JP2001125879A - System and method for storing data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a data storage system assigning the memory device of the other network connected computer as a cache device. SOLUTION: An operation managing application 121 network-mounts cache disks 2b and 2c of a host computer (B) 1b and a host computer (C) 1c, and prepares cache device list information indicating the result of the network- mounting for ensuring as the cache device of an optical disk auto-changer 3. When performing access to the network cache, the cache capacity is obtained by referring to the cache device list information, and the physical address of the cache is calculated form a cache line index so that the pertinent cache device can be specified.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data storage system and a data storage method for a computer in which a magnetic disk device is assigned as a cache device of an optical disk autochanger, for example, and more particularly to a cache device for a memory device of another computer connected to a network. The present invention relates to a data storage system and a data storage method that can be assigned as a storage device.

[0002]

2. Description of the Related Art With the recent increase in the capacity of data storage media and the speed of data communication, various server systems for processing I / O requests from clients connected to a network have been developed. In a server system of this type, data is stored in an optical disk autochanger or the like that exchangeably stores a large number of optical disks, and a magnetic disk device is used as a cache device of the optical disk autochanger, so that large-capacity data with high access performance can be stored. The system has been realized.

[0003]

When the cache area becomes insufficient due to an increase in I / O requests, it is necessary to newly add a magnetic disk device to be assigned as a cache device of the optical disk autochanger.

[0004] However, the magnetic disk device assigned as a cache device of the optical disk autochanger has conventionally been intended only for a magnetic disk device in a computer on which a server system operates.
Even if there is a magnetic disk device that is not widely used by other computers connected to the network, this magnetic disk device cannot be assigned as a cache device of an optical disk autochanger. That is, there has been a problem that device resources of other computers connected to the network cannot be effectively used as a cache.

The present invention has been made in view of such circumstances, and provides a data storage system and a data storage method capable of allocating a memory device of another network-connected computer as a cache device. The purpose is to:

[0006]

In order to achieve the above-mentioned object, a data storage system according to the present invention allocates a memory device of another computer connected to a network to a cache memory as required, or conversely, The cache device is released from the cache device. For this purpose, at least one or more of the second computer connected to the network as a cache device of the first computer.
A data storage system that uses a memory device of the computer, wherein the first computer has a shortage of cache area, and a cache candidate specifying unit that specifies a memory device of the second computer as a candidate for a cache device. A cache selecting unit that selects any one of the memory devices specified by the cache candidate specifying unit; and a cache allocation that automatically allocates the memory device selected by the cache selecting unit as a cache device. Means for automatically releasing a memory device allocated from the cache device when the memory device allocated as a cache device by the cache allocation means becomes unnecessary. In which was to so that.

According to the present invention, when the cache area runs short, a memory device of another computer connected to the network is automatically allocated as a cache device. On the other hand, when the allocated memory device becomes unnecessary, Since this memory device is automatically released from the cache device, the memory device of another computer connected to the network can be effectively used. In addition, since the conflict between the transfer processing to the medium and the search / registration processing due to the shortage of the cache area can be avoided, it is possible to prevent the read / write performance from deteriorating due to the conflict and the throughput from decreasing due to the medium exchange.

[0008] Further, the data storage system of the present invention comprises:
It is preferable to include a means for measuring the response time of a memory device of another computer and hierarchizing the memory device in a multi-stage manner from the top in ascending order of the measured response time.

In the data storage system according to the present invention, it is possible to automatically construct a cache system optimal in performance by employing a hierarchical cache configuration.

[0010] The data storage system of the present invention further comprises:
It is preferable to have means for managing the reliability of the memory devices of the own computer and other computers, and arranging specific data specified in advance in the memory device considered to have the highest reliability.

In the data storage system of the present invention, for example, important data such as a cache management table for managing the entire cache can be consciously arranged in the most reliable memory device.
It is possible to increase the reliability of the entire data storage system.

[0012]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a connection form of a network system according to an embodiment of the present invention.

As shown in FIG. 1, in this network system, a plurality of computers including a host computer (A) 1a, a host computer (B) 1b, and a host computer (C) 1c are connected to each other via a network A. I have. This host computer (A) 1a,
The host computer (B) 1b and the host computer (C) 1c are computers connected to each other by a network protocol such as TCP / IP, and each has a magnetic disk device (cache (A) 2a, cache ( B) 2b and a cache (C) 2c).

