JPH10254638A - Storage device system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent performance decline at the time of a R/W device fault in a storage device utilizing a DVD and taking redundant constitution. SOLUTION: At the time of mounting a physical medium group 1860 to a physical R/W device group 1800 provided with a fault physical R/W device 1810, a physical data medium 1820 is given priority over a physical parity medium 1850 and is mounted. The physical medium group 1860 not to be written in the physical R/W device group 1800 provided with the fault physical R/W device 1810 is mounted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

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

[0002]

2. Description of the Related Art In a computer system, as a known example closest to the invention, the following Patterso is shown.
n papers are known.

[0003] A. C. M. Sig Modco
Conference Proceedings, 1988, 6
Moon, page 109-116 (D. Patterson, et.
al: ACase for Redundant Arrays of Inexpensive Disks
(RAID), ACM SIGMOD conference proceeding, Chicago,
IL, June 1-3, 1988, pp. 109-116) The article by Patterson discloses a technique relating to data arrangement on a disk array. The disk array is a mechanism for realizing high performance and high reliability of the disk system. In a disk array, in order to improve performance, physically a plurality of disk devices appear as one disk device to a processing device. On the other hand, for higher reliability, when a failure occurs in a disk device storing data, redundant data for recovering data is stored in another disk device. Normally, data which is a unit of read / write of a disk device is called a record, but in Patterson's paper, several record arrangement methods are proposed. However, when a disk array is used, a record which is a read / write unit viewed from the processing device may have a different data length from a record actually recorded on the disk device. Hereinafter, the former is called a logical record and the latter is called a physical record.

[0004] Hereinafter, some record arrangement methods proposed in Patterson's paper will be described. In the first arrangement method, the logical records, that is, the records viewed from the processing device side, are m (m ≧ m) on the disk device.
This is an arrangement method of dividing into physical records and storing the divided physical records. Hereinafter, this arrangement method is referred to as a division arrangement method. When the divided arrangement is used, one logical record can be transferred to and from m disk devices, so that the same effect as apparently increasing the data transfer speed by m times can be obtained.

Next, a method of creating redundant data in a divided arrangement will be described. In the divided arrangement, n (n ≧ 1) redundant data are created for m physical records obtained by dividing a logical record, and each of the redundant data is stored in a disk device as one physical record (there are n in total). I do. Less than,
A physical record that stores data that is directly read / written by the processing device is called a data record, and a physical record that stores redundant data is called a parity record. A group including m data records and n parity records is called a parity group. Normally, if the number of parity records in a parity group is n, n
Even if a failure occurs in up to disk units, the data of the parity group can be recovered. The second arrangement method is an arrangement method in which a logical record, which is a read / write unit viewed from the processing device, is stored as one physical record, that is, one data record, on a disk device. Hereinafter, this is referred to as a non-divided arrangement. Therefore,
Logical records are equivalent to data records. (Since a data record or a parity record is assigned to each physical record, a physical record and a logical record are not necessarily equivalent. That is, one logical record is one physical record, but one physical record is (It is not necessarily one logical record but may be a parity record.) The feature of the non-divided arrangement is that the read / write processing can be executed for each disk device constituting the disk array. . (If the division arrangement method is adopted, it is necessary to occupy a plurality of disk devices for reading / writing.)
Therefore, by adopting the non-divided arrangement, it is possible to improve the multiplicity of the read / write processing that can be executed in the disk array, thereby realizing an improvement in performance. Even in the non-divided arrangement, n parity records are created from m data records and stored in the disk device. However, in the case of the split arrangement, a set of data records in the parity group forms one logical record as viewed from the processing device, whereas in the non-split arrangement, each of the data records is viewed from the processing device. It becomes a completely independent logical record. In a computer system, a magnetic tape, an optical storage device, and the like are often used as storage devices other than the disk device. Particularly recently, DVD (Digit
alVideoDisk) is attracting attention. These storage devices are characterized by a storage medium and R / W (Rea
d / Write) device is separated, a storage medium is loaded into an arbitrary R / W device, and data on the storage medium is read / written. Generally, these media are referred to as portable media. In a large-scale computer system, a library is introduced to easily manage a very large number of portable media. The library includes, in addition to the storage medium and the R / W device, a storage for storing a large number of storage media, a robot for transferring the storage medium between the storage and the R / W device, and the like. . Since the data handled by the computer system is becoming larger and larger, the need for improving its availability is very high. Therefore, it is effective to realize high availability by applying the concept proposed in Patterson's paper even in the storage device system configured by the portable medium as described above.

