200, 300, 400. Each disk node 100, 200, 300, 400 executes validation / invalidation and software update processing in accordance with instructions from the management node 600.

  When the software version is updated to “V3” during operation in the eighth state (ST8) in FIG. 5, the software update processing is performed by the disk node 100, the disk node 200, the disk node 300, the disk It is executed in the order of the node 400. In this case, the connection relationship between the disk nodes 100, 200, 300, and 400 and the storage apparatuses 810, 820, and 830 follows the state transitions shown in FIGS.

  By the way, when the update process as shown in FIGS. 4 and 5 is executed, the disk nodes that manage each storage device are different before and after the update process. Therefore, in order to continue the operation of the multi-node storage system, it is necessary to appropriately notify the access node 700 that the disk node as the access destination for data access has been changed. Therefore, the function and processing procedure of the update process will be described in detail, including the process of notifying that the access destination to the access node 700 has changed along with the software update process.

  First, the principle for accessing data in the storage apparatus from the access node 700 will be described. In this embodiment, a virtual disk is defined in the access node 700. In the virtual disk, a storage area is divided into data units called segments, and a storage area in any storage device is associated with each segment. Then, the access node 700 determines the disk node that manages the data to be accessed via the virtual disk, and accesses the data in the storage apparatuses 810, 820, and 830.

  FIG. 6 shows the data structure of the virtual disk. In the present embodiment, a virtual disk identifier “LVOL-X” is assigned to the virtual disk 60. Storage IDs “DS-A”, “DS-B”, and “DS-C” are assigned to the storage devices 810, 820, and 830 connected via the network, respectively, for identification of the individual devices. .

  Each of the storage devices 810, 820, and 830 constitutes a RAID 5 storage system. The storage function provided by each of the storage apparatuses 810, 820, and 830 is divided into a plurality of slices 811a to 811e, 821a to 821e, and 831a to 831e and managed.

  The virtual disk 60 is configured in units of segments 61 to 64. The storage capacity of the segments 61 to 64 is the same as the storage capacity of a slice that is a management unit in the storage apparatuses 810, 820, and 830. For example, if the storage capacity of the slice is 1 gigabyte, the storage capacity of the segment is also 1 gigabyte. The storage capacity of the virtual disk 60 is an integral multiple of the storage capacity per segment. Each of the segments 61 to 64 includes a set (slice pair) of primary slices 61a, 62a, 63a, and 64a and secondary slices 61b, 62b, 63b, and 64b.

  Two slices belonging to the same segment belong to different disk nodes. In the area for managing individual slices, there are flags in addition to the virtual disk identifier, segment information, and slice information constituting the same segment, and a value representing primary or secondary is stored in the flag.

  In the example of FIG. 6, the identifier of the slice in the virtual disk 60 is indicated by a combination of “P” or “S” alphabets and numbers. “P” indicates a primary slice. “S” indicates a secondary slice. The number following the alphabet represents what number segment it belongs to. For example, the primary slice of the first segment 61 is indicated by “P1”, and the secondary slice is indicated by “S1”.

  By defining such a virtual disk 60, when the disk node that manages the storage device is changed, it is only necessary to change the designation of the disk node that manages the slice assigned to each segment. .

  Next, functions necessary for updating software of a disk node in a multi-node storage system that performs data access using a virtual disk will be described.

FIG. 7 is a block diagram showing the function of each node. The control node 500 includes a virtual disk management unit 510 and a virtual disk metadata storage unit 520.
The virtual disk management unit 510 manages slices in the storage apparatuses 810, 820, and 830. For example, the virtual disk management unit 510 transmits a metadata acquisition request to the disk nodes 100, 200, 300, and 400 when the system is activated. Then, the virtual disk management unit 510 generates virtual disk metadata from the metadata returned in response to the metadata acquisition request, and stores the virtual disk metadata in the virtual disk metadata storage unit 520.

  The virtual disk metadata storage unit 520 is a storage device that stores virtual disk metadata generated based on metadata collected from the disk nodes 100, 200, 300, and 400. For example, a part of the RAM storage area in the control node 500 is used as the virtual disk metadata storage unit 520.

  The management node 600 has an update control unit 610. The update control unit 610 controls software update processing to be executed by the disk nodes 100, 200, 300, and 400. Details of the function of the update control unit 610 will be described later.

  The access node 700 has a virtual disk access control unit 710. The virtual disk access control unit 710 performs data access to the disk node that manages the specified data in response to an access request specifying the data in the virtual disk 60 from the user. Specifically, the virtual disk access control unit 710 specifies a block in the virtual disk 60 in which data to be accessed is stored. Next, the virtual disk access control unit 710 specifies a segment corresponding to the specified block. Further, the virtual disk access control unit 710 specifies a disk node corresponding to a primary slice that constitutes a segment and a slice in the disk node. Then, the virtual disk access control unit 710 outputs an access request to the specified slice to the specified disk node.

  The disk node 200 includes a communication unit 210, a data access unit 220, a data management unit 230, a software (SW) update unit 240, a metadata storage unit 250, and a connected storage management unit 260.

  The communication unit 210 performs data communication with other disk nodes 100, 300, 400, the control node 500, the management node 600, and the access node 700 via the network 10. In response to an access request from the access node 700, the data access unit 220 accesses data in the storage apparatus having a valid connection relationship among the storage apparatuses 810 and 820. The data management unit 230 manages data in the storage apparatus having a valid connection relationship. The SW update unit 240 updates software executed on the disk node 200. The metadata storage unit 250 stores, for each slice in the storage apparatus having a valid connection relationship, information (metadata) regarding the segment to which the slice is assigned. The connected storage management unit 260 manages the validity / invalidity of the connection between the two storage apparatuses 810 and 820. Details of the function of the disk node 200 will be described later.

  Similar to the disk node 200, the other disk nodes 100, 300, 400 are also communication units 110, 310, 410, data access units 120, 320, 420, data management units 130, 330, 430, SW update units 140, 340, 440, metadata storage units 150, 350, and 450, and connection storage management units 160, 360, and 460. The components of the disk nodes 100, 300, and 400 have the same functions as the components of the same name of the disk node 200.

