JPH09244954A - Information storage device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain a safe shut-down operation within the time limit of a computer OS without increasing the capacity, price and casing size of a storage battery that is necessary for supplying the electric power via a no-break power unit by providing a control means which controls the information storage capacity of the 1st to 3rd storage means. SOLUTION: This storage device includes of the 1st storage layers 10 and 11 consisting of a RAM cache 12 which caches the RAM serving as a volatile memory, etc., a 2nd storage layer 30 consisting or a magnetic fixed disk(HD) cache 21, etc., and a 3rd storage layer 30 consisting of a MO 33, etc. A clean part is always secured in the cache 21 serving as a nonvolatile memory, so that the dirty data of the cache 12 can be completely stored in the cache 21 without accessing the MO 33. In a shut-down mode, the data of the cache 12 are stored in the clean part of the cache 21. Then a clean part is also secured in the cache 12 in a start-up mode.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information storage device using a hierarchical storage device which stores data by combining a magneto-optical disk as a plurality of portable information storage media and an autochanger, and uses a magnetic fixed disk as a cache. Regarding

[0002]

2. Description of the Related Art In order to realize a large-capacity storage device at low cost, Cheyenne HSM manufactured by Cheyenne and HC manufactured by Novell
Like SS, a portable storage medium such as a magneto-optical disk (hereinafter referred to as MO) or a magnetic tape having a low storage bit unit price is combined with an autochanger to store data, and a magnetic fixed disk (hereinafter referred to as HD). ) Has been commercialized in recent years.

FIG. 21 shows the concept of a hierarchical storage device, in which a semiconductor memory (RAM) and a second storage layer are used as the first storage layer.
3 is an example of a three-layered structure in which HD is used as the storage layer of, and MO and an autochanger are used as the third storage layer.

These hierarchical storage devices are connected to a computer or the like, and when the computer reads / writes data, the data on the MO is cached on the HD. Dirty data for HD capacity (write from host updates data on HD, MO
The upper data is the amount of outdated data), and these hierarchical storage devices are controlled so that the usage rate does not exceed a certain fixed value (watermark). The setting of this watermark is left to the user or the administrator and is set to about 90% to 95% of the size of HD.

Usually, the semiconductor memory with a built-in file system installed in a computer is used to cache the data in the external storage, so that the overall memory is of three layers: semiconductor memory, HD, MO (autochanger). It is a storage device (system).

When the power of this system is turned off and terminated, the memory cannot be retained unless power is supplied (this is called volatile), and the dirty data of the semiconductor cache is retained even when power is not supplied. It must be possible to write to HD (which is called non-volatile). At this time, if there is no free space in HD, it is necessary to write data in MO. Writing to this MO may require replacement of the MO by an autochanger. This MO
Replacement is time consuming (10 MOs to replace 1 MO
Approximately seconds), more power than semiconductor memory and HD is required.

In addition, an uninterruptible power supply (hereinafter referred to as UPS) may be inserted between the system and the power supply in preparation for a power failure. The UPS has a storage battery inside, and normally supplies the power of the power supply to the system. When the power supply from the power supply drops, the storage battery inside the UPS supplies the power to the system and notifies the system of a power failure. . The system that has received the notification of the power failure receives the power supply from the UPS and performs a stop process called "system shutdown" to move the data in the volatile memory to a safe non-volatile state. Even if a power failure occurs due to this process, the data in the system will not be lost.

However, when power is supplied from the UPS to the autochanger to safely shut down, the required capacity of the storage battery becomes large, and the price and the housing become large. Furthermore, some computer operating systems have a limit on the time until shutdown (for example, Microsoft's Wi
ndows-NT3.5® is 2 minutes), there may be no time to replace the MO using the autochanger.

[0009]

As described above, when power is supplied from the uninterruptible power supply (UPS) to the autochanger to safely shut down, the required capacity of the storage battery becomes large, and the price and the housing are increased. There was a problem that it would be a big one. Further, some computer OSs have a limit on the time until shutdown, which causes a problem that the time for exchanging an MO using the autochanger is lost.

