JP4218636B2 - Storage device for storing portable media - Google Patents

Description

  The present invention relates to a highly available storage device system for portable media. In particular, the present invention relates to a storage device system in which each component has redundancy in order to improve availability.

(1) Description of RAID A disk array is a technology that realizes higher speed than a single disk by operating a plurality of disk devices (hereinafter simply referred to as disks) in parallel. However, when n disks are arranged, the failure probability deteriorates n times. Therefore, “RAID (Redundant Arrays of Inexpensive Disks)” is known as a technique for achieving both high speed and high reliability. RAID is detailed in Non-Patent Document 1.

  RAID realizes high-speed control by operating a plurality of disks in parallel, and by storing redundant data called parity in a specific disk called parity disk, it is possible to store data in one unit. This is a disk control method that can reproduce the data of the failed disk from the parity of the other disk and the parity disk even if the disk fails, and can improve the disk failure reliability.

RAID has levels 1 to 5 depending on the method of storing the parity. FIG. 2 shows a data layout diagram of a level 4 type RAID as an example of these. In this example, if there are 5 disks, 4 are data disks 0 to 3, and 1 is a parity disk, the data is in the order of disk 0, disk 1, disk 2, and disk 3 in a certain data block. Are distributed and stored in order. This data block unit is called a stripe, and this distributed control is called striping. The parity P0, which is redundant data, is generated by calculating the exclusive OR (hereinafter referred to as XOR) of the data D0-3 of the same stripe stored in the disks 0-3. That is,
P0 = D0 + D1 + D2 + D3 (formula a) where + indicates an XOR operation.

Also, when disk 0 storing D0 fails,
D0 = D1 + D2 + D3 + P0 (Formula b)
Thus, D0 of the failed disk 0 can be reproduced.

  Further, the data and parity stripe group in the same column for generating parity is referred to as a “parity group”. In the case of FIG. 2, D0, D1, D2, D3, and P0 form the same parity group. In addition to RAID 4, levels 3 and 5 generate parity in the same manner as in equation a above.

  FIG. 1 shows the configuration of a RAID device. 1 is a host computer, 2 is a RAID device, and both are connected by a host interface 11. The RAID device 2 is composed of RAID control means 3 for performing RAID control and a plurality of disk devices 210 to 214. The RAID control means 3 controls the plurality of disk devices 210 to 214, and the data arrangement shown in FIG. Read / write data. The RAID control means 3 controls the generation of the parity and the restoration of data of the failed disk device.

(2) Description of Portable Media (DVD) In a computer system, there are a magnetic tape storage device, an optical disk storage device, and the like as a storage device that is often used in addition to a disk device. The characteristics of these storage devices are that the storage medium (also referred to as storage medium) and the recording / reproducing device are separated, and the storage medium is loaded into an arbitrary recording / reproducing device, and data on the storage medium is read and written. That is the point. Generally, these media are called “portable media”. Hereinafter, it will be simply referred to as “media”. In a large-scale computer system, a library apparatus is introduced in order to easily manage a very large number of media. In addition to storage media and recording / playback devices, the library device includes a storage for storing a large number of storage media, and a transport device for transferring the storage media between the storage and the recording / playback device. .

  As one of portable media, there is a DVD (Digital Video Disk). DVD is a kind of optical disk. It is a portable medium with large capacity, low cost and rewritable characteristics, and was developed mainly for the purpose of application to digital video devices. However, DVDs can also be used in computers, and their application to various computers that take advantage of their characteristics, such as distribution of programs and data, backup of data, and replacement of magnetic disks, has attracted attention.

(3) Description of RAIL Since the data handled in the computer system is becoming larger and larger, the need for improving its availability is very high. Therefore, even in a storage device system composed of portable media as described above, it is effective to realize high availability by applying the concept as proposed in the Paper of Patterson.

  Non-patent document 2 is a technology that applies such a concept to portable media. This document proposes RAIL (Redundant Arrays of Inexpensive Libraries) in which redundancy is provided by combining a plurality of ordinary library devices composed of a DVD, a recording / reproducing unit, a conveying unit, and the like.

  RAIL is considered to have no redundancy because one library device that is a component of RAIL is a normal library device (Inexpensive Libraliy). The RAIL concept is not based on the idea of providing redundancy to individual means constituting the library device, but is that the library device itself is considered as one disk device and the unit of redundancy is a library device. That is, even if one library device fails, it is stored in the failure library device from the DVD data stored in other library devices in the same manner as (1) RAID (formula b). DVD data can be reproduced. As described above, RAIL is a technology that realizes high reliability of a library apparatus using portable media.

  In the present specification, for the sake of easy explanation, the above-described RAID4 type RAID is considered. Of course, the same applies to other RAID levels. Similar to RAID, data is divided into a fixed block size and distributed on different library media. The division unit at this time is “stripes” as in RAID, and the data stripes and parity stripes in the same column for generating parity are collectively referred to as “parity groups”. Since it is RAID 4, the parity is concentrated on the same medium, but the medium storing the parity is referred to as “parity medium”.

  A group of library devices that provide redundancy of library devices is referred to as a “RAIL group”. A media group constituting the ECC group is referred to as a “media group”.

"A Case for Redundant Arrays of Inexpensive Disks (RAID)"; In Proc. ACM SIGMOD, June 1988 (published by the University of California, Berkeley) Comdex 96: DVD application (Alan E. Bell (IBM Research Division): DVD Applications, COMDEX 96, Nov. 20, 1996)

■ Issues
(1) Media replacement at the time of library device replacement In the above RAIL device, the library device is a device that can store tens to hundreds of media, but once the library device is replaced, it is stored in the faulty library device. All media must be re-stored in the appropriate location of the new library device (the same location as the storage location in the failed library device). If the media storage position is wrong, a “media group” that is an ECC group configured with media stored in another library apparatus cannot be configured correctly. As a result, there is a problem that the consistency of data stored in the media group is lost, correct read / write cannot be performed, and data is lost.