[0006] As a technology in which such a concept is applied to a portable medium, Comdex 96: DVD application (Alan E. Bell (IBM Research Division): DVD App)
lications, COMDEX 96, Nov. 20, 1996). In this document, RAI having redundancy by combining a plurality of ordinary libraries composed of DVD, R / W device, robot, etc.
L (Redundant Arrays of Inexpensive Libralies) has been proposed.

[0007]

SUMMARY OF THE INVENTION Using a DVD, a DVD
Consider a library device in which media and read / write devices have a redundant configuration. If a parity group is composed of m data records and n parity records, these records will be distributed over m + n DVD media, so that m + n units are stored inside the library device for efficient access. And at least one set of R / W devices.

If one of the m + n R / W devices fails, m + n media cannot be set there. Therefore, if an arbitrary m + n-1 media are set, and a read request for a data record included in a media that has not been set is received,
A total of m data records and parity records constituting a parity group including a data record to be read are read out to restore data stored in the data record to be read, or one medium is removed from the R / W device. It is necessary to set the medium including the data record for which the read request has been made in the R / W device and to read the data record for which the request has been made, as compared with reading the data record normally. There is a problem that it takes a long time.

Similarly, if one of the m + n R / W devices fails, m + n media cannot be set there. Then, if any m + n-1 media are set, a write request comes to a data record of m media of the m + n media (for example, a sequential write request to m continuous data records comes). At this time, in order to obtain consistency of the (m + n) parity groups, after writing to (m + n-1) media, the media that has not been set must be set in the R / W device and written to this media. However, there is a problem that it takes a very long time to replace the medium.

Alternatively, in order to prevent performance degradation, it is possible to give up consistency of m + n parity groups and to achieve consistency of only m + n-1 parity groups. However, in this case, redundancy is reduced. There is a problem that reliability is reduced.

A first object of the present invention is to provide a technique that does not reduce the time required for the read / write processing even when the R / W device breaks down.

A second object of the present invention is to provide a technique that does not lower the reliability during write processing even if the R / W device breaks down.

[0013]

In order to achieve the first object of the present invention, a medium storing a data record is preferentially given to an R / W device without mixing a data record and a parity record in the medium. set. In this way, all data records that can be read-requested are R
Since it is on the medium mounted on the / W device, there is no need to reconstruct the target data record from the parity record and other data records, so that it is possible to prevent performance degradation. In addition, when a write request comes to a group of R / W devices including a failed R / W device, the parity record can be prevented from being updated, so that the performance does not decrease.

In order to achieve the second object of the present invention, a group of media to which write requests do not frequently come is set to a group of R / W devices including a failed R / W device. However, in this case, it is assumed that there are two or more sets of R / W device groups. There are, for example, the following two methods for recognizing that a write request does not come often.

The mount request issued by the host computer is divided into two types: read-only and writable. When a read-only mount comes, no write processing will be performed on the mounted media thereafter.

A counter is provided on the medium for the read processing and the write processing that have been issued so far for this medium. When the number of times of the write process is smaller than a certain ratio as compared with the number of times of the read process, it is possible to determine that the possibility of writing is low.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION

(1) First Embodiment The first embodiment relates to a case in which when a medium is mounted, whether or not a write request is issued to the medium is declared.

FIG. 1 shows a configuration example of a storage system 1010 to which the present invention is applied. Storage system 1010
Is a controller 1020, a large number of media 1110
, A plurality of physical R / W devices 1810, a media transport robot 1090 that transports media 1110 between the rack 1120 and the physical R / W device 1810, and a robot rail 1100. In this embodiment, the medium 1110 is assumed to be a DVDRAM medium, and the physical R / W device 1810 is assumed to be a readable / writable DVDR / W device. However, the present invention is not limited to a DVD as a matter of course. Controller 1
020, a host interface 1030 for exchanging with the host computer 1000, a drive interface 1070 for exchanging with the physical R / W device 1810, a CPU 1040 for executing a control program,
It comprises a memory 1050 that stores a control program and also functions as a data buffer, and a robot interface 1060 that controls the media transport robot 1090.