The host computer (A) 1a is provided with an optical disk autochanger 3 for accommodating a large number of optical disks in an exchangeable manner. The host computer (A) 1a is provided with an I / O from a client computer (not shown). In response to a request, the optical disk autochanger 3 is accessed via a cache area to store data in the optical disk autochanger 3 or to read out the stored data. According to the present invention, not only the cache (A) 2a but also the cache (B) of the host computer (B) 1b connected via the network A as a cache area of the optical disk autochanger 3 of the host computer (A) 1a. 2b and the cache (C) 2c of the host computer (C) 1c can be appropriately utilized. This will be described in detail below.

FIG. 2 is a schematic configuration diagram of the network system. Focusing on the host computer (A) 1a, an internal bus B is laid on the host computer (A) 1a, and the CPU 11, the semiconductor memory 12, the network interface 13, and the storage device interface 14 are connected to the internal bus B. ing.

The CPU 11 controls the overall operation of the host computer (A) 1a, and operates each unit in the host computer (A) 1a in accordance with the description of various modules stored in the semiconductor memory 12.

The semiconductor memory 12 is a memory device serving as a main memory of the host computer (A) 1a.
It stores various modules executed by the CPU 11 and various processing data.

The network interface 13 controls data communication with another computer via the network A. Then, the storage device interface 14
Manages data input / output to / from the magnetic disk device 2a and the optical disk autochanger 3.

The CP stored in the semiconductor memory 12
As various modules executed by U11, in the host computer (A) 1a, the operation management application 121, the file system 122, the hierarchical storage control module 123, the cache device manager 1
24, a device driver 125 and a network manager 126.

The operation management application 121 is a module for managing a cache device, such as mounting a cache device connected to the network A and securing a cache area.

The file system 122 is located in an intermediate layer between various application modules and the device driver 125, and performs file management.

The hierarchical storage control module 123 is a module that performs data cache control in response to an I / O request from the file system 122.

The cache device manager 124
This is a service program for managing a cache device in accordance with an instruction from the operation management application 121 or the hierarchical storage control module 123.

The device driver 125 is a driver module for controlling I / O devices connected to the host computer (A) 1a.

Then, the network manager 126
Is a module for accessing a storage device connected to the network A.

First Embodiment First, as a first embodiment, an example in which a local storage device or a plurality of storage devices connected to a network are automatically assigned as cache devices as needed will be described.

In order to automatically allocate the local storage device or a plurality of storage devices connected to the network as cache devices as needed, it is necessary to show these storage devices as one cache disk. .

For this purpose, first, the operation management application 121 secures a cache area in the host computer (A) 1a, which is a local host, and in the host computer (B) 1b and the host computer (C) 1c.

For example, host computer (A) 1a
/ Mnt / localhost /
cache_a, and the host computer (B)
1b and the cache disks (2b, 2c) of the host computer (C) 1c are mounted on the network at / mnt / host_b and / mnt / host_c, and the cache area of the host computer (B) 1b is set to / mnt / host_b / cache_b. Change the cache area of the computer (C) 1c to / mnt /
Secured as a file of host_c / cache_c.

The optical disk autochanger 3 and the built-in drive connected to the host computer (A) 1a are mounted as device files such as / dev / acdev0 and mo_drv0 so that they can be accessed via the device driver 125. I do. It is also assumed that the cache disk (2a) connected to the host computer (A) 1a is also mounted on the device file. Then, the cache device list information and the mount information of the optical disc changer 3 and the like are stored in a file in the format shown in FIG. In the case of the first embodiment, the cache device list information is, for example, as shown in FIG.

Next, a method of accessing the network cache will be described. Here, for simplicity,
First, a description will be given on the assumption that each cache device is fixedly allocated.

When accessing the network cache, the cache capacity is acquired by referring to the cache device list information set by the operation management application 121.

For example, assuming that the number of cache lines of the caches 2a to 2c is n, the total number of cache lines is 3 × n lines.

As for the cache control, when the cache device is used as a simple buffer,
There are various cases where the access frequency is managed and the highly active data is left in the cache, but there is no particular description here.

When the cache is accessed using the cache line index as a key, the physical address to the cache is calculated by the following equations (1) and (2).

Cache device idx = cache line index / n (1) expression Block address = (cache line index mod n) * line_size (2) expression where line_size represents the cache line size (unit: block), Denotes integer division, mod denotes remainder of integer division, and * denotes multiplication. Also. The cache device idx indicates 0: cache 2a 1: cache 2b 2: cache 2c, and the block address indicates a block address from the beginning of each cache address.