  FIG. 8 is a diagram illustrating an example of the data structure of the storage apparatus. In addition to the slices 812a, 812b, 812c,..., A plurality of metadata 813a, 813b, 813c,. The metadata 813a, 813b, 813c,... Stored in the storage device 810 is read by the data management unit of the disk node at the time of activation of the disk node whose connection is valid, and stored in the metadata storage unit. Stored.

  For example, when the connection relationship with the storage device 810 is valid in the disk node 200, the metadata 813a, 813b, 813c,... Of the storage device 810 is read by the data management unit 230, and the metadata storage unit 250. The data structure of the metadata storage unit 250 that stores the metadata 813a, 813b, 813c,... In the storage device 810 will be described below.

  FIG. 9 is a diagram illustrating an example of the data structure of the metadata storage unit. The metadata storage unit 250 stores a metadata table 251. The metadata table 251 includes columns for storage ID, slice ID, status, virtual disk ID, segment ID, paired storage ID, and paired slice ID. Information arranged in the horizontal direction in the metadata table 251 is associated with each other to form one record indicating the metadata.

In the storage ID column, identification information (storage ID) of the storage device 810 is set.
In the slice ID column, identification information (slice ID) in the storage device 810 of the slice corresponding to the metadata is set.

  In the status column, a status flag indicating the status of the slice is set. If the slice is not assigned to a virtual disk segment, the status flag “F” is set. When the virtual disk segment is assigned to the primary storage, the status flag “P” is set. When the virtual disk segment is allocated to the secondary storage, the status flag “S” is set.

In the virtual disk ID column, identification information (virtual disk ID) for identifying the virtual disk to which the segment corresponding to the slice belongs is set.
In the segment ID column, identification information (segment ID) of a segment to which a slice is assigned is set.

In the pair storage ID column, identification information (storage ID) of a storage device having a pair slice (another slice belonging to the same segment) is set.
In the pair slice ID column, identification information (slice ID) for identifying the pair slice within the storage apparatus to which the slice belongs is set.

  The metadata stored in the metadata storage units 150, 250, 350, 450 of each disk node 100, 200, 300, 400 is transmitted to the control node 500 in response to a request from the control node 500. The control node 500 creates virtual disk metadata that defines the allocation of slices to the virtual disk 60 based on the metadata collected from the disk nodes 100, 200, 300, and 400. Specifically, the virtual disk management unit 510 of the control node 500 classifies the collected metadata according to the virtual disk ID. Then, the virtual disk management unit 510 sets a set of metadata having the same virtual disk ID as virtual disk metadata corresponding to the virtual disk indicated by the active disk ID. The created virtual disk metadata is stored in the virtual disk metadata storage unit 520.

  FIG. 10 is a diagram illustrating an example of the data structure of the virtual disk metadata storage unit. The virtual disk metadata storage unit 520 stores a virtual disk metadata table 521. The virtual disk metadata table 521 includes columns for storage ID, disk node IP address, slice ID, state, virtual disk ID, segment ID, paired storage ID, and paired slice ID. Information arranged in the horizontal direction in the virtual disk metadata table 521 is associated with each other to form one record indicating metadata. Information set in each column other than the disk node IP address of the virtual disk metadata table 521 is the same type of information as the column of the same name in the metadata table 251.

  In the column of the disk node IP address, the IP address of the disk node that manages the storage device indicated by the storage ID is set. That is, each storage device can be connected to two disk nodes, but only one connection relationship is valid. Therefore, the IP address of a disk node having a valid connection relationship with each storage device is set in the disk node IP address column.

  The contents of the virtual disk metadata table 521 are transmitted to the access node 700 by the virtual disk management unit 510 in response to a metadata request from the access node 700. In the access node 700, the virtual disk access control unit 710 holds the same metadata as the virtual disk metadata table 521 shown in FIG. When there is data access to the virtual disk from the user, the virtual disk access control unit 710 stores the IP address of the disk node that manages the access target data and the data based on the held metadata. The slice ID of the current slice can be determined.

Next, software update processing procedures in the system configured as described above will be described.
FIG. 11 is a sequence diagram illustrating the procedure of the software update process. The process shown in FIG. 11 is a process during the transition from the first state (ST1) shown in FIG. 4 to the second state (ST2). In the example of FIG. 11, data access from the access node 700 occurs even during the update process. In the following, the process illustrated in FIG. 11 will be described in order of step number.

  [Step S11] The update control unit 610 of the management node 600 transmits a software update instruction to the disk node 400 in response to an operation input from the system administrator. As a result, the software update process is started by the SW update unit 440 of the disk node 400.

  [Step S12] The virtual disk access control unit 710 of the access node 700 determines the IP address of the disk node that manages the access target data in response to a data access request from the user. In the example of FIG. 11, data access to data stored in the storage device 830 has occurred. In this case, the virtual disk access control unit 710 transmits an access request to the disk node 300.

  [Step S13] The data access unit 320 of the disk node 300 performs I / O access to the storage apparatus 830 in response to an access request from the access node 700. For example, if it is a data write access request, the data access unit 320 writes data to the storage device 830. If the access request is for reading data, the data access unit 320 reads data from the storage device 830.

  [Step S14] The storage apparatus 830 inputs and outputs data in response to an I / O access from the disk node 300. When data is written, after the data is written in the storage device 830, a write completion response is returned to the disk node 300. When data is read, the storage device 830 reads the data, and a read completion response is returned to the disk node 300 together with the read data.

  [Step S15] The data access unit 320 of the disk node 300 returns a completion response corresponding to the access request. In the case of data read access, data is included in the response content.

  [Step S16] After that, it is assumed that the software update process in the disk node 400 is completed. Then, the SW update unit 440 of the disk node 400 transmits an update completion response to the management node 600.

[Step S <b> 17] The update control unit 610 of the management node 600 transmits a request for invalidating the connection with the storage apparatus 830 to the disk node 300.
[Step S18] The connected storage management unit 360 of the disk node 300 sets the state of the connection port connected to the storage device 830 to invalid. Then, information indicating disconnection is transmitted from the disk node 300 to the storage apparatus 830.

[Step S <b> 19] The storage apparatus 830 performs a process of cutting off the connection with the disk node 300, and finally transmits a completion response to the disk node 300.
[Step S20] The connected storage management unit 360 of the disk node 300 transmits a connection invalidation completion response to the management node 600.