The storage system 1010 has a disk array structure of the non-divided arrangement described in the prior art, and a parity group is constituted by three data records and one parity record. Also, three physical data media 1820 having only data records and one physical parity medium 1850 having only parity records are collectively referred to as a physical media group 1860. The three data records and one parity record that make up the parity group are arranged on three physical data media 1820 and one physical parity medium 1850 that make up the physical media group 1860, respectively.

In the present embodiment, one parity record is provided for three data records. However, it is needless to say that the present invention can be applied to another configuration.

FIG. 2 shows the correspondence between the logical media 1830 and the actual physical media as viewed from the host computer 1000. From the figure, the logical media 1830 is 3
It corresponds to a physical media group 1860 composed of one physical data medium 1820 and one physical parity medium 1850. The address space of the logical medium 1830 is a combination of the address spaces of the three physical data media 1820. When accessing, the number of the logical medium 1830 and the internal logical address 2
000, which physical data media 18
Which physical address 2010 on the device 20 should be accessed is uniquely determined. Here, the association between the logical medium 1830 and the plurality of physical media 2020 is performed with reference to a logical physical medium conversion table 1200 described later. The logical address 2000 indicates which physical data medium 1820
Which physical address 2010 corresponds to the CPU 104 on the controller 1020 using a simple calculation formula.
0 is calculated.

In the storage system 1010, four physical R / W devices 1810 are collectively managed as a physical R / W device group 1800, and the host computer 1
Logical media 1830 visible from 000 (logically)
The logical R / W device 1840 to be stored is associated with the logical physical R / W device group conversion table 1600 described later.

The host computer 1000 issues a mount request by designating the logical medium 1830 and the logical R / W device 1840. Logical media 1830 is logical R /
Mounting on the W device 1840 means that the logical media 18
The four physical media 2020 constituting the logical
/ W devices 1840 constituting four physical R / W devices 18
10 means to mount each. Similarly, the host computer 1000 is the logical R / W device 1
An unmount (unmount) request is issued by designating the physical media 2020 mounted on the four physical R / W devices 1810 constituting the logical R / W device 1840, respectively. Means to unmount.

FIG. 3 shows a logical-physical media conversion table 1200 for associating the logical media 1830 with the physical media 2020. The logical-physical media conversion table 1200 includes logical-physical media conversion table entries 1210 for the logical media 1830 that can be stored in the storage system 1010. The logical-physical media conversion table entry 1210 is a logical medium 1
The physical data media number a 1220, the physical data media number b 1230, and the physical data media number respectively store the numbers of the four physical media 2020 corresponding to the physical media group 1860 corresponding to 830.
c 1240, a physical parity media number 1250, a parity invalid flag 1260 indicating that the physical parity medium 1850 does not store correct data, and a physical R mounting the physical media group 1860.
/ W device group number 1270. In the present embodiment, the logical media 1830 and the physical media 20
The response of 20 shall not change forever. For this reason,
Physical data media number a 1220, physical data media number b 1230, physical data media number c 124
0 and the physical parity media number 1250 are not rewritten.

FIG. 4 shows a logical R / W device 1840 and a physical R / W device.
Logical / physical R to be associated with the W device group 1800
14 shows a / W device group conversion table 1600. As shown in the figure, the logical R / W device number 1620 changes to the physical R / W
The device group number 1270 can be obtained. Also,
The value -1 is stored in the entry of the logical R / W device 1840 not associated with the physical R / W device group 1800.

FIG. 5 shows a certain physical R / W device group 180.
0 physical R / W device group management table 14
00 and is composed of physical R / W device group management table entries 1410 of the physical R / W device group 1800 existing in the storage system 1010.

The physical R / W device group management table entry 1410 contains the physical R / W device group 1800
, The logical media number 1420 mounted on the logical R / W device 1840 corresponding to the logical R / W device 1840, and the four physical R / W devices 1810 constituting the physical R / W device group 1800
Physical R / W device entry a 143 corresponding to
0, physical R / W device entry b 1440, physical R / W device entry c 1450, physical R / W device entry d 14
60, and the number of failed R / W devices 1470 indicating the number of failed physical R / W devices 1810 out of the four physical R / W devices 1810. The physical R / W device group 180
When the logical medium 1830 is not mounted on 0, the value -1 is stored in the logical medium number 1420. Physical R / W device entry a 1430, physical R / W device entry b 1440, physical R / W device entry c 14
50, the physical R / W device entry d 1460, stores the number of the physical medium 2020 actually mounted on each physical R / W device 1810. During a failure, store the value -2.