(2) Media exchange while the RAIL device is online The host computer to which the RAIL device is connected is generally a computer called a server, and provides online services for the library in response to requests from a large number of client computers. Since such a server is required to have high availability such as 24-hour continuous operation, it is usually impossible to stop RAIL for media recovery even when a failure library device of RAIL fails. Therefore, it is a problem that the RAIL is also performed while the RAIL is online together with the replacement of the failed library device with the new library device.

■ Purpose
(1) First object The first object of the present invention is to provide a plurality of recording / playback apparatuses, a plurality of storage units for storing a plurality of media, at least one medium transport unit, and at least one or more units. In the redundant library device including the RAIL device having the media input means, when a failure occurs in any one storage means (the library device in the RAIL device), the failure storage means (in the RAIL device) It is necessary to exchange a large number of media in conjunction with the replacement of the faulty library device. At that time, a method for confirming the storage location of the media and a method for correcting it when it is erroneously stored are provided. Thus, data loss due to inconsistency of the “media group” due to erroneous storage is prevented.

(2) Second object The second object of the present invention is to check the storage position and correct processing at the time of media replacement of the first object. In other words, it is possible to provide a method in which the read / write service of the redundant library apparatus and the recovery process of the faulty storage means (the faulty library apparatus in the RAIL apparatus) can be performed in parallel.

  In a redundant library device connected to a host computer, a plurality of recording / reproducing means arranged on an array for recording / reproducing data on a plurality of portable media and a storage position capable of storing a plurality of portable media can be exchanged. A plurality of storage means arranged on the array, at least one input means for inputting portable media into the storage means, and transporting the portable media between the input means, the storage means, and the recording / reproducing means An “media group” that is an ECC group is composed of a plurality of media stored in a plurality of different storage means, including at least one transport unit and a controller that controls a plurality of recording / reproducing apparatuses to operate in parallel. Then, redundant data generated from data of other media is stored in at least one media in the same media group.

  The media also includes a media group information storage unit that stores identification numbers of media groups constituting the media group, storage unit numbers for storing media, and storage location numbers of storage units for storing media.

The controller controls a host connection means for connecting to a host computer, at least one storage device connection means for connecting a plurality of recording / reproducing means and transferring data to a medium, and a transport means. At least one transport device connecting means, parity data to be stored in the parity medium, redundant data operation means for regenerating the data of the failed media and the media to be stored in the failed library device, Central control means for controlling the entire controller, storage means for detecting failure in any one storage means, failure detection means, and media exchange when replacing the failure storage means with new storage means When referring to the media storage location number of the media group information stored in each media, Located comprises a media inspection means for controlling so as to house the media.

  With the above configuration, when a failure occurs in one of the arbitrary storage means constituting the redundant library device, the controller detects the failure storage means, closes the failure storage means, and media of other normal storage devices The operation shifts to the degenerate operation in which the read / write processing is performed while reproducing the data of the failure storage means from the data and the parity.

  Thereafter, the user replaces the failure storage means with a new storage means.

  Thereafter, the user inputs one of the media stored in the failure storage unit into the input unit.

  The controller transports the inserted media, loads it into one of the plurality of recording / reproducing means that is not used for online processing from the host computer, and checks and checks the media storage position number stored in the media. The loaded media is transported and stored in the storage position indicated by the media storage position number, and it is confirmed whether or not all the media have been stored in the new storage means. Repeat the steps.

  When the storage of all the media is completed, the controller transports the media of the media group currently used for online processing from the host from the new storage means among the plurality of media stored in the new storage means, and performs online processing. Are loaded in unused recording / reproducing means, and the normal operation is restored from the degenerate operation.

  As described above, since the storage position information of the storage means in which the medium is to be stored is stored inside the medium, the controller can check it and store it in an appropriate position, so that the storage position is wrong. There is no.

  In addition, the above method has a problem in that a large number of media must be sequentially inserted manually, which increases the time required for humans to be present. Therefore, after all the media has been continuously inserted into the storage means, or the human has moved all the media of the fault storage means directly to the storage means of the new storage means and stored it, the controller sequentially It is possible to confirm the storage positions of all media. In this case, since the confirmation work does not involve any manual work, it is possible to save labor.

  In addition, if all the media are input at once as described above, there may be an error in the media position at the time of input due to human error. In this case, the media inspection unit of the controller rearranges the media to an appropriate position while referring to the storage position number of the media group information stored in the media. At this time, if another medium is already stored when trying to be stored in the correct position, the conveying means is temporarily controlled to temporarily save the other medium, and then the medium is stored in the correct position. Since the place where the current media was originally stored is vacant in the save place, the place may be used, or another temporary storage place may be prepared. The sorting method includes a heuristic method and a well-known sorting method such as “quick sort”. In this way, if all media are replaced at once, the storage location of all media can be inspected, and if it is wrong, it can be corrected, so that media can be stored in an appropriate location. it can.

  Further, since the degenerate operation is performed during the storage means exchange process, at least one recording / reproducing means is not used for the read / write process. Therefore, this recording / reproducing apparatus can be used for the inspection work of the storage position of the inserted medium as described above, and the redundant library apparatus performs “online processing” in which a read / write request from the host computer is processed, and fault storage means. And media replacement / confirmation / storage processing can be executed in parallel.

  According to the present invention, when a failure occurs in any one of the storage means constituting the redundant library device, when a large number of media stored in the failure storage means are replaced with new storage means, Checking the storage position of the media and correcting the position in the event of incorrect storage, it is possible to prevent data loss due to media group inconsistency due to incorrect storage. There is an effect that can be done.

  In addition, according to the present invention, the storage position confirmation and correction processing at the time of media replacement of the first effect can be performed while the redundant library device is online, so that an online service for performing read / write processing and fault storage means The recovery process can be performed in parallel.

  Furthermore, according to the present invention, it is possible to minimize the time for manually loading media in the process of exchanging media to the new storage means while realizing the first and second effects described above. effective.

  Furthermore, according to the present invention, the media group in the currently used media group is detected first in the media inspection to immediately restore the currently used media group and return from the degenerate operation to the normal operation. In addition, since it is possible to postpone confirmation of the storage position of other media and one correction, there is an effect that it is possible to shorten the degenerate operation time in which performance degradation occurs.