  [Step S21] The update control unit 610 of the management node 600 transmits a request to validate the connection with the storage apparatus 830 to the disk node 400 for which the software update process has been completed.

  [Step S22] The connected storage management unit 460 of the disk node 400 sets the state of the connection port connected to the storage device 830 to be valid. Then, a connection request is transmitted from the disk node 300 to the storage apparatus 830.

[Step S <b> 23] The storage apparatus 830 establishes a connection with the disk node 400 and transmits a final completion response to the disk node 400.
[Step S <b> 24] The connected storage management unit 460 of the disk node 400 transmits a response to the connection completion of connection validation to the management node 600.

  [Step S25] The update control unit 610 of the management node 600 transmits to the control node 500 an IP address change notification indicating that the IP address of the disk node that manages the storage device 830 is changed to the IP address of the disk node 400. .

  [Step S26] The virtual disk management unit 510 of the control node 500 responds to the IP address change notification from the virtual disk metadata table 521 in the virtual disk metadata storage unit 520 to the storage ID “DP-C. Search for metadata. Then, the virtual disk management unit 510 sets the IP address of the disk node 400 to the disk node IP address of the corresponding metadata. Thereafter, the virtual disk management unit 510 transmits an IP address change process completion response to the management node 600.

  [Step S27] After that, it is assumed that a data access request to the data stored in the storage device 830 is input to the access node 700. At this point, the access node 700 recognizes that the storage device 830 is managed by the disk node 300. Therefore, the virtual disk access control unit 710 of the access node 700 transmits an access request to the disk node 300.

[Step S28] The data access unit 320 of the disk node 300 responds an error to the access node 700 because the storage device 830 is not a management target.
[Step S <b> 29] The virtual disk access control unit 710 of the access node 700 transmits a metadata request to the control node 500.

[Step S30] The virtual disk management unit 510 of the control node 500 responds to the access node 700 with the metadata stored in the virtual disk metadata storage unit 520.
[Step S31] The virtual disk access control unit 710 of the access node 700 stores the metadata acquired from the control node 500. Further, the virtual disk access control unit 710 recognizes the latest state of the slice allocation to the segment of the virtual disk 60 based on the acquired metadata. Then, the virtual disk access control unit 710 recognizes the IP address of the disk node 400 that manages the storage apparatus 830 based on the acquired metadata, and transmits an access request to the disk node 400.

  [Step S32] The data access unit 420 of the disk node 400 performs I / O access to the storage apparatus 830 in response to an access request from the access node 700.

[Step S <b> 33] The storage apparatus 830 inputs and outputs data in response to an I / O access from the disk node 400, and responds to the result to the disk node 400.
[Step S34] The data access unit 420 of the disk node 400 returns a completion response corresponding to the access request. In the case of data read access, data is included in the response content.

  In this manner, the software update process and the connection validation / invalidation process for the storage apparatus are executed in conjunction with each other. As a result, the software update process can be executed by sequentially setting the disk nodes to the unused state while continuing the service in the multi-node storage system.

Next, the function of the update control unit 610 for issuing software update instructions and instructions for enabling / disabling connection to the storage apparatus will be described in detail.
FIG. 12 is a block diagram illustrating detailed functions of the update control unit. The update control unit 610 includes a version information storage unit 611, a connection information storage unit 612, a disk node IP address storage unit 613, a copy source software storage unit 614, a communication unit 615, an update target determination unit 616, a connection switching instruction unit 617, an update An instruction unit 618 and a connected disk node change notification unit 619 are provided.

  The version information storage unit 611 is a storage function that stores software versions installed in the disk nodes 100, 200, 300, and 400. For example, a part of the storage area of the HDD 603 is used as the version information storage unit 611.

  The connection information storage unit 612 is a storage function that stores information on connection relations between the disk nodes 100, 200, 300, 400 and the storage apparatuses 810, 820, 830, and valid connections. For example, a part of the storage area of the HDD 603 is used as the connection information storage unit 612.

  The copy source software storage unit 614 has a storage function for storing new versions of software (programs and data) executed by the disk nodes 100, 200, 300, and 400. For example, a part of the storage area of the HDD 603 is used as the copy source software storage unit 614.

  The communication unit 615 performs data communication using a predetermined communication protocol via the network 10. In the present embodiment, the communication unit 615 performs communication using TCP / IP (Transmission Control Protocol / Internet Protocol).

  The update target determination unit 616 determines a disk node to be updated in response to an operation input instructing to upgrade the software executed in the disk nodes 100, 200, 300, and 400. Specifically, the update target determination unit 616 refers to the version information storage unit 611 and recognizes the version of software currently installed in each disk node 100, 200, 300, 400. In addition, the update target determination unit 616 refers to the connection information storage unit 612 to determine to which disk node each storage device 810, 820, 830 is effectively connected. Then, the update target determination unit 616 sets a disk node to which any storage apparatus is not effectively connected and on which an older version of software is installed as an update target.

  The update target determination unit 616 notifies the update instruction unit 618 of a software update request to be updated. When an update completion response is received from the update instruction unit 618, the update target determination unit 616 sequentially determines the disk nodes to be updated until all the disk nodes become new versions.

  If the storage device is effectively connected to the disk node to be updated, the update target determination unit 616 requests the connection switching instruction unit 617 to invalidate the connection. In addition, when there is a storage device that has lost a valid connection relationship with the disk node due to invalidation of the connection, the update target determination unit 616 establishes the connection relationship validation between the storage device and the updated disk node. This is requested to the switching instruction unit 617.

  Furthermore, upon receiving a connection switching completion response from the connection switching instruction unit 617, the update target determination unit 616 updates the version information in the version information storage unit 611 and the connection information in the connection information storage unit 612. Then, the update target determination unit 616 notifies the connection disk node change notification unit 619 of the changed connection relationship.

  The connection switching instruction unit 617 transmits a connection switching instruction to the corresponding disk node in response to a request for validation or invalidation of the connection relationship between the storage apparatus and the disk node from the update target determination unit 616.

  In response to the disk node software update request from the update target determination unit 616, the update instruction unit 618 transmits an update instruction to the disk node. At this time, if the new version of update software is not stored in the disk node to be updated, the update instruction unit 618 reads the new version of software from the copy source software storage unit 614 and transmits it to the disk node.