When the cache device is accessed via the file system 122, the file pointer is calculated from the block address by the equation (3).

File pointer = block address * block size ... (3) In addition, the usage of the cache device is monitored by the hierarchical storage control module 123, and when the usage of the cache device is detected, the lower The data is ejected to the layer device (here, the optical disc autochanger 3).

FIG. 5 is an operation flow relating to cache control in the hierarchical storage control module 123.

First, the configuration of the cache device is checked with reference to the device information file created as described above (step A1). Next, an I / O request from the file system 122 is received (step A2), and a cache is searched from the access address (step A).
3).

If the cache hits, the physical address of the cache is accessed by equation (2). On the other hand, in the case of a cache miss, if there is a free area in the cache (YES in step A4), the cache access address is calculated (step A5), and the cache contents are updated (step A6). If there is no free space in the cache (NO in step A4),
The data is reflected on the medium of the optical disk autochanger 3 (step A9), and after the cache area is opened, the processing of steps A5 to A6 is executed.

Then, after the cache content update processing is completed, the cache counter indicating the cache usage is updated (step A7), and a reply of the processing end is returned to the file system 122 (step A8).

With the above processing, a plurality of storage devices including a network-connected storage device can be seen as one cache disk and accessed.

Next, an example in which a local storage device or a plurality of storage devices connected to a network are automatically assigned as cache devices as needed based on the above-described principle will be specifically described.

In the case of the network system having the connection form as shown in FIG.
At 21, a cache device information list having a format as shown in FIG. 6 is created. In FIG. 6, S indicates a device that is fixedly assigned to the cache in advance, and C
Indicates a device that is dynamically allocated as a cache depending on the state of the system during operation. U indicates that the device set to C is actually being used as a cache.

For example, when this cache device information list is created as shown in FIG.
ost: indicates that the device of / user / cache is fixedly allocated before the operation starts, and
host: / usr2 and host_b: / usr /
cache, host_c: indicates that the device of / usr / cache is dynamically allocated later.

In FIG. 7, the localhos
When the cache area of t: / user / cache (Chache0) runs short, localhost: / u
An attempt was made to secure sr2 as a cache, but it could not be secured (Chache-), and host_b: // as the next candidate
usr / cache is secured as a cache (Ch
ace1), followed by host_c: / us
This indicates that r / cache has been secured as a cache (Chase 2).

FIGS. 8 and 9 show an operation flow relating to the cache control of the first embodiment.

For example, if the cache area of this cache 0 runs short during the data registration / search operation using the cache device of cache 0 (NO in step B1 in FIG. 8), the hierarchical storage control module 1
23 requests the cache device manager 124 to allocate a cache device (step B in FIG. 8).
2).

The cache device manager 124
When this cache device allocation request is received (step C1 in FIG. 9), a cache device that can be dynamically allocated to the cache is searched by referring to the cache device information list (step C2 in FIG. 9).

First, a locally connected device is searched, and if there is no area that can be used as a cache, a device connected to the network is searched as a next candidate.

Here, the local device loc
alhost: / usr2 cannot be secured as a cache and is connected to host computer (B) 1b.
It is assumed that er / cache is detected as a cache candidate.

The cache device manager 124
/ User / cach of host computer (B) 1b
e directory on local disk to / mnt / hos
Mount on t_b (Step C4 in FIG. 9) and Cach
Assign to e1.

Next, the cache device manager 12
No. 4 checks the cache capacity of / mnt / host_b (step C5 in FIG. 9), secures a cache area up to 0 × 100,000 bytes, updates the cache device list (step C6 in FIG. 9), and stores the hierarchical storage. The control module 123 is notified that it has been secured (step C7 in FIG. 9).

If a dynamically assignable cache is not found (NO in step C3 in FIG. 9),
This is notified (step C7 in FIG. 9), and the hierarchical storage control module 123 executes a stage-out to the medium of the optical disk autochanger 3 in order to free the cache area of Cache0 (step B3 in FIG. 8).
NO, step B4).

When the cache usage rate decreases and a dynamically allocated cache device becomes unnecessary, the hierarchical storage control module 123 sends a cache device release request to the cache device manager 124. And the cache device manager 1
24 releases the corresponding cache area and updates the cache device information list.

As described above, in the first embodiment, an example has been described in which a local storage device or a plurality of storage devices connected to a network are automatically assigned as cache devices as needed.