Therefore, according to the present invention, it is possible to safely shut down within the time limit of the computer OS without increasing the capacity, price and housing of the storage battery required to supply power from the uninterruptible power supply. An object is to provide an information storage device.

[0011]

According to another aspect of the present invention, there is provided an information storage device comprising: a first storage means made of a volatile storage medium; and a second storage means made of a non-volatile storage medium.
An information storage device configured to include a third storage unit including a plurality of portable storage media and storing information, wherein:
Calculation means for calculating a value obtained by subtracting the information storage capacity of the first storage means from the information storage capacity of the second storage means, and the information storage capacity of the second storage means corresponding to the value calculated by the calculation means. As a result, the amount of information stored in the second storage means that is not stored in the third storage means is reduced, and an overwritable area is provided in at least the first storage area in the second storage means. The storage means comprises a control means for controlling so as to secure the same amount as the information storage capacity of the storage means.

The information storage device of the present invention is the above-mentioned claim 1.
When powering off the first storage means of the described information storage device, the information stored in the first storage means is stored in the first storage means.
Second control means for controlling the storage in the overwritable area of the second storage means secured by the control of the control means. It is composed of

The information storage device of the present invention comprises a first storage means composed of a volatile storage medium, a second storage means composed of a non-volatile storage medium, and a third storage means composed of a plurality of portable storage media. In an information storage device configured to store information and configured to store information, a value is set to secure an overwriteable area in the second storage means for storing the information stored in the first storage means. It comprises a setting means and a changing means for changing the value set by the setting means in order to change the information storage capacity of the overwritable area secured in the second storage means.

The information storage device of the present invention comprises a first storage means comprising a plurality of volatile storage media, a second storage means comprising a non-volatile storage medium, and a third storage means comprising a plurality of portable storage media. In an information storage device configured to store information by storing means, first power supply means for supplying power to a plurality of volatile storage media of the first storage means, and second storage means. The second power supply means for supplying power to the third storage means, and the overwritable area corresponding to the total information storage capacity of the plurality of volatile storage media of the first storage means are provided as the second overwrite area. First control for securing in storage means
Control means and the first power supply means from the first power supply means.
When the power supply to the first storage means is cut off, the information stored in the plurality of volatile storage media of the first storage means is transferred to the first storage means.
And second control means for performing control to save in the overwritable area of the second storage means secured by the control of the control means.

The information storage device of the present invention comprises a first storage means composed of a volatile storage medium, a second storage means composed of a non-volatile storage medium, and a third storage means composed of a plurality of portable storage media. In the information storage device configured to store information, the information storage capacity of the second storage means is used to store information stored in the first storage means that is not stored in the second storage means. From the calculation means for calculating the value obtained by subtracting the information amount and the information storage capacity of the second storage means corresponding to the value calculated by the calculation means, the second
Of the information not stored in the third storage means by the information stored in the second storage means, and the overwritable area is stored in at least the first storage means in the second storage means. It is composed of control means for performing control to secure the same amount as the capacity.

The information storage device of the present invention comprises a first storage means composed of a volatile storage medium, a second storage means composed of a non-volatile storage medium, and a third storage means composed of a plurality of portable storage media. In the information storage device configured to store information, the information storage capacity of the second storage means is used to store information stored in the first storage means that is not stored in the second storage means. A first control is performed to secure an overwritable area corresponding to the amount of information in the second storage means.
When the control means and the first storage means are powered off,
The information stored in the first storage means and not stored in the second storage means is stored in the overwritable area of the second storage means secured by the control of the first control means. And a second control means for performing control.

The information storage device of the present invention comprises a first storage means which is a volatile storage medium, a second storage means which is a non-volatile storage medium, and a third storage means which is a plurality of portable storage media. In the information storage device configured to store information, the overwritable area for storing the information stored in the first storage means and not stored in the second storage means is Setting means for setting the value to be secured in the second storage means, and the value set by the setting means for changing the information storage capacity of the overwritable area secured in the second storage means And the changing means for changing.