  When the connection disk node change notification unit 619 receives the changed connection relationship from the update target determination unit 616, the connection disk node change notification unit 619 transmits an IP address change notification of the disk node that manages the storage device to the control node 500 along the new connection relationship. To do.

The functions of the update control device 1 shown in FIG. 1 are included in the update control unit 610 in FIG. The correspondence between functions is as follows.
The function of the connection information storage unit 1 a of the update control device 1 is included in the connection information storage unit 612 of the update control unit 610. The functions of the update target selection unit 1b and the connection information update unit 1e of the update control device 1 are included in the update target determination unit 616 of the update control unit 610. The function of the update instruction unit 1 c of the update control device 1 is included in the update instruction unit 618 of the update control unit 610. The function of the connection switching instruction unit 1 d of the update control device 1 is included in the connection switching instruction unit 617 of the update control unit 610.

Next, information held in the update control unit 610 will be described in detail.
FIG. 13 is a diagram illustrating a data structure example of the version information storage unit. The version information storage unit 611 has a version information table 611a. The version information table 611a has columns for disk node ID and version. In the disk node ID column, disk node identification information (disk node ID) is set. In the version column, the version of software installed in the disk node is set. Note that the version information table 611a is provided for each piece of software to be managed by the update control unit 610.

  FIG. 14 is a diagram illustrating an example of a data structure of the connection information storage unit. The connection information storage unit 612 has a connection information table 612a. The connection information table 612a has columns for a disk node ID, a first connection destination device ID, a second connection destination device ID, and a valid connection.

  In the disk node ID column, disk node identification information (disk node ID) is set. In the column of the first connection destination device ID, identification information (storage ID) of the storage device connected to the connection port whose disk node port number is “1” is set. In the second connection destination device ID column, identification information (storage ID) of the storage device connected to the connection port whose disk node port number is “2” is set. The port number of the connection port in which the connection relationship is valid is set in the valid connection column.

FIG. 15 is a diagram illustrating a data structure example of the disk node IP address storage unit. The disk node IP address storage unit 613 has an IP address table 613a.
The IP address table 613a has columns for disk node ID and IP address. In the disk node ID column, disk node identification information (disk node ID) is set. In the IP address column, the IP address of the disk node indicated by the corresponding disk node ID is set.

Next, a procedure of system update processing executed by the update control unit 610 of the management node 600 will be described in detail.
FIG. 16 is a flowchart showing the procedure of the system update process. In the following, the process illustrated in FIG. 16 will be described in order of step number. This process is started when an operation input instructing system update is performed.

  [Step S41] The update target determination unit 616 determines whether the system update has been completed. When the system update is completed, the software versions of the disk nodes 100, 200, 300, and 400 indicated by the version information storage unit 611 are all the same as the software versions stored in the copy source software storage unit 614. This is the case. When the system update is completed, the process ends. If the system update has not been completed, the process proceeds to step S42.

  [Step S42] The update target determination unit 616 searches for a disk node in an unused state that is disconnected from the storage device. Specifically, the update target determination unit 616 refers to the connection information table 612a in the connection information storage unit 612. Then, the update target determination unit 616 searches for a disk node whose valid connection is blank, and determines that the corresponding disk node is not used.

  [Step S43] The update target determination unit 616 determines whether it is necessary to upgrade an unused disk node. Specifically, the update target determination unit 616 refers to the version information table 611a in the version information storage unit 611, and determines the software version of the unused disk node. Then, the update target determining unit 616 determines that version upgrade is necessary if the software version of the unused disk node is not the new version. On the other hand, if the software version of the unused disk node is a new version, the update target determination unit 616 determines that version upgrade is unnecessary. If an upgrade is necessary, the process proceeds to step S44. If version upgrade is unnecessary, the process proceeds to step S47.

  [Step S44] The update instruction unit 618 transmits an update instruction to an unused disk node. Specifically, the update target determination unit 616 acquires from the IP address table 613 a in the disk node IP address storage unit 613 the IP address of the disk node determined to be upgraded. Then, the update target determination unit 616 passes a software update request including the acquired IP address to the update instruction unit 618. The update instruction unit 618 outputs an update instruction with the IP address acquired from the update target determination unit 616 as a destination.

  As a result, the software update process is executed in the unused disk node. When the update process is completed, an update completion response is returned from the disk node.

  [Step S45] The update instruction unit 618 receives an update completion response from an unused disk node. Then, the update instruction unit 618 notifies the update target determination unit 616 that the software update process for the unused disk node has been completed.

  [Step S46] The update target determination unit 616 updates the version information table 611a in the version information storage unit 611. In other words, the update target determining unit 616 updates the software version of the disk node for which update processing has been completed to the latest version.

  [Step S47] The update target determination unit 616 searches for a disk node to be updated next. Specifically, the update target determination unit 616 refers to the connection information table 612a in the connection information storage unit 612, and extracts another disk node connected to the same storage device as the disk node for which the update process has been completed. . Then, the update target determination unit 616 sets an unprocessed disk node that has not been detected in step S42 among the extracted disk nodes as a disk node to be updated next.

  [Step S48] The connection switching instruction unit 617 instructs the disk node to be updated next to invalidate the connection. Specifically, the update target determination unit 616 acquires the IP address of the disk node to be updated next from the IP address table 613 a in the disk node IP address storage unit 613. Then, the update target determination unit 616 passes a connection relation invalidation request including the acquired IP address to the connection switching instruction unit 617. The connection switching instruction unit 617 outputs a connection invalidation instruction with the IP address acquired from the update target determination unit 616 as the destination.

  As a result, in the disk node to be updated next, the currently valid connection is changed to an invalid state. When the invalidation processing is completed, a response indicating the completion of invalidation is returned from the disk node.

  [Step S49] The connection switching instruction unit 617 receives the invalidation completion response from the disk node to be updated next. Then, the connection switching instruction unit 617 notifies the update target determination unit 616 that the connection invalidation processing of the disk node to be updated next is completed.

  [Step S50] The update target determination unit 616 updates the connection information table 612a in the connection information storage unit 612. Specifically, the update target determination unit 616 sets a valid connection corresponding to the disk node subjected to the connection invalidation process to a blank (no valid connection).