(Second Embodiment) Next, as a second embodiment, an example in which the configuration of a cache is configured according to the access performance of each network device will be described.

In the second embodiment, as shown in FIG. 10, a multi-stage hierarchical cache is constituted by a plurality of cache devices connected to a network.

In the cache control, it is assumed that data with a high access frequency is allocated to the upper cache side (Cache 0 in FIG. 10).

The stage out path in FIG. 10 is a path for transferring data with low access frequency to the lower cache device, and the stage in path is a path for transferring data with high access frequency to the upper cache device. .

The operation management application 121 dummy-writes a predetermined amount of data to each cache device connected to the network A, and measures the response time. Then, the operation management application 121
Sorts the devices in order from the one with the shortest measured response time, and outputs a recommended cache configuration list indicating the recommended configuration of the cache as shown in FIG.

The tiered storage control module 123 performs tiered storage control by allocating devices from the higher-order cache in the order of devices having shorter response times according to the recommended cache configuration list.

The cache configuration list allows the user to freely reset the cache configuration list from the operation application 121. The timing of measuring the response time can be measured not only immediately before changing the cache configuration but also periodically during operation. These pieces of measurement information are sorted in units of measurement time and stored as a cache configuration recommendation list. As a result, the user can select the cache configuration recommendation list at the peak of the registration / search operation from the operation management application 121, and can perform the cache configuration according to the usage status.

As described above, in the second embodiment, an example in which the configuration of the cache is configured according to the access performance of each network device has been described.

(Third Embodiment) Next, as a third embodiment, an example in which data is arranged according to the reliability of a device connected to a network will be described.

In the network system of the connection form shown in FIG. 1, the cache device (2a) of the host computer (A) 1a is a duplicated nonvolatile memory device or a magnetic disk device of a RAID configuration. In this case, it is preferable to place important data on this device.

Important data is, for example, data such as a management table holding a list structure of a cache used in hierarchical storage control. If this data is destroyed, all data on the cache will be lost. Data should be placed on highly reliable devices.

Therefore, in the third embodiment, the operation management application 121 can specify to which device important data is to be allocated, and the setting result is added to the cache device information list shown in the first embodiment. I do. FIG. 12 is a diagram showing a cache device information list to which the setting result of the important data arrangement is added.

In FIG. 12, data is the upper limit of the data area size, and tbl is the upper limit of the management table area size.

Then, the hierarchical storage control module 123
Refers to this cache device information list and determines to which cache device the cache management table area is allocated.

By arranging important data in a highly reliable device, the risk of losing all cache data due to a hardware failure or the like can be reduced.

As described above, in the third embodiment, an example has been described in which data is arranged according to the reliability of a device connected to a network.

[0074]

As described above in detail, according to the data storage device and the data storage method of the present invention, when the cache memory runs short, the memory device of another computer connected to the network is automatically set as the cache device. On the other hand, when the allocated memory device becomes unnecessary, the memory device is automatically released from the cache device, so that the memory device of another network-connected computer can be effectively used. Become. In addition, since the conflict between the transfer processing to the medium and the search / registration processing due to the shortage of the cache area can be avoided, it is possible to prevent the read / write performance from deteriorating due to the conflict and the throughput from decreasing due to the medium exchange.

Also, the response time of the memory device of another computer is measured, and the memory device is hierarchized in order from the shortest in response time to the cache device having the optimum performance. It is possible to build.

Further, by providing means for managing the reliability of the memory devices of the own computer and other computers and arranging specific data specified in advance in the memory device considered to have the highest reliability, for example, the entire cache is provided. Since important data such as a cache management table for managing data can be consciously arranged in a memory device having the highest reliability, the reliability of the entire data storage system can be improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing a connection form of a network system according to an embodiment of the present invention.

FIG. 2 is an exemplary configuration diagram of the network system according to the embodiment;

FIG. 3 is a view showing a storage format of cache device list information and mount information of an optical disc changer 3 and the like according to the first embodiment;

FIG. 4 is a view exemplifying cache device list information according to the first embodiment;

FIG. 5 is an operation flow relating to cache control in the hierarchical storage control module of the first embodiment.

FIG. 6 is an exemplary view showing a format of a cache device information list created by the operation management application of the first embodiment.

FIG. 7 is a view exemplifying a cache device information list according to the first embodiment;

FIG. 8 shows a first example of the cache control according to the first embodiment.
Operation flow.