FIG. 6 shows the flow of the mounting process 5000. When issuing the mount request, the host computer 1000 specifies the number of the logical medium 1830 and the number of the logical R / W device 1840. First, in step 5010, a mounting target physical R / W device group determination process 5 described later
Perform 200. In step 5020, it is checked whether or not there is a physical R / W device 1810 that is not mounted and has not failed in the physical R / W device group 1800 to be mounted. If not, the process jumps to step 5010. If present, the found physical R / W device 1810
Is a physical R / W device 1810 to be mounted. Then, in step 5040, the logical media 18 to be mounted
It is checked whether or not any of the physical data media 1820 constituting the 30 has not been mounted. If not, the parity invalid flag 126 of the logical / physical media conversion table entry 1210 is checked in step 5050.
At 0, it is checked whether the parity is valid. If it is invalid, the mount processing 5000 ends. If the parity is valid, it is checked in step 5060 whether or not there is a physical R / W device 1810 that is not mounted and has not failed in the physical R / W device group 1800 to be mounted. Ends. If so, in step 5070, the found physical R / W
The physical parity medium 1850 is mounted on the device 1810. In step 5080, it is checked whether or not the mount has been established. If the mount has been established, in step 5090, the number of the physical parity medium 1850 is stored in the physical R / W device entry 2030 corresponding to the physical R / W device 1810. Then, the mount processing 5000 ends.

In step 5020, the physical R to be mounted
If it is determined that there is no physical R / W device 1810 that is not mounted and has not failed in the / W device group 1800, all media 1110 mounted by the current mount processing 5000 are unmounted in step 5150. Then, the set table is returned to its original state, and the routine jumps to step 5010.

If it is determined in step 5080 that the mounting has failed, in step 5130 the found physical R / W device 1810 is determined to have failed, and the number of failed R / W devices in the physical R / W device group 1800 is 1470. To the physical R / W device 1810 corresponding to the physical R / W device 1810.
The value-2 is stored in the W device entry 2030. And
In step 5140, the physical R / W device group 180
It is checked whether the number of failed R / W devices 1470 of 0 has become 2 or more, and if so, the process jumps to step 5150. If the number of failed R / W devices 1470 is one or less, the process jumps to step 5020.

In step 5040, a physical data medium 1820 constituting the logical medium 1830 to be mounted
If it is determined that any of the media has not been mounted yet, the physical data medium 1820 found in step 5100 is replaced with the physical R / W device 18 to be mounted.
Mount on 10. Then, in step 5110, it is checked whether or not the mount has been completed. If the mount has failed, the process jumps to step 5130.
Then, the number of the mounted medium 1110 is stored in the physical R / W device entry 2030 corresponding to the physical R / W device 1810, and the process jumps to step 5020.

Thus, the physical parity medium 1
Since the physical data medium 1820 is mounted with priority over 850, the failed physical R / W device 1
If there is (only one) 810, only the physical data medium 1820 will be mounted.

FIG. 7 shows a flow of the mounting target physical R / W device group determination processing 5200. First, step 521
If it is 0, it is determined whether or not it is a read-only mount. If not, the process jumps to step 5240. If it is a read-only mount, in step 5220 the physical R / W
Check the device group management table 1400, and check the failure R / W
The number of devices 1470 is value 1 and the logical media number 142
The physical R / W device group 1800 in which the value -1 is stored in 0, that is, not mounted, is searched. And
It is determined whether or not it is step 5230, and if it is, jump to step 5260. If not, the physical R / W device group management table 1400 is checked in step 5240, and the physical R / W device group in which the number of failed R / W devices 1470 is 0 and the value -1 is stored in the logical media number 1420 is stored. Look for 1800.

Then, it is determined whether or not step 5250 has been reached, and if not, the process ends with an error. If there is, at step 5260, the found physical R / W
Step 5 sets the device group 1800 as a mount target.
At 270, the logical / physical R / W device group conversion table 1
Logical R / W device number 16 for which 600 mount requests were made
The number of the physical R / W device group 1800 to be mounted is set in the entry corresponding to No. 20. Then, in step 5280, the physical R / W device group 1 to be mounted in the physical R / W device group management table 1400
Logical media number 1420 of the entry corresponding to 800
Is set to the number of the logical medium 1830 that has made the mount request.

As described above, when the mount is a read-only mount, the physical R / W device group 1800 including the failed physical R / W device 1810 is preferentially mounted, and the writable mount is used. Sometimes, the physical R / W device group 1800 including the failed physical R / W device 1810 is not targeted for mounting.