  [Step S51] The connection switching instruction unit 617 instructs the unused disk node to validate the connection. Specifically, the update target determination unit 616 acquires the IP address of an unconnected disk node from the IP address table 613 a in the disk node IP address storage unit 613. Also, the update target determination unit 616 refers to the connection information table 612a of the connection information storage unit 612, and determines the storage devices connected to both the unused disk node and the disk node to be updated next.

  Further, the update target determination unit 616 acquires the port number of the connection port connected to the storage device connected to both of the unused disk nodes as the port number to be validated. Then, the update target determination unit 616 passes the connection relationship validation request including the acquired IP address and the validation target port number to the connection switching instruction unit 617. The connection switching instruction unit 617 outputs a connection activation instruction that includes the IP address acquired from the update target determination unit 616 as the destination and includes the port number to be validated as data.

  As a result, in the unused disk node, the currently invalid connection is changed to the valid state. When the activation process is completed, an activation completion response is returned from the disk node.

  [Step S52] The connection switching instruction unit 617 receives a validation completion response from a disk node that has not been used. Then, the connection switching instructing unit 617 notifies the update target determining unit 616 that the connection node enabling process for the unused disk node has been completed.

  [Step S53] The update target determination unit 616 updates the connection information table 612a in the connection information storage unit 612. Specifically, the update target discriminating unit 616 sets the port number of the valid connection port in the valid connection column corresponding to the disk node that has undergone the connection validation process. Thereafter, the process proceeds to step S41.

Next, the function of the disk node for performing software update processing and connection validation / invalidation processing in accordance with instructions from the management node 600 will be described in detail.
FIG. 17 is a block diagram showing detailed functions of the disk node. FIG. 17 shows the SW update unit 240, the connected storage management unit 260, and functions in more detail.

The SW update unit 240 includes an update software storage unit 241, an update execution unit 242, and an update completion notification unit 243.
The update software storage unit 241 has a storage function for storing software to be updated. For example, a part of the storage area of the HDD built in the disk node 200 is used as the update software storage unit 241.

  Upon receiving an update instruction from the management node 600, the update execution unit 242 acquires update target software from the update software storage unit 241, and performs update processing of the OS and the like. Specifically, when receiving an update instruction, the update execution unit 242 determines whether or not a new version of software stored in the copy source software storage unit 614 of the management node 600 is stored in the update software storage unit 241. Judging. If not stored, the update execution unit 242 acquires the new version of software from the copy source software storage unit 614 of the management node 600 and stores it in the update software storage unit 241. Thereafter, the update execution unit 242 executes update processing.

  When the update execution unit 242 executes the update process, the data providing service in the disk node 200 is temporarily stopped. For example, the OS is restarted during the update process. During the OS suspension period, the data access unit 220 stops providing data in the storage apparatus.

When the update processing by the update execution unit 242 is completed, the update completion notification unit 243 transmits an update completion notification to the management node 600.
The connection storage management unit 260 includes two device connection units 261 and 262, a connection switching unit 263, and a switching completion notification unit 264.

  The device connection unit 261 is connected to one host connection unit 815 of the two host connection units 814 and 815 provided in the storage apparatus 810. The device connection unit 262 is connected to one host connection unit 824 of the two host connection units 824 and 825 provided in the storage apparatus 820. The host connection units 814, 815, 824, and 825 of the storage apparatuses 810 and 820 are interfaces for connecting the storage apparatuses 810 and 820 to a computer. Connection port numbers are allocated to the device connection units 261 and 262, respectively. In the example of FIG. 17, the connection port number of the device connection unit 261 is “1”, and the connection port number of the device connection unit 262 is “2”.

  The device connection units 261 and 262 have a connection flag indicating validity / invalidity of the connection relationship. The value of the connection flag is set by the connection switching unit 263. If the connection relationship is valid, the device connection units 261 and 262 perform data access in the connected storage apparatus in response to an access request from the data access unit 220 or the like. If the connection relationship is invalid, the device connection units 261 and 262 do not access the storage apparatus and return an error in response to an access request from the data access unit 220.

  The connection switching unit 263 switches the storage device that is effectively connected in accordance with an instruction from the connection switching instruction unit 617 of the management node 600. Specifically, upon receiving an activation request designating a connection port from the connection switching instruction unit 617, the connection switching unit 263 sets the connection flag in the device connection unit of the corresponding connection port to “valid”. When the connection switching unit 263 receives an invalidation request designating a connection port from the connection switching instruction unit 617, the connection switching unit 263 sets the connection flag in the device connection unit of the corresponding connection port to “invalid”.

The switching completion notification unit 264 transmits a switching completion notification to the management node 600 when the connection validation or invalidation processing by the connection switching unit 263 is completed.
The data access unit 220 can access data in the storage device via a device connection unit that is enabled for connection. In addition, the data management unit 230 manages data in the storage apparatus via a device connection unit that is enabled for connection. For example, when one device connection unit is invalidated and the other device connection unit is validated, the data management unit 230 reads the metadata from the storage device connected to the validated device connection unit, and the metadata Store in the storage unit 250.

Next, software update processing and connection switching processing in the disk node 200 will be described in detail.
FIG. 18 is a flowchart showing the procedure of the software update process. In the following, the process illustrated in FIG. 18 will be described in order of step number.

[Step S61] The update execution unit 242 of the SW update unit 240 receives a software update instruction from the management node 600.
[Step S62] The update execution unit 242 determines whether or not a new version of software indicated by the update instruction is stored in the update software storage unit 241. If the new version of software is not stored, the update execution unit 242 requests the management node 600 for the new version of software, and acquires the new version of software stored in the copy source software storage unit 614. Then, the update execution unit 242 stores the acquired new version of software in the update software storage unit 241.

  [Step S <b> 63] The update execution unit 242 uses the software stored in the update software storage unit 241 to update the software in the disk node 200. If the system needs to be restarted during the update process, all functions in the disk node 200 are stopped by the OS. For this reason, the data providing service to the access node 700 by the data access unit 220 is also stopped. When the OS is started first, each function in the disk node 200 shown in FIG. 17 is also restarted by the OS.

  [Step S64] When the update completion notification unit 243 completes the software update by the update execution unit 242, the update completion notification unit 243 transmits an update completion response to the management node 600. Thereafter, the software update process ends.