  An embodiment of the present invention will be described. The configuration of the redundant library apparatus according to the present embodiment assumes the above-described conventional RAIL apparatus for the sake of clarity. That is, the RAIL apparatus has a plurality of independent library devices arranged in an array, and each library device includes recording / reproducing means, transport means, storage means, and input means independently as shown in FIG. Yes. However, the present invention is not only applicable to such a RAIL apparatus, but also a general redundant library apparatus in which each of the recording / reproducing means, transport means, storage means, and input means has an arbitrary redundancy. It should be noted that this is also applicable to.

■ First embodiment (Method of inserting one sheet at a time from the slot)
(1) Configuration of RAIL Device A first embodiment will be described. FIG. 3 is a system configuration diagram of the present invention.

  1 is a host computer, and 5 is a RAIL device, which are connected by a host interface (hereinafter simply referred to as I / F) 11.

  In the RAIL device 5, 6 is a RAIL controller, and 400 to 404 are library devices. Recording / reproducing means (described later) of each library apparatus is connected to the RAIL controller 6 through recording / reproducing I / Fs 50 to 54. Further, the transport control means (described later) of each of the library apparatuses 400 to 404 is connected to the RAIL controller 6 by a transport control I / F 55.

(2) Configuration of Library Device FIG. 4 is a configuration diagram of the library device. 40 is a recording / reproducing unit that loads portable media and reads / writes data from / to this medium, and 41 stores portable media from the input unit 43 to the storage unit 42, or from the storage unit 42 to the recording / reproduction unit 40. Is a storage means for storing a plurality of portable media, 43 is an input means for storing the portable media from the outside in the library device 4, and 44 is a host Transport control means 45 for operating the transport means 41 in response to a command from the system to move the media or operating the input means, 45 is a DVD, CD-ROM, CD-R, MO (magneto-optical disk) Or other portable media (hereinafter simply referred to as media). The storage means 42 is provided with storage locations for n (integers greater than or equal to n) 0, and logical location numbers 0 to (n−1) that can uniquely identify locations. Is added. When the medium is moved to a specific storage position by the transport means 41, the logical number of this storage position is used. Hereinafter, this number is referred to as a “storage position number”.

  As described above, the recording / reproducing means 40 is connected to the RAIL controller 6 through the recording / reproducing I / F 50 (or 51 to 54). Further, the transport control means 44 is connected to the RAIL controller 6 by a transport control I / F 55.

  In the present embodiment, as shown in FIG. 4, an example in which the recording / reproducing means 40 is provided inside the library apparatus 4 is shown, but there is a configuration in which the recording / reproducing means 40 is separated from other means. The following description will be made with reference to FIG. 4, but the present invention can also be used in a library device having such a separation configuration.

  In the description of the RAIL apparatus, the symbols shown in FIG. 8 are used to indicate the respective constituent elements of the plurality of library apparatuses 400 to 404.

(3) Configuration of RAIL Controller FIG. 5 is a configuration diagram of the RAIL controller 6. 61 is a host connection means for receiving commands from the host computer and transferring data via the host I / F 11, and 640 to 644 are records of command transmission and data transfer to the recording / reproducing means 40 of the library devices 400 to 404 The storage device connection means 62, which is performed via the reproduction I / Fs 50 to 54, absorbs the difference in data transfer speed between the recording / reproduction means 40 of the library device and the host computer 1, and temporarily stores recently used data. Data storage means 63 configured as a disk cache that can be stored and transferred when the host computer 1 re-accesses the same data without operating the library devices 400 to 404. Redundant data (parity) to be stored on the media is generated, the media stored in the failed library device, the failed library device or media Redundant data calculation means for reproducing the data stored in the storage device 65, a transport device connection for issuing a command to the transport control means 44 via the transport control I / F 55 in order to control the transport means 41 of the library device 4 Means 60 is a central control means for controlling the means constituting the RAIL controller 6.

  In the present embodiment, the storage device connection means and each library device are connected in a one-to-one manner. However, even if n (n> 0 integer) library devices are connected to one storage device connection means. I do not care. Further, although the five library devices are connected to the transport device connecting means, the number of connected devices is arbitrary. Further, although the storage device connecting means and the transport device connecting means are described as different ones, they can be mixed in the same connecting means. As a suitable connection means for mixing, SCSI (Small Computer Systems Interface) is often used. Although the following description will be made based on FIG. 5, the present invention can be similarly used even when the configuration of the connecting means is different.

(4) Configuration of Central Control Unit As a preferred embodiment, the central control unit 6 can be realized using a microprocessor unit (MPU), its execution program, and a memory for storing programs and data. FIG. 6 is a block diagram of the central control means 6. 601 is a host connection control unit that controls the host connection means 61, and 602 is a command control unit that analyzes a command received from the host computer 1 and regenerates the command to the recording / reproducing device 40 of each library device 400 to 404 603, a data storage unit control unit for determining an area for storing data to be read / written by the same command in the data storage unit 62, and determining a hit / miss of the disk cache, and 604, redundant data (parity) at the time of writing ), Or the redundant data calculation control unit 605 for controlling the redundant data calculation means 63 when reproducing the data of the faulty media or the media stored in the fault library, the command control unit 602 generates It is a storage device connection control unit that controls the storage device connection means 640 to 644 in order to issue commands to the recording / reproducing means 40.

  Also, 608 is a transport control unit that controls the transport device connection means 65 to issue a command for operating the transport means 41 of each library device 400 to 404, and 607 is a failure in the library devices 400 to 404 609 is a media inspection unit that inspects the media so that the correct media is stored in the correct storage position, and 610 is the configuration of the RAIL device 5, the management information of the media, and the media group A configuration information holding unit 606 that holds various management information of the RAIL device 5 used by the central control unit 60, such as management information and failure information of the library device, is provided when a failure occurs in the library device or media. A fault communication unit that notifies the user of fault information or receives a fault recovery command from the user.

(5) Media Configuration FIG. 7 shows the configuration of the media 45 used in the RAIL apparatus of the present invention. Reference numeral 4500 denotes a media group information storage unit stored in the RAIL controller 5 in a specific area of the media. The part 4500 is provided at a specific logical address (for example, logical address 0) on the medium, and this area needs to be hidden from the host computer 1.