FIG. 8 shows the flow of the unmount processing 5300. When issuing the unmount request, the host computer 1000 specifies the number of the logical R / W device 1840. First, in step 5310, the logical / physical R / W device group conversion table 1600 is checked, and the requested logical
Physical R / W device group 1 corresponding to / W device 1840
Get 800 number. Next, in step 5320, the physical R / W device group 1800 obtained in step 5310
, The physical R / W device group management table entry 1410 to be unmounted is determined. Then, in step 5330, the logical media number 142
Store the value -1 in 0. Next, in step 5340, a search is made for a physical R / W device entry 2030 in which a value other than value-1 or value-2 is stored. Step 5350 if any
In step 5360, the medium 1110 is unmounted from the physical R / W device 1810, and the value -1 is stored in the physical R / W device entry 2030 in step 5360.
Jump to zero.

If it is determined that step 5340 is not satisfied,
The logical R / W device 1840 requested in the logical / physical R / W device group conversion table 1600 in step 5370.
The value -1 is stored in the entry corresponding to the number, and the unmount processing 5300 ends.

FIG. 9 shows the flow of the read processing 5400.
The host computer 1000 is a logical R / W device 184
A read request is issued by designating a number of 0, a logical address, and a transfer length. First, at step 5410, the logical R /
The number of the W device 1840 is converted into the number of the physical R / W device group 1800 by looking at the logical physical R / W device group conversion table 1600. Next, at step 5420,
From the entry of the number of the specified logical medium 1830 in the logical / physical medium conversion table 1200 and the specified logical address, the physical medium 202 to be read is specified.
The number of 0 and the physical address to be read are obtained. Then, in step 5430, the physical R / W obtained in step 5410
The four physical Rs of the physical R / W device group management table entry 1410 corresponding to the number of the device group 1800
It checks which of the / W device entries 2030 stores the number of the physical medium 2020 to be read, and determines that it is the physical R / W device 1810 to be read. Then, in step 5440, the physical address to be read obtained in step 5420 is read from the physical R / W device 1810 to be read, and
A read request is issued by designating the requested transfer length.
Further, read data is obtained in step 5450, and transferred to the host computer 1000 in step 5470.

FIG. 10 shows the flow of the write processing 5500. The host computer 1000 is a logical R / W device 1
A write request is issued by designating the number of 840, the logical address, and the transfer length. First, the number of the logical R / W device 1840 specified in step 5510 is converted into the number of the physical R / W device group 1800 by referring to the logical / physical R / W device group conversion table 1600. Next, step 5520
The physical medium 2 to be written is obtained from the entry of the number of the specified logical medium 1830 in the logical / physical medium conversion table 1200 and the specified logical address.
The number of 020 and the physical address to be written are obtained.

Then, in step 5530, step 55
Physical R / W device group management table entry 1 corresponding to the number of physical R / W device group 1800 obtained in step 10
It is checked which of the four physical R / W device entries 2030 of 410 stores the number of the physical medium 2020 to be written, and the physical R / W device 1810 to be written is checked.
And In step 5540, the corresponding logical media number 1420 is obtained by looking at the entry of the number of the physical R / W device group 1800 obtained in step 5510 in the physical R / W device group management table 1400, and the corresponding logical media conversion table 1200 is obtained. It is determined whether the parity invalid flag 1260 of the entry to be turned on is on. If it is determined that the parity invalid flag 1260 is off, then in step 5550, step 5 of the physical R / W device group management table 1400 is performed.
It is determined whether the number of failed R / W devices 1470 of the entry of the number of the physical R / W device group 1800 obtained in 510 is 0. If this is 0, it means that the physical parity medium 1850 is mounted. If the number of failed R / W devices 1470 is 0, step 55
At 60, the physical R / W device 1810 to be written
The data previously written at the physical address to be written obtained in step 5520
The data is read out on 0, and is set as old data 2040.

Next, at step 5570, the physical R /
Step 551 of W device group management table 1400
The physical R / W device 1810 corresponding to the physical R / W device entry d 1460 of the entry of the number of the physical R / W device group 1800 obtained at 0 is previously written to the physical address to be written obtained at step 5520 at the write address. The read data is read onto the memory 1050, and this is set as the old parity 2050. Then, in step 5580, an exclusive logical loop is formed between the data to be written received from the host computer 1000, the old data 2040, and the old parity 2050, and this is set as a new parity 2060. Further, in step 5590, the old parity 2
The new parity 2060 is written to the address of the physical R / W device 1810 where 050 is stored. Then, in step 5600, the physical R / W device 1 to be written
The data to be written received from the host computer 1000 is written to the physical address to be written at 810, and the write processing ends.