FIG. 19 is a flowchart illustrating a procedure of connection switching processing. In the following, the process illustrated in FIG. 19 will be described in order of step number.
[Step S71] The connection switching unit 263 receives from the management node 600 a connection switching instruction for instructing invalidation or validation of the connection.

  [Step S72] The connection switching unit 263 determines whether or not the connection switching instruction is an invalidation instruction. If it is an invalidation instruction, the process proceeds to step S73. If it is an activation instruction, the process proceeds to step S76.

  [Step S73] In the case of an invalidation instruction, the connection switching unit 263 sets the state of the currently connected connection port to invalid. As a result, the connection between the device connection unit corresponding to the invalid connection port and the storage apparatus is blocked.

  [Step S74] The data management unit 230 recognizes that there is no storage device to be managed due to invalidation of the connection, and discards the metadata in the metadata storage unit 250.

[Step S <b> 75] The switching completion notification unit 264 transmits an invalidation completion response to the management node 600. Thereafter, the process ends.
[Step S76] In the case of an activation instruction, the connection switching unit 263 sets the state of the connection port corresponding to the port number included in the activation instruction to be valid. Thereby, the connection between the device connection unit corresponding to the valid connection port and the storage apparatus is established.

  [Step S77] The data management unit 230 recognizes that the storage device is newly managed as a result of the connection being validated, and reads metadata from the storage device having a valid connection relationship. Then, the data management unit 230 stores the read metadata in the metadata storage unit 250.

[Step S78] The switching completion notification unit 264 transmits an activation completion response to the management node 600. Thereafter, the process ends.
As described above, the software of each of the disk nodes 100, 200, 300, and 400 can be updated while continuing the service of the multi-node storage system. In addition, when the connection to the storage apparatus is switched, it is possible to change the disk node to which the access request is output only by changing the IP address in the metadata. Therefore, the period during which the storage device cannot be accessed due to connection switching is short.

  By the way, in the multi-node storage system, the control node 500 has a function of monitoring the state of each of the disk nodes 100, 200, 300, and 400 and changing the allocation of slices to virtual disks when the state changes. For example, when any failure occurs in one of the disk nodes 100, 200, 300, and 400, the control node 500 performs a recovery process. The recovery process is a process of allocating a new slice to a segment whose duplex state has collapsed due to the occurrence of a storage device that cannot be accessed due to a failure, and recovering the duplex state. As a part of such monitoring processing, the control node 500 has a function of collecting metadata from the disk nodes 100, 200, 300, and 400 in operation. If this metadata collection function is used, the access node 700 can be made aware of the change of the data access destination when the connection is switched without performing the IP address changing process.

  FIG. 20 is a sequence diagram illustrating the procedure of the software update process. In addition, each process of step S81-S90 shown in FIG. 20 is the same as the process of step S11-S20 shown in FIG. 11, respectively. Each process of steps S93 to S96 shown in FIG. 20 is the same as the process of steps S21 to S24 shown in FIG. Each process of steps S101 to S108 shown in FIG. 20 is the same as the process of steps S27 to S34 shown in FIG. Accordingly, processing different from the processing in FIG. 11 among the processing shown in FIG. 20 will be described below.

First, the processing of steps S91 to S92 will be described along with step numbers.
[Step S <b> 91] The data management unit 330 of the disk node 300 that has completed connection invalidation transmits a connection disconnection notification of the storage device 830 to the control node 500.

  [Step S92] Upon receiving the disconnection notification, the virtual disk management unit 510 of the control node 500 stores the metadata corresponding to the disconnected storage device in the virtual disk metadata table in the virtual disk metadata storage unit 520. Delete from 521. Then, the virtual disk management unit 510 transmits a connection disconnection notification confirmation response to the disk node 300.

Next, the process of step S97-step S100 is demonstrated along a step number.
[Step S97] The data management unit 430 of the disk node 400 that has completed the connection activation transmits an activation notification to the control node 500.

[Step S98] The virtual disk management unit 510 of the control node 500 transmits a confirmation response to the activation notification to the disk node 400.
[Step S99] The data management unit 430 of the disk node 400 reads the metadata from the storage device 830 and transmits it to the control node 500.

  [Step S100] The virtual disk management unit 510 of the control node 500 associates the metadata sent from the disk node 400 with the identification information (for example, IP address) of the disk node 400, and stores it in the virtual disk metadata storage unit 520. Are registered in the virtual disk metadata table 521. Then, the virtual disk management unit 510 transmits a confirmation response indicating that the metadata has been received to the disk node 400.

  By reconfiguring the virtual disk metadata table 521 of the control node 500 in this way, thereafter, in response to a metadata request from the access node 700, metadata indicating the connection relationship after connection switching is returned to the access node 700. can do. That is, the connection switching result can be recognized by the access node 700 using the metadata collection function in the control node 500. The metadata collection function in the control node 500 is used for new disk node addition processing, recovery processing in the event of a disk node failure, and the like (for example, in the disk node or control node). Processing priority is set high). Therefore, by using the metadata collection function, it is possible to safely and reliably switch the disk node that manages the storage apparatus.

  The above processing functions can be realized by a computer. In that case, a program describing processing contents of functions that the control node 500, the management node 600, and the disk nodes 100, 200, 300, and 400 should have is provided. By executing the program on a computer, the above processing functions are realized on the computer. The program describing the processing contents can be recorded on a computer-readable recording medium. Examples of the computer-readable recording medium include a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory. Examples of the magnetic recording device include a hard disk device (HDD), a flexible disk (FD), and a magnetic tape. Optical disks include DVD (Digital Versatile Disc), DVD-RAM, CD-ROM (Compact Disc Read Only Memory), CD-R (Recordable) / RW (ReWritable), and the like. Magneto-optical recording media include MO (Magneto-Optical disc).

  When distributing the program, for example, a portable recording medium such as a DVD or a CD-ROM in which the program is recorded is sold. It is also possible to store the program in a storage device of a server computer and transfer the program from the server computer to another computer via a network.

  The computer that executes the program stores, for example, the program recorded on the portable recording medium or the program transferred from the server computer in its own storage device. Then, the computer reads the program from its own storage device and executes processing according to the program. The computer can also read the program directly from the portable recording medium and execute processing according to the program. Further, each time the program is transferred from the server computer, the computer can sequentially execute processing according to the received program.