  The data content stored in the unit 45 includes a media group information identifier 4501 indicating that the media group information is stored, 4502 an identification number of the media group, 4503 a number of the library apparatus 4 storing the media, Reference numeral 4504 denotes a storage position number of the storage means 42 of the library apparatus 4, and reference numeral 4505 denotes data for managing information such as media initialization date and error information.

(6) Media Group Configuration FIG. 8 shows the media group configuration. Media 450 to 454 are stored in recording / reproducing means 4000 to 4004. An ECC group composed of a plurality of media for distributing data and a medium for storing the parity is referred to as a “media group”. In the RAID 4 configuration, media 450 to 453 for storing data are referred to as data media 0 to 4 and media 454 for storing parity are referred to as parity media.

  Similarly to the media stored in the recording / reproducing means 4000 to 4004, the media stored in the storage means of the library apparatuses 400 to 404 constitute a media group between different library apparatuses 400 to 404. A series of identification numbers is added to each media group, and this information is stored in the media group information storage unit 4500 of each media as described above. This media group number must be the only absolute number so as not to be confused with media used in other RAIL devices or other media groups previously used in this RAIL device. Therefore, it is desirable to generate the number by combining data such as the serial number of the RAIL device and the creation date and time of the media group in order to guarantee uniqueness.

  In this embodiment, a plurality of media constituting a media group are stored in the same storage position number of different library devices. That is, the media of the same media group are stored horizontally and in a line. Accordingly, when exchanging the media group to be mounted, there is an effect that the average time (average media group exchange time) for exchanging the media group by operating the transport unit is minimized. If a media group is stored at an arbitrary position, the media group exchange time is limited by the media that takes the longest time to transport, so the average media group exchange time deteriorates. In addition, this storage method also has an effect of facilitating media management by the RAIL controller. Of course, it is also possible for the RAIL controller 6 to manage and associate the relation with each other in any storage position.

(7) Initializing media (Registering media group information)
The operation of this embodiment will be described. First, an operation for storing a new medium in the library apparatus will be described. The user inputs new media into the input means 430 to 434 of each library apparatus 400 to 404. The user operates the host computer and issues a command for instructing initialization of the media. The host connection means 61 of the RAIL controller 6 receives this command and notifies the central processing means 60 of it. The host connection control unit 601 of the central processing unit 60 receives this command, and the command control unit 602 analyzes the command. Subsequently, the processing is transferred to the transport control unit 608, which issues commands to the transport control units 440 to 444 of the respective library devices 400 to 404 to move the media from the input unit 430 to 434 to the recording / reproducing unit 4000 to 4004. To do. Each of the library devices 400 to 404 receives the command, and sets the media in the recording / reproducing devices 4000 to 4004.

  Subsequently, the media inspection unit 609 of the central control means 60 commands the recording / playback devices 4000 to 4004 to read the media group information from the address of the media in the media group information storage unit via the storage device connection control unit 605. Issue. The read data is transferred from the recording / reproducing means 4000 to 4004 to the designated area of the data storing means 62 via the recording / reproducing I / F 50 to 54 and the storage device connecting means 640 to 644. The media inspection unit 609 inspects the read data. Since this media is a new media, the read data does not include the media group identification information 4501. The media inspection unit 609 recognizes that these media are new media, refers to the configuration information holding unit 610, and determines the storage positions of the media in the storage units 420 to 424 of the library apparatuses 400 to 404. This place may be an empty storage position. Next, the part 609 constructs a media group with these media, generates the media group identification number 4502 by the above method, and stores the media group identification number together with the determined storage position in the configuration information holding part 610. . Next, the unit 609 writes the media group information identifier 4101, media group number 4502, library device number 4503, library storage location number 4504, and other information 4505 via the storage device connection control unit 605. A command is issued, and media group information is stored in each medium. Subsequently, the same unit 609 writes 0 in the data storage unit of all media, thereby ensuring consistency between the data of the media group and the parity. This completes the initialization of the media, and notifies the host computer 1 of the end of the initialization instruction command.

  Thereafter, the host computer 1 can handle this media group as a logical volume, and executes the logical format of this logical volume. When the logical format is completed, the same media group can be used from the host computer 1.

  In the following description, the central control means 60 controls the transport means 41 of the library apparatus 4 to move the media between the recording / reproducing means 40, the storage means 42, and the input means 43. The method is the same as above. is there. Therefore, unless otherwise specified, description of the conveyance control is omitted as “conveyance control is performed”.

  In the following description, the central control means 60 issues a data read / write command to the medium to the recording / reproducing means 40 of the library apparatus 4, and the method is the same as described above. Therefore, the description of the recording / reproducing control is omitted as “read / write control is performed” unless otherwise specified.

(8) Initialization of Configuration Information Holding Unit When the RAIL device is turned on, nothing is recorded in the configuration information holding unit 610 of the central control means 60, so it is necessary to initialize it. Furthermore, it is necessary to confirm that the media is stored in the correct storage position while maintaining the relationship of the media group. This is because the media group configured between the library devices is restored to the original media storage position when the power is turned off and then restarted even if the media group maintained the correct positional relationship during the previous operation. This is because there is no guarantee. Since the media of the RAIL device is a portable medium as described above, the media may be moved by a human when the power is shut off.

  The media inspection unit 609 controls the transport control unit 608 to perform transport control, further controls the storage device connection control unit 605 to perform read / write control, and controls the storage means 420 to 424 of each library device 400 to 404. All the storage media are mounted sequentially, the media group information is read, and the information is stored in the configuration information holding unit 610. Here, “mount” refers to an operation of taking out media of a designated media group from the storage means of each library device and setting the media in the recording / reproducing device. When information is stored in the configuration information holding unit 610, it is confirmed that the media of different library devices stored in the same storage position are correctly the media of the same media group, and the information is stored in association with each other.

(9) Media Mounting Next, media mounting operation will be described. It is assumed that a media mount request is issued from the host computer 1. This mount request assumes that the host computer 1 manages the media group as a logical volume and is directly specified by the media group number.