FIG. 9 shows a second example of the cache control according to the first embodiment.
Operation flow.

FIG. 10 is a conceptual diagram for explaining a multi-stage hierarchical cache composed of a plurality of cache devices connected to the network according to the second embodiment;

FIG. 11 is a view exemplifying a cache recommended configuration list indicating a recommended cache configuration output by the operation management application according to the second embodiment;

FIG. 12 is a view showing a cache device information list to which a result of setting important data arrangement according to the third embodiment is added.

[Explanation of symbols]

 1a, 1b, 1c Host computer 2a, 2b, 2c Cache disk 3 Optical disk autochanger 11 CPU 12 Semiconductor memory 13 Network interface 14 Storage device 121 Operation management application 122 File system 123 Hierarchical storage control module 124 cache device manager 125 device driver 126 network manager A network B internal bus

Claims (12)

[Claims] 1. A data storage system using at least one or more networked memory devices of a second computer as a cache device of a first computer, wherein the first computer is a candidate for a cache device. A cache candidate designating unit for designating a memory device of the second computer, and a cache for selecting any one of the memory devices designated by the cache candidate designating unit when a cache area is insufficient. A data storage system comprising: a selection unit; and a cache allocation unit that automatically allocates the memory device selected by the cache selection unit as a cache device. 2. The system according to claim 1, wherein the first computer further includes a cache release unit that automatically releases the memory device from the cache device when the memory device allocated as a cache device by the cache allocation unit becomes unnecessary. The data storage system according to claim 1, comprising: 3. A data storage system using at least one or more memory devices of a second computer networked as a cache device of a first computer, wherein the first computer is a candidate for a cache device. A cache candidate designating unit that designates a memory device of the second computer, and a cache allocating unit that assigns a memory device designated by the cache candidate designating unit as a cache device by tiering the memory device in multiple stages. Data storage system characterized. 4. The first computer further comprises measuring means for measuring a response time of a memory device of the second computer, wherein the cache allocating means is a memory device designated by the cache candidate designating means. 5. The data storage system according to claim 4, wherein the data is hierarchized in ascending order of the response time measured by said measuring means. 5. The first computer, comprising: a measuring unit for periodically measuring a response time of the second memory device; a measurement result managing unit for managing a measurement result of the measuring unit for each time zone; A measurement result selection unit for selecting a measurement result in any one of time periods from among the measurement results managed by the measurement result management unit, wherein the cache allocating unit includes a memory designated by the cache candidate designation unit. 5. The device according to claim 4, wherein devices are hierarchized in ascending order of response time in a selection result selected by said measurement result selection means.
A data storage system as described. 6. The reliability management means for managing the reliability of the memory device of the first computer and the memory device of the second computer; 6. The data storage system according to claim 1, further comprising: a data arranging unit arranged in a memory device having high reliability. 7. In a data storage system of a computer connected to at least one or more other computers via a network, a cache management table for holding a mounting state of a memory device of the other computer as a cache device of the own computer. A cache address translation unit that processes address translation for accessing memory devices of the plurality of other computers that are physically different logically as one cache device based on the cache management table. Data storage system characterized. 8. A data storage method for a computer connected to at least one or more other computers via a network, the method comprising: specifying a memory device of the other computer as a candidate for a cache device; And a step of selecting one of the specified memory devices from the specified memory device, and allocating the selected memory device as a cache device. 9. The data storage method according to claim 8, further comprising the step of releasing the memory device allocated as a cache device from the cache device when the memory device becomes unnecessary. 10. A data storage method for a computer that is connected to at least one or more other computers via a network, the method comprising: specifying a memory device of the other computer as a candidate for a cache device; A method of tiering devices in multiple stages and assigning them as cache devices. 11. A data storage method for a computer connected to at least one or more other computers via a network, the method comprising: specifying a memory device of the other computer as a candidate for a cache device; Measuring the response time of a device; and tiering the specified memory device in a hierarchical manner from the top in the order of the shortest measured response time and assigning it as a cache device. . 12. A data storage method for a computer connected to at least one or more other computers via a network, the method comprising: specifying a memory device of the other computer as a candidate for a cache device; Periodically measuring the response time of the device; managing the measurement results by time zone; selecting a measurement result of any of the time zones from the managed measurement results; and A method of assigning a memory device as a cache device after hierarchizing the memory device in the order of shorter response time from the highest in the selected result and as a cache device.