The present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the gist of the present invention.
The main technical features of the embodiment described above are as follows.

(Supplementary Note 1) A plurality of storage devices and a plurality of disk nodes larger in number than the storage devices are provided, each of the storage devices is connected to at least two disk nodes, and only one connection is the storage device. It is defined as an effective connection relationship used for data input / output in the device, and one or two storage devices are connected to the disk node, and are connected to a plurality of the disk nodes. An update control program for causing a computer to control software update processing of a plurality of the disk nodes,
The computer,
Reference is made to connection information storage means for storing connection relationships between the plurality of storage devices and the plurality of disk nodes, and the effective connection relationships, and the disk nodes that do not have the effective connection relationships are updated disk When the update of the software of the disk node to be updated is completed by selecting as a node, the disk node that has not been updated among the disk nodes connected to the same storage apparatus as the disk node to be updated is next Update target selection means for determining as a selected disk node and selecting the next selected disk node as the update target disk node after the connection relationship of the next selected disk node is invalidated;
Update instruction means for instructing the selected update target disk node to update the software each time the update target disk node is selected by the update target selection means;
When the next selected disk node is determined, an instruction to switch the effective connection of the storage device that is effectively connected to the next selected disk node from the next selected disk node to the update target disk node is performed. A connection switching instruction means for outputting to the disk node and the update target disk node;
Connection information updating means for updating information on valid connections in the connection information storage means according to the connection relation validated or invalidated in accordance with an instruction from the connection switching instruction means;
Update control program to function as.

(Supplementary note 2)
When the connection relationship between the update target disk node and the storage device is validated according to the instruction from the connection switching instruction means, access for accessing the storage device connected by the validated connection relationship Functioning as a connection disk node change notification means for notifying an access node that accesses the storage apparatus that the request destination has been changed to the update target disk node connected by an activated connection relationship The update control program according to supplementary note 1 characterized by the above.

  (Supplementary Note 3) The storage switching instruction means refers to address storage means for storing address information for uniquely identifying each of the disk nodes on the network, and the storage connected by an activated connection relation The update control program according to appendix 2, characterized in that the access node is notified of the address of the disk node to be updated that is connected by an enabled connection relationship as a destination of an access request for accessing the device. .

  (Supplementary Note 4) The connection switching instruction means is connected by an activated connection relationship via a control node that provides address information of the disk node that is effectively connected to the storage device to the access node. The update control program according to appendix 3, wherein the address of the disk node to be updated is notified to the access node.

(Supplementary Note 5) The update target selecting means refers to a version information storage means for storing the version of the software to be updated installed in a plurality of the disk nodes after selecting the update target disk node, and If the version of the software installed in the target disk node is the same as the version after the update process, determine the next selected disk node without waiting for the completion of the update of the software of the target disk node,
The update instructing unit instructs the update target disk node to update the software only when the version of the software installed in the update target disk node is different from the version after the update process. The update control program according to attachment 1.

  (Supplementary Note 6) The connection switching instruction means transmits a valid connection invalidation instruction to the next selected disk node, and the connection relation with the next selected disk node is related to the update target disk node. The update control program according to supplementary note 1, wherein an instruction for validating a connection relationship with the invalidated storage apparatus is transmitted.

(Supplementary Note 7) A plurality of storage devices and a plurality of disk nodes larger in number than the storage devices are provided, and each of the storage devices is connected to at least two disk nodes, and only one connection is the storage device. It is defined as an effective connection relationship used for data input / output in the device, and one or two storage devices are connected to the disk node, and software update processing of a plurality of the disk nodes An update control device for controlling
Reference is made to connection information storage means for storing connection relationships between the plurality of storage devices and the plurality of disk nodes, and the effective connection relationships, and the disk nodes that do not have the effective connection relationships are updated disk When the update of the software of the disk node to be updated is completed by selecting as a node, the disk node that has not been updated among the disk nodes connected to the same storage apparatus as the disk node to be updated is next Update target selecting means for determining as a selected disk node and selecting the next selected disk node as the update target disk node after the connection relationship of the next selected disk node is invalidated;
Update instruction means for instructing the selected update target disk node to update the software each time the update target disk node is selected by the update target selection means;
When the next selected disk node is determined, an instruction to switch the effective connection of the storage device that is effectively connected to the next selected disk node from the next selected disk node to the update target disk node is performed. Connection switching instruction means for outputting to the disk node and the update target disk node;
A connection information updating unit that updates information of a valid connection in the connection information storage unit according to a connection relationship that is validated or invalidated according to an instruction of the connection switching instruction unit;
An update control device.

(Supplementary Note 8) A plurality of storage devices and a plurality of disk nodes larger in number than the storage devices are provided, and each of the storage devices is connected to at least two disk nodes, and only one connection is the storage device. It is defined as an effective connection relationship used for data input / output in the device, and one or two storage devices are connected to the disk node, and are connected to a plurality of the disk nodes. An update control method for causing a computer to control software update processing of a plurality of the disk nodes,
The computer is
Reference is made to connection information storage means for storing connection relationships between the plurality of storage devices and the plurality of disk nodes, and the effective connection relationships, and the disk nodes that do not have the effective connection relationships are updated disk Select it as a node,
Each time the update target disk node is selected, the update target disk node is instructed to update the software,
When the update of the software of the update target disk node is completed, among the disk nodes connected to the same storage apparatus as the update target disk node, the disk node that has not been updated is determined as the next selected disk node And
When the next selected disk node is determined, an instruction to switch the effective connection of the storage device that is effectively connected to the next selected disk node from the next selected disk node to the update target disk node is performed. Output to the disk node and the update target disk node,
According to the connection relation that has been validated or invalidated, update the information of the valid connection in the connection information storage means,
After the connection relationship of the next selected disk node is invalidated, the next selected disk node is selected as the update target disk node.
An update control method characterized by the above.