  The host connection means 61 of the RAIL controller 6 receives the mount request command from the host computer 1 and notifies the central control means 60 of it. The central control means 60 obtains the command, analyzes it, and recognizes that it is a media group mount request. The central control means 60 performs transport control, and stores the media currently mounted on the recording / reproducing means 4000 to 4004 in the appropriate storage positions of the storage means 420 to 424 for the respective library devices 400 to 404. Subsequently, the central control means 60 performs conveyance control, takes out the media of the media group for which the mount request has been made from the storage means 420 to 424, moves to the recording / reproducing means, and mounts it. At this time, the storage position can be obtained by referring to the configuration information holding unit 610.

  Also, if the address space of the RAIL device looks flat from the host computer, that is, if it looks like one logical volume, the mount request may not be issued. In this case, it is necessary to store the address space visible to the host computer and the correspondence between each media group in the configuration information holding unit 610, recognize the address accessed by the command control unit 602, and mount a different media group as necessary. There may be. In this case, the mount request is generated inside the RAIL controller 6, but other than that is the same as the above.

(10) Read / Write Command Processing Next, the operation when a read / write request from the host to the mounted media group is issued will be described. When a read command is issued from the host computer 1, the command control unit 602 of the central control means 60 analyzes the command. Since data distribution as shown in FIG. 2 is performed, it is calculated in which medium the read request data from the host computer is stored. Such calculation is called address conversion. As a result of the address conversion, if D2 in FIG. 2 is access data, the data medium 2 stores the data. Therefore, the command control unit 602 generates a read command to the recording / reproducing device 4002 of the library device 402. Further, the data storage means control unit 603 reserves an area for storing the read data in the data storage means 62. The storage device connection control unit 605 issues the above read command to the storage device connection means 642, and the storage device connection means 642 transfers the data from the recording / playback device 4002 to the data storage means, and upon completion, the storage device connection means 642 sends the data to the central control means 60. Notice. The host connection control unit 601 of the central control means 60 issues a data transfer command to the host connection means 61, and the host connection means 61 transfers the data stored in the data storage means 62 to the host computer 1, and upon completion, the central control The means 60 is notified. The host connection control unit 601 of the central control unit 60 controls the host connection unit 61, transmits an end report to the host computer, and the read command processing ends.

  When a write command is issued from the host computer 1, there are two differences from the read command: the transfer direction is different and the parity is calculated. Similar to the read, when D2 in FIG. 2 is write data, the corresponding parity P1 also needs to be updated. Here, there are the following two methods of updating the parity: (Formula c) (Formula d).

New P0 = D0 + D1 + New D2 + D3 (Formula c)
However, + indicates an exclusive OR operation.

New P0 = old D2 + new D2 + old P0 (formula d)
In the method of (Formula c), it is necessary to read the data of D0, D1, and D3 to the data storage means 62 prior to the generation of parity. In the method of (formula d), it is necessary to read the old D2 and the old P0 in the same manner before generating the parity. The read control of these data is the same as the operation of the read command. If D0, D1, and D3 are already stored in the data storage means 62 (or the old D2 and the old P0 are stored), it is not necessary to perform these read controls.

  Regardless of the method (Equation c) or (Equation d), when all the data is prepared in the data storage means 62, the redundant data calculation control unit 604 activates the redundant data calculation means 63, The calculation of (Expression c) or (Expression d) is executed, and when it is finished, the central control means 60 is notified. The command control unit 602 of the central control means 60 generates a D2 write command and a new P0 write command, issues both write commands to the recording / reproducing means 4002 and 4004 as described above, and the storage device connection means 642 and 644 Both data are transferred from the data storage means 62 to the recording / reproducing means 4002 and 4004 and stored in the data medium 451 and the parity medium 454, respectively. When the transfer ends, the central control means 60 issues an end report to the host computer 1 as with the read command, and the write command processing ends.

(11) Failure Detection of Library Device Next, a case where a failure occurs in one library device 400 constituting the RAIL device 5 as shown in FIG. 8 will be described with reference to the flowchart of FIG. First, the failure of the library apparatus will be described. There are roughly two types of failure detection. The first is a case where the library apparatus itself detects a failure. The library itself detects that the recording / reproducing means 40 has become unreadable / writable or the conveying means 41 has become inoperable, and error information is detected. Is reported to the RAIL controller 6. The second is a case where the library apparatus stops operating and the central control means 60 detects it. In any case, when the library failure detection unit 607 of the central control means 60 recognizes these events, this library device is detected as a failure library (1000 in FIG. 9).

(12) Degeneration Operation The library failure detection unit 607 shifts the RAIL controller 6 to the degeneration operation (1001). Here, degenerate operation refers to reading / writing data of all media stored in a library device when a failure occurs in one library device, and data and parity of media in other library devices. It is the operation performed using. For example, at the time of a read request to the data D0 stored in the data medium 0 (450) of the library device 400, the data media 1 to 3 ( 451 to 453) and the data D1, D2, D3 stored in the parity media 454 and the parity P0 to D0 stored in the parity medium 454 are reproduced. When a write request for D0 is made, P0 is generated from D0 to be written and D1, D2, and D3, and only P0 is updated. The library failure detection unit 607 records the transition to the degenerate operation in the configuration information holding unit 610, and thereafter, the command control unit 602 performs address conversion for the degenerate operation as described above. In addition, the configuration information holding unit 610 records that all the media in the library device 4000 cannot be used.

(13) Failure Notification and Replacement of Library Device The library failure detection unit 607 controls the failure communication unit 606 to notify the host computer 1 that a failure has occurred in the library device 400 via the host connection means 61. At this time, a message prompting the user to replace the library device 400 'with a new one and an indicator are displayed to inform the user (1002). When the user replaces the library device, the host computer 1 notifies the RAIL controller 6 of the replacement completion, and the failure communication unit 606 notifies this to the library failure detection unit 607 (1003).

(14) Inserting Media Next, the same unit 607 passes the processing to the media inspection unit 609. The unit 609 controls the fault communication unit 606 to prompt the host computer 1 to take out one medium of the fault library device 400 and put it into the input means 430 ′ of the new library device 400 ′ (1004). Further, the media inspection unit 609 controls the conveyance, moves the conveyance unit 410 ′ to the input unit 430 ′, and waits (1005).