If it is determined in step 5540 that the parity invalid flag 1260 is on, then step 5
At 610, the logical-physical media conversion table 120
The parity invalid flag 1260 of the entry of the number of the logical medium 1830 for which the write request of 0 has been made is turned on, and the process jumps to step 5600.

At step 5550, the number of failed R / W devices is 14
If it is determined that 70 is not 0, step 561 is executed.
Jump to zero.

As described above, when the parity is invalid or the physical parity medium 1850 is not mounted, the access to the physical medium 2020 is only a quarter of the number of times, so that the write processing can be performed at high speed. Can be processed. However, the physical parity media 185
There is a problem that the parity is forcibly invalidated even when 0 is not mounted. Therefore, as much as possible, the physical R / W including the failed physical R / W device 1810
The device group 1800 wants to avoid performing write processing. In the present embodiment, the physical R / W device
Physical R / W device group 180 including the I / W device 1810
0 solves this problem by setting it as a mount target only in a read-only mount.

(2) Second Embodiment In the second embodiment, the host computer does not notify the presence / absence of writing on the medium to be mounted.
This is a case in which statistical information is used to make an independent decision.
Here, only the portions different from the first embodiment will be described.

In the second embodiment, the mount request issued by the host computer 1000 to the storage system 1010 does not have a read-only function, and is treated as all writable.

FIG. 11 shows a logical-physical media conversion table 1200 for associating the logical media 1830 with the physical media 2020. The logical-physical media conversion table 1200 includes logical-physical media conversion table entries B1210 for the logical media 1830 that can be stored in the storage system 1010. The logical-physical media conversion table entry 1210 stores the numbers of four physical media 2020 corresponding to the physical media group 1860 corresponding to a certain logical media 1830, respectively. , Physical data media number c 1240, physical parity media number 1250
A parity invalid flag 1260 indicating that the physical parity medium 1850 does not store correct data,
A physical R / W device group number 1270 on which the physical media group 1860 is mounted, a read count 1280 and a write count 1290 for the logical medium 1830 so far. In this embodiment, it is assumed that the correspondence between the logical medium 1830 and the physical medium 2020 does not change forever. Therefore, the physical data media number a 1220, the physical data media number b 1230,
The physical data media number c 1240 and the physical parity media number 1250 are not rewritten. The read count 1280 and the write count 1290 are used to calculate a read / write ratio to the logical medium 1830.

FIG. 12 shows the flow of the mounting target physical R / W device group determination processing B. First, step 5820
Then, the number of reads 1280 and the number of writes 1290 of the logical-physical medium conversion table entry B corresponding to the logical medium 1830 for which the mount request is made are checked to calculate the write ratio, and it is determined whether the write ratio is high. Here, the criterion is used to determine whether the write ratio is high or low depending on whether the number of times of writing is 1/10 or more of the number of times of reading. However, it goes without saying that the present invention can be applied regardless of how the number itself is changed. If it is determined that the write ratio is high, the process jumps to step 5880. If it is determined that the write ratio is not high, step 5840 is executed.
Then, the physical R / W device group management table 1400 is checked, and the number of failed R / W devices 1470 is value 1 and the value -1 is stored in the logical media number 1420, that is, the physical R / W not mounted is not mounted. Search for the device group 1800.

Then, it is determined whether or not step 5860 has been reached, and if so, the flow jumps to step 5920. If not, the physical R / W device group management table 1400 is checked in step 5880, and the physical R / W device group in which the number of failed R / W devices 1470 is 0 and the value -1 is stored in the logical media number 1420 is stored. 180
Look for 0. Then, it is determined whether or not step 5900 has been reached, and if not, an error is terminated. If there is, in step 5920, the found physical R / W device group 1800 is set as a mount target, and step 59
40, the logical / physical R / W device group conversion table 16
Logical R / W device number 162 for which a mount request of 00 has been issued
The number of the physical R / W device group 1800 to be mounted is set in the entry corresponding to 0. Then, in step 5960, the physical R / W device group management table 1
400 physical R / W device groups 18 to be mounted
Logical media number 1420 of the entry corresponding to 00
Is set to the number of the logical medium 1830 that has made the mount request.