It is a figure which shows the outline | summary of embodiment. It is a figure which shows the structural example of the multinode storage system of this Embodiment. It is a figure which shows the hardware structural example of the management node used for this Embodiment. It is a 1st figure which shows the state transition in a software update process. It is a 2nd figure which shows the state transition in a software update process. It is a figure which shows the data structure of a virtual disk. It is a block diagram which shows the function of each node. It is a figure which shows the example of a data structure of a storage apparatus. It is a figure which shows the example of a data structure of a metadata memory | storage part. It is a figure which shows the example of a data structure of a virtual disk metadata storage part. It is a sequence diagram which shows the procedure of a software update process. It is a block diagram which shows the detailed function of an update control part. It is a figure which shows the example of a data structure of a version information storage part. It is a figure which shows the example of a data structure of a connection information storage part. It is a figure which shows the example of a data structure of a disk node IP address memory | storage part. It is a flowchart which shows the procedure of a system update process. It is a block diagram which shows the detailed function of a disk node. It is a flowchart which shows the procedure of a software update process. It is a flowchart which shows the procedure of a connection switching process. It is a sequence diagram which shows the procedure of a software update process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Update control apparatus 1a Connection information storage means 1b Update object selection means 1c Update instruction means 1d Connection switching instruction means 1e Connection information update means 2a, 2b,..., 2m, 2n Storage devices 3a, 3b, 3c,. , 3m, 3n, 3o disk nodes

Claims (7)

A plurality of storage devices and a plurality of disk nodes more than the storage device are provided, each of the storage devices is connected to at least two disk nodes, and only one connection is data in the storage device Is defined as an effective connection relationship used for input / output of the disk, and one or two storage devices are connected to the disk node, and a plurality of computers connected to the plurality of disk nodes are connected to the disk node. An update control program for controlling the software update process of the disk node,
The computer,
Reference is made to connection information storage means for storing connection relationships between the plurality of storage devices and the plurality of disk nodes, and the effective connection relationships, and the disk nodes that do not have the effective connection relationships are updated disk When the update of the software of the disk node to be updated is completed by selecting as a node, the disk node that has not been updated among the disk nodes connected to the same storage apparatus as the disk node to be updated is next Update target selection means for determining as a selected disk node and selecting the next selected disk node as the update target disk node after the connection relationship of the next selected disk node is invalidated;
Update instruction means for instructing the selected update target disk node to update the software each time the update target disk node is selected by the update target selection means;
When the next selected disk node is determined, an instruction to switch the effective connection of the storage device that is effectively connected to the next selected disk node from the next selected disk node to the update target disk node is performed. A connection switching instruction means for outputting to the disk node and the update target disk node;
Connection information updating means for updating information on valid connections in the connection information storage means according to the connection relation validated or invalidated in accordance with an instruction from the connection switching instruction means;
Update control program to function as. Said computer further
When the connection relationship between the update target disk node and the storage device is validated according to the instruction from the connection switching instruction means, access for accessing the storage device connected by the validated connection relationship Functioning as a connection disk node change notification means for notifying an access node that accesses the storage apparatus that the request destination has been changed to the update target disk node connected by an activated connection relationship The update control program according to claim 1.   The connection switching instructing unit refers to an address storage unit that stores address information for uniquely identifying each of the disk nodes on the network, and accesses the storage device connected by the validated connection relation. 3. The update control program according to claim 2, wherein the access node is notified of an address of the update target disk node connected by an enabled connection relationship as a transmission destination of an access request for the access. The update target selecting means refers to version information storage means for storing the version of the software to be updated installed in a plurality of the disk nodes after selecting the update target disk node. If the version of the installed software is the same as the version after the update process, the next selected disk node is determined without waiting for the completion of the update of the software of the disk node to be updated,
The update instructing unit instructs the update target disk node to update the software only when the version of the software installed in the update target disk node is different from the version after the update process. The update control program according to claim 1.   The connection switching instruction means transmits a valid connection invalidation instruction to the next selected disk node, and the connection relationship with the next selected disk node is invalidated to the update target disk node. The update control program according to claim 1, wherein an instruction to validate the connection relationship with the storage apparatus is transmitted. A plurality of storage devices and a plurality of disk nodes more than the storage device are provided, each of the storage devices is connected to at least two disk nodes, and only one connection is data in the storage device An update that is defined as an effective connection relation used for input / output of a disk, and one or two storage devices are connected to the disk node, and controls software update processing of a plurality of the disk nodes. A control device,
Reference is made to connection information storage means for storing connection relationships between the plurality of storage devices and the plurality of disk nodes, and the effective connection relationships, and the disk nodes that do not have the effective connection relationships are updated disk When the update of the software of the disk node to be updated is completed by selecting as a node, the disk node that has not been updated among the disk nodes connected to the same storage apparatus as the disk node to be updated is next Update target selecting means for determining as a selected disk node and selecting the next selected disk node as the update target disk node after the connection relationship of the next selected disk node is invalidated;
Update instruction means for instructing the selected update target disk node to update the software each time the update target disk node is selected by the update target selection means;
When the next selected disk node is determined, an instruction to switch the effective connection of the storage device that is effectively connected to the next selected disk node from the next selected disk node to the update target disk node is performed. Connection switching instruction means for outputting to the disk node and the update target disk node;
A connection information updating unit that updates information of a valid connection in the connection information storage unit according to a connection relationship that is validated or invalidated according to an instruction of the connection switching instruction unit;
An update control device. A plurality of storage devices and a plurality of disk nodes more than the storage device are provided, each of the storage devices is connected to at least two disk nodes, and only one connection is data in the storage device Is defined as an effective connection relationship used for input / output of the disk, and one or two storage devices are connected to the disk node, and a plurality of computers connected to the plurality of disk nodes are connected to the disk node. An update control method for controlling a software update process of the disk node,
The computer is
Reference is made to connection information storage means for storing connection relationships between the plurality of storage devices and the plurality of disk nodes, and the effective connection relationships, and the disk nodes that do not have the effective connection relationships are updated disk Select it as a node,
Each time the update target disk node is selected, the update target disk node is instructed to update the software,
When the update of the software of the update target disk node is completed, among the disk nodes connected to the same storage apparatus as the update target disk node, the disk node that has not been updated is determined as the next selected disk node And
When the next selected disk node is determined, an instruction to switch the effective connection of the storage device that is effectively connected to the next selected disk node from the next selected disk node to the update target disk node is performed. Output to the disk node and the update target disk node,
According to the connection relation that has been validated or invalidated, update the information of the valid connection in the connection information storage means,
After the connection relationship of the next selected disk node is invalidated, the next selected disk node is selected as the update target disk node.
An update control method characterized by the above.

Images