(15) Media inspection and storage The user waits for the media to be inserted (1006), and when the media is inserted, the media inspection unit 609 performs transport control and moves the inserted media to the recording / reproducing means 4000 ′. Mount (1007). The unit 609 performs read control, and reads the media group information stored in the medium to the data storage means 62 (1008). The media inspection unit 609 inspects the library device number 4503 of the same information to confirm that the media is the library device 400 (1009). If the media is in a different library device, the transport process is performed and the media is ejected (1010). If it is a correct medium of the library apparatus, the library storage position number 4004 is inspected (1011), and a carrying process is performed so that the medium is stored at the indicated position (1012). The media inspection unit 609 notifies the host computer 1 that the storage of one piece of media has been completed, and issues a notification that prompts an instruction whether to insert the next media or end the input processing (1013). If the next medium is to be inserted, the process returns to (14) Media insertion (1004) and the process is repeated until all the media have been stored.

(16) Return from degenerate operation to normal operation If a notification of completion of storage of all media is received from the host computer 1 (1014), the media inspection unit 609 returns the processing to the library failure detection unit 607. The library failure detection unit 607 confirms the currently mounted media group with reference to the configuration information holding unit 610, and mounts the media of the same media group on the new library device 4000 ′. Further, the same unit 607 recovers from the degenerate operation to the normal operation and records it in the configuration information holding unit 610. Further, the media stored in all the library devices are recorded in the usable state in the configuration information holding unit 610. Subsequently, the host computer 1 is notified of normal recovery, and the recovery from the library failure state is terminated (1016).

(17) Effect As described above, in the RAIL device, when a failure occurs in the library device, and the medium is replaced with a new library device, the media stored in all the failed library devices must be re-stored in the new library device. Although it is essential, according to the present embodiment, the medium can be stored in the correct storage position and can be restored to the state before the occurrence of the failure. This has the effect of preventing the loss of data caused by it.

  In addition, since the media of the same media group can always be stored in the same storage location number of the library apparatus, the average replacement time can be minimized when the media group is replaced, and high-speed transport performance can be realized.

  In addition, while one library device fails and the RAIL device provides services in a degenerate operation, the media of the failed library device can be stored in the correct position of the new library device in the background, so a highly available RAIL device There is an effect that can be realized.

■ Second Embodiment (Continuously Putting in Appropriate Order. Method for Device to Check and Sort Later)
Next, a second embodiment will be described. In the first embodiment, the operator inputs media one by one and each time the RAIL device stores the media in the correct position. Although it is possible to securely store it in the correct position, on the other hand, it is necessary for the operator to witness this work for a long time, and when the number of media stored in the library device reaches several hundreds, the required time becomes long. End up. The present embodiment provides a method for solving this problem.

  In the configuration of this embodiment, in addition to the configuration of the first example, a media storage management table 6101 shown in FIG. The media storage management table 6101 needs to be prepared according to the storage position that can be stored for each library device. When the power is turned on, all media are set to “empty”. Record. When a library device fails, all the tables of the library device are set to “free”.

  Only the difference in operation from the first embodiment will be described below.

(1) Replacement of Fault Library FIG. 10 is a flowchart showing the operation when a fault occurs in the library apparatus according to this embodiment, and will be described below with reference to this figure.

  The central control means 60 detects that a failure has occurred in the library device 400 (1000), shifts to degenerate operation (1001), notifies the user of the failure (1002), and the user uses a new library device 400 ′. Connection and replacement are completed (1003).

(2) Media exchange (sequential continuous input)
Next, the media inspection unit 609 of the central control means 60 initializes the counter n to 0 (1017). Then, the host computer 1 is instructed to insert a new medium via the failure communication unit 606, and the operator inserts one medium (1018). The media inspection unit 609 performs transport control, temporarily stores the inserted media in the storage position n, and sets “temporary storage” in the position n of the media storage management table 6101 (1019). The unit 609 requests confirmation through the fault communication unit 606 whether all media have been inserted (1020). If there is still media, the counter n is incremented (1021). Repeat the temporary storage. When this process ends, the number n of stored media is determined.

(3) Inspection of media storage management table Subsequently, the media inspection unit 609 initializes another counter k to 0 (1022). With reference to the media storage management table of the configuration information holding unit 610 shown in FIG. 11, it is confirmed whether the correct media has already been moved and stored in the storage position k (1023). In the case of k = 0, since no medium has been inspected yet, “temporary storage” is set in the media storage management table 6101, so this determination is false. In addition, when k = j (n ≧ j ≧ 1), there is a case where the movement described later has already been performed and k is “normal storage” when k <j. At this time, the process proceeds to 1035, k is incremented (1035), k <n is confirmed, and if k <n, the process returns to 1023 and repeats the process (3) (1036).

(4) Movement of media If the result of the inspection of the media storage management table as to whether the media has been moved is false, the media temporarily stored in the storage position k is transported and loaded into the recording / reproducing apparatus 4000 ′ (1024 ). Subsequently, the media inspection unit 609 of the central control unit 60 reads the media group information and confirms that the library storage position number 4504 is i (an integer of i ≧ 0) (1025).

  If this media should not be stored in this library device 400 ', or if it is a wrong media such as a new media, this media is transported and ejected (1027) “Free” is set at position k of the media storage management table (1028).

  If k = i (1029), the position is correct as it is, so that the same medium is stored in the storage position j by conveyance control (1030), and the position “k” (= i) of the medium storage management table 6101 is “ “Normal storage” is set (1031). Thereafter, the process proceeds to 1035, k is incremented (1035), k <n is confirmed, and if k <n, the process returns to 1023 and the processes of (3) and (4) are repeated (1036).

  If k ≠ i (1029), the conveyance control is performed, and the medium currently stored in the storage position i is temporarily moved to the storage position k and the position k in the medium storage management table is set to “temporary”. Set to “storage” (1032), the medium to be stored in the storage position j loaded in the recording / reproducing apparatus 4000 ′ is stored in the correct storage position i by the conveyance control (1032), and “ “Normal storage” is set (1034). Thereafter, the process returns to 1023 without incrementing k, and the above processes (3) and (4) are repeated until the correct medium is stored in the storage position k.