As described above, when a mount request is issued to the logical medium 1830 whose write ratio is not high, the physical R / W device group 1800 including the failed physical R / W device 1810 is preferentially mounted. When a mount request is issued to the logical medium 1830 having a high write ratio, the physical R / W device group 1800 including the failed physical R / W device 1810 is not a mounting target.

FIG. 13 shows the flow of read processing B6200 and write processing B6260.

In read processing B 6200, first, read processing 5400 is performed in step 6220, and read processing is performed in step 622.
The number of reads 128 of the logical-physical media conversion table entry B corresponding to the logical medium 1830 mounted on the logical R / W device 1840 that has met the read request at 0
The value 1 is added to 0, and the read processing B6200 ends. Here, in order to determine the mounted logical medium 1830, the logical physical R / W device group management table 16
Looking at the entry corresponding to the logical R / W device 1840 that has requested the read of 00, the physical R / W device group number 16
10, the physical R / W device group management table 14
It is sufficient to see the logical media number 1420 of the entry corresponding to the physical R / W device group number 1610 obtained for 00.

In the write process B 6260, first, the write process 5509 is performed in step 6280, and
The number of writes 129 of the logical-physical medium conversion table entry B corresponding to the logical medium 1830 mounted on the logical R / W device 1840 that has received the read request at 0
The value 1 is added to 0, and the write processing B6260 ends. The method of finding the mounted logical medium 1830 is the same as that of the read processing B6200.

[0054]

According to the present invention, it is possible to prevent the performance of the read / write process from deteriorating when the R / W device fails.
A storage device having a redundant configuration using D can be realized.

Further, it is possible to realize a storage device using a DVD and having a redundant configuration, which can prevent a decrease in the reliability of the write process when the R / W device fails.

[Brief description of the drawings]

FIG. 1 shows a configuration example of a storage device system to which the present invention is applied.

FIG. 2 shows a correspondence relationship between a logical medium, a physical medium, a logical R / W device, and a physical R / W device.

FIG. 3 shows a logical-physical media conversion table in the first embodiment.

FIG. 4 shows a logical / physical R / W device group conversion table.

FIG. 5 shows a physical R / W device group management table.

FIG. 6 shows a flow of a mounting process.

FIG. 7 shows a physical R / W to be mounted in the first embodiment.
13 shows a device group determination process.

FIG. 8 shows a flow of an unmount process.

FIG. 9 shows a flow of a read process.

FIG. 10 shows a flow of a write process.

FIG. 11 shows a logical-physical media conversion table in the second embodiment.

FIG. 12 shows a physical R / R to be mounted in the second embodiment.
13 shows W device group determination processing B.

FIG. 13 shows a flow of a read process B and a write process B in the second embodiment.

[Explanation of symbols]

1000: host computer, 1010: storage system, 1020: controller, 1030: host interface, 1040: CPU, 1050: memory, 1060: robot interface, 107
0: Drive interface, 1090: Media transport robot, 1100: Rail for robot, 1110:
Media, 1120: rack, 1800: physical R / W device group, 1810: physical R / W device, 1820: physical data media, 1830: logical media, 184
0: Logical R / W device, 1850: Physical parity medium

Claims (5)

[Claims] 1. A storage device system having two or more read / write devices capable of reading / writing from / to a removable storage medium, comprising: at least one data storage storage medium for storing data; A storage device system, wherein when mounting one or more redundant data storage media for storing generated redundant data in the read / write device, the data storage media is preferentially mounted. 2. A storage device system having two or more read / write device groups each including two or more read / write devices capable of reading / writing to / from a removable storage medium, wherein the failed read / write device is provided. A storage device system, wherein a write request is not issued to the read / write device group. 3. A storage device system having two or more read / write device groups each including two or more read / write devices capable of reading / writing to / from a removable storage medium, wherein the failed read / write device is provided. A storage system, wherein a storage medium that is unlikely to be written is mounted on the read / write device group. 4. The storage device system according to claim 3, wherein a mount request from the host computer is read-only and writable to determine the storage medium having a low possibility of being written. Storage system characterized by being divided into: 5. The storage system according to claim 3, wherein the number of times the storage medium is read from the storage medium is determined in the storage medium in order to determine the storage medium that is unlikely to be written. A storage device system for storing the number of times of writing to a storage device.