(5) Returning from the degenerate operation to the normal operation When the above processing is executed for all k <n, all media can be stored in the correct storage position. After that, as in the first embodiment, the media inspection unit 609 returns the processing to the library failure detection unit 607, and the library failure detection unit 607 confirms the currently mounted media group with reference to the configuration information holding unit 610. The media of the same media group is also mounted on the new library device 4000 ′. Further, the same unit 607 recovers from the degenerate operation to the normal operation and records it in the configuration information holding unit 610. Further, the media stored in all the library devices are recorded in the usable state in the configuration information holding unit 610. Subsequently, the host computer 1 is notified of normal recovery, and the recovery from the library failure state is terminated (1016).

  In the above description, the method of performing media movement processing (sorting processing) to the correct position using a heuristic method has been shown. However, as another method, first, the media group information of all media is read, After all the current storage positions and the regular storage positions of the media are recognized, the sorting process can be performed using an algorithm such as “quick sort” whose sorting is known.

(6) Effect As described above, according to the present embodiment, the same effect as that of the first embodiment can be realized while shortening the time for the operator to attend to all media inputs.

■ Third embodiment (Human hands can be used. The device will check and sort later)
In the second embodiment, a method in which all the media are sequentially inserted one by one by one person has been shown. However, this method can be performed without manpower as compared with the first embodiment, and the witness time can be shortened. However, it still has a problem that it takes a long time to attend because it has to throw in hundreds of media. Therefore, a method for further reducing the time for input will be described as a third embodiment.
(1) Method This method is a method in which all the media are once taken out from the library device when the failed library device 400 is removed, and stored offline in a new library device 400 ′ by a human. Thereafter, the new library apparatus 400 ′ is connected, and the sorting process is performed by performing the steps 1023 to 1036 of FIG.

  When humans manually store media, there is a risk that storage errors will occur. However, by using the sorting processing method of the second embodiment together, even if there is a storage error, it can be corrected later.

(2) Effects According to the third embodiment, it is possible to minimize the time required for manually inserting media into the new library apparatus, and to achieve the same effects as in the first and second embodiments. Can do.

Fourth Embodiment In the first to third embodiments described above, when media are executed together with replacement of a fault library device, it is confirmed that all media are stored in the correct storage position. It was described as returning from normal operation to degenerate operation.

  However, this method takes a long time to operate in a degenerate operation. Since the degenerate operation has a problem that the performance is deteriorated with respect to the normal operation, it is a problem to shorten the degenerate operation time of the first to third embodiments. This embodiment will explain the method.

  Although it is possible to apply to all of said 1st-3rd embodiment, hereafter, a difference is demonstrated based on 2nd embodiment typically.

  Of the media groups (referred to as current media groups) currently used for “online processing”, which is the processing of read / write requests from the host, at the stage where all the media is temporarily stored in the new library device Search the media of the new library device. Specifically, the media inspection unit 609 confirms the current media group number with reference to the configuration information holding unit 610, and determines the storage location of the media group from the media storage information of other library devices. Check. Subsequently, the conveyance control is performed, one medium at the confirmed storage position is taken out and loaded into the recording / reproducing means 4000 ′, the media group information is inspected, and it is confirmed whether the medium is in the current media group. If the confirmation can be made, the current media group has been restored, and at this point, the normal operation is restored from the degenerate operation.

  Also, if the loaded media is not the media of the current media group as a result of the inspection, the media temporarily stored in an arbitrary order is loaded into the recording / reproducing means 4000 ′ one by one and the media group information is inspected. This search process is repeated for all media until the current media group number is found, and when it is found, the normal operation is restored from the degenerate operation.

  Thereafter, when the read / write request from the host does not come for a certain period of time and the use frequency of the RAIL device is low, the reordering described in the second embodiment is performed by checking the media storage position information of the media using the recording / playback device of the library device Perform the process. Also, if there is an access request from the host during this rearrangement operation, the operation may temporarily shift to the degeneration operation, or the media storage location of the current media group is known. Therefore, this media may be mounted when accessed.

  This embodiment uses the fact that in many cases, humans sequentially insert media when storing the media, so there is a high possibility that the media are arranged in the same position as before. In addition, even if the order has changed, since the rearrangement is postponed, the processing time for restoring the current media group can be greatly shortened.

  In the above description, the second embodiment has been described, but the same applies to the third embodiment. In the first embodiment, if the user knows the current media group, the same effect can be obtained by storing the media first. Even if you do not know, if you return to normal operation from the degenerate operation when the media of the current media group is stored, the effect will be less than the second and third embodiments, but the first embodiment alone Compared with the case where it implements, there exists an effect that the average time of the degeneration operation time during the replacement | exchange operation | work of a medium can be reduced to a half.

  Further, in the above description, a method of rearranging is shown. However, when there is a high possibility that media are arranged in the same position as before, a method of not rearranging may be considered. In this case, each time the media group to be mounted is exchanged, it is confirmed whether or not the corresponding medium is stored in the correct position. If it is wrong, there is a demerit that you have to search all the worst media, but if there is a high possibility that it is correct, immediately after the temporary storage of the media is over, it will immediately go from degenerate operation to normal operation It has the effect of being able to return.

(5) Fifth Embodiment In the above-described embodiment, as described at the beginning of the description of the embodiment, the above-described conventional RAIL apparatus has been described. However, the present invention is not only applicable to such a RAIL apparatus, but also a general redundant library apparatus in which each of the recording / reproducing means, transport means, storage means, and input means has an arbitrary redundancy. It can also be applied.

  For example, there is an apparatus shown in FIG. 12 as such a redundant library apparatus. In this example, the redundant library apparatus includes one controller 6, six recording / reproducing units 4000 to 4005, three transport units 410 to 412, five storage units 420 to 424, and two units. Input means 430 to 431 are provided.

  For example, it is assumed that the RAIL group is composed of five of the six recording / reproducing apparatuses, and the other one is a spare recording / reproducing apparatus. Also, all three transport means can be transported from any storage means to any recording / reproducing means, and from any input means to any storage means. In addition, since all the recording / reproducing means, transport means, storage means, and input means have redundancy, even if a failure occurs in one of all means, the other can be used or degraded. The online service can be continued by driving.

  In such a redundant library apparatus, the first to fourth embodiments can be similarly applied. When a failure of an arbitrary storage means occurs, the medium to the correct storage position is replaced simultaneously with the replacement of the new storage means. Can be exchanged. However, when there are more units than the recording / reproducing means constituting the RAIL group as in this embodiment, it is not necessary to shift to the degenerate operation.

It is a block diagram of the RAID apparatus of a prior art. It is a block diagram of data striping in RAID and RAID. It is a block diagram of the RAIL apparatus of this invention. It is a block diagram of the library apparatus of embodiment of this invention. It is a block diagram of the RAIL controller of embodiment of this invention. It is a block diagram of the central control means of embodiment of this invention. FIG. 3 is a configuration diagram of media and media group information according to the embodiment of this invention. It is a block diagram of the media group of embodiment of this invention. It is a flowchart of 1st embodiment of this invention. It is a flowchart of 2nd embodiment of this invention. It is a block diagram of the media storage management table of 2nd embodiment of this invention. It is a block diagram of the redundant library apparatus of 5th embodiment of this invention.

Explanation of symbols

1: host computer, 2: RAID device, 3: RAID control means, 4: library device, 5: RAID device, 6: RAID controller, 40: recording / reproducing means, 41: transport control means, 42: storage means, 43: Input means 44: Transport control means 45: Media 60: Central control means 61: Host connection means 62: Data storage means 63: Redundant data calculation means 64: Storage device connection means 65: Transport Device connection means, 606: failure notification unit, 607: library failure detection unit, 608: transport control unit, 609: media inspection unit, 610: configuration information holding unit, 4500: media group information storage unit, 4501: media group information identifier , 4502: Media group identification number, 4503: Library device number, 4504: Library storage position number

Claims (7)

A plurality of recording / reproducing means for performing data recording / reproducing processing on portable media;
A plurality of storage means each storing a plurality of portable media;
One or more transport means for transporting portable media between the storage means and the recording / reproducing apparatus;
Control means for controlling the plurality of recording / reproducing means,
A media group is composed of a plurality of portable media each storing either data belonging to the same parity group or parity used for restoration of the data ,
The parity group is a group composed of the data and the parity.
The control means uses a recording / reproducing means to store in the portable medium a storage means identification number indicating a storage means in which the portable medium is stored, and a storage indicating a storage position of the portable medium in the storage means A storage system that records position information . The storage device system according to claim 1 ,
The storage system further includes means for recognizing a storage means identification number and storage position information recorded on the portable medium,
When the portable medium is inserted into the storage device system, the control means uses the conveying means to move the portable medium to the storage position indicated by the storage means identification number and the storage position information recorded on the portable medium. A storage system for storing portable media. A plurality of storage devices each having recording / reproducing means for reading and writing data to / from portable media and storage means for storing a plurality of portable media;
Control means for controlling the plurality of storage devices,
The control means has detection means,
A media group is composed of a plurality of portable media stored in storage units of different storage devices, and data and parity used for reproducing the data are different for different portable media belonging to the same media group. Recorded,
When a failure occurs in the first storage device, the control means detects the failure of the first storage device by the detection means and stores the first storage device in the portable medium of the first recording device . the readout process of the data to the data, by processing by restoring the first data using other stored data and parity to the portable media belonging to the same media group as movable transportable media, the control The means processes the output command of the data received from the computer,
When the first recording device is replaced with a second recording device and portable media of the first recording device is inserted,
The control means reads out and reads out the storage position information of the portable medium recorded in the loaded portable medium using the recording / reproducing means while processing the input / output command of the data received from the computer. A storage system that controls to store the portable medium in a predetermined storage position in the storage unit of the second storage device based on storage position information. The storage device system according to claim 3 ,
The portable media of the first recording device are loaded one by one,
The control means accepts input of the next portable medium after storing the one portable medium in the storage means based on the storage position information recorded on the portable medium,
After receiving a notification from the computer that a plurality of portable media of the first recording device has been inserted into the second storage device, the control means is a medium that is the target of a data input / output command. The portable media stored in the second storage device belonging to the group is loaded into the recording / reproducing means of the second storage device, and the portable media and other portable media belonging to the media group are used. A storage system that processes input / output commands received from a computer. The storage device system according to claim 3 ,
The control means has information holding means,
The control means stores the plurality of transportable media of the first storage device in the storage means of the second storage device in the order of input, and each of the plurality of transported media input is in the order of input. The stored information is stored in the information holding means,
Based on the information holding means, the portable media stored in the order of loading is determined, the portable media stored in the order of loading is read out using the recording / playback means, and the storage position information of the portable media is read out. Re-storing the portable medium in a regular storage position based on the position information, storing in the information holding means that the portable medium has been re-stored in a regular storage position;
Before the plurality of portable media in the first storage device that has been input are stored in the regular storage position, a read process for the first data stored in the input portable media is performed. The control means processes the output command of the data received from the computer by processing the first data using the data and parity stored in other portable media belonging to the same media group. ,
The inserted plurality of portable medium of the first storage device, after finishing housed in normal retracted position, said control means belongs to a media group that is the subject of the data input-output commands of the A storage device system for loading a portable medium stored in a second storage device into a recording / reproducing unit and processing an input / output command using the portable medium and another portable medium belonging to the media group. Recording / reproducing means for reading / writing data from / to portable media, storage means for storing a plurality of portable media, and transport means for transporting portable media between the storage means and the recording / reproducing means, A plurality of library devices, and control means for controlling the plurality of library devices,
A plurality of portable media belonging to different library devices constitutes a media group, and data and parity used for data recovery are recorded on different portable media belonging to the same media group,
The portable medium stored in the storage unit of the library apparatus includes identification information of the library apparatus in which the portable medium is stored, and a storage position indicating the storage position of the portable medium in the storage unit of the library apparatus And a storage device system in which information is recorded.
The storage device system according to claim 6 ,
When the portable medium is inserted into the storage device system, the control means stores the portable medium in a predetermined storage position based on the identification information and the storage position information of the library device recorded on the portable medium. The storage device system is controlled so as to be stored in the storage device.

Images