According to an eighth aspect of the present invention, there is provided an information storage device comprising a first storage means composed of a plurality of volatile storage media.
In an information storage device configured to include second storage means composed of a non-volatile storage medium and third storage means composed of a plurality of portable storage media to store information, a plurality of volatilization of the first storage means is performed. Power supply means for supplying power to a flexible storage medium, second power supply means for supplying power to the second storage means and the third storage means, and the first storage means. First control means for performing control to secure in the second storage means an overwriteable area corresponding to the information amount of the information not stored in the second storage means in the stored information; When the power supply from the power supply means to the first storage means is cut off, the information stored in the plurality of volatile storage media of the first storage means is stored in the second storage means. No information was secured by the control of the first control means And a second control means for performing control of storing the overwrite area of the serial second storage means.

The information storage device of the present invention is the above-mentioned claim 8.
The information storage device described, and connected to the information storage device,
A fourth storage means composed of a non-volatile storage medium and a control for storing the information stored in the first storage means in the fourth storage means when the information storage device is powered off. The first control means and the second control means for controlling the information stored in the fourth storage means to be stored in the second storage means when the information storage device is activated. .

[0020]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings. First, a first embodiment will be described. FIG. 1 shows a schematic configuration of the information storage device according to the first embodiment.

That is, the information storage device 1 is entirely controlled by the CPU 2. Further, the information storage device 1 includes the first storage layers 10 and 11 including the RAM caches 12 and 12 having the RAM as the volatile memory as the cache and the RAM cache management units 13 and 13, and the magnetic fixed disk (as the nonvolatile memory). Hereinafter, a second storage layer 20 including an HD cache 21 and an HD cache management unit 22 that use HD as a cache, an autochanger 31, an MO 33, ..., An MO drive 34, and an autochanger management unit (hereinafter, referred to as an AC management unit). 36)
The third storage layer 30 is composed of three layers.

The first storage layers 10 and 11 have the same structure, and the information storage device 1 has two sets of the first storage layers. Further, the RAM cache management units 13, 13 of the first storage layers 10, 11 are connected to the bus 4 via the external I / O 14, 14.
0, the HD cache management unit 22 of the second storage layer 20 is connected to the bus 40 via the external I / O 23, and the AC management unit 36 of the third storage layer 30 is connected to the bus 40.

In the first storage layer 10 (and 11),
The RAM cache 12 is provided with a block address for each block (for example, 1 Kbyte). RAM
The cache management unit 13 has a RAM cache block count 13a, a RAM cache usage count 13b, and a RAM cache management table 13c.
Manages the data, and transfers and controls data with other storage layers. Data transfer is performed in block units.

In the second storage layer 20, the HD cache 21 is provided with block addresses for each block. The HD cache management unit 22 uses the number of HD caches 22a, the number of HD cache blocks 22b, the watermark 22d, and the HD cache management table 22.
c, manages data in the HD cache 21, and performs data transfer and control with other storage layers. Data transfer is performed in block units.

In the third storage layer 30, MO numbers, which identify the MO 33, are assigned to the MOs 33, ... Corresponding to the slots 35 in the autochanger 31. The AC management unit 36 has the MO number of the MO 33 being loaded, and the AC management unit 36 controls the autochanger 31 to access the MO 33 between the MO drive 34 and the slot 35.
To exchange the blocks on the MO 33 to transfer and control data with other storage layers. Data transfer is performed in block units.

These first, second and third storage layers 10, 11,
The buses 20 and 30 are connected to each other by a bus 40 (or network). In addition, the first storage layers 10 and 11
The uninterruptible power supply (UP)
S) 41 is connected, and power is supplied for several minutes even after a power failure. Even if a power failure occurs using this power, the system can be safely stopped by the shutdown (described later) without losing the data in the volatile memory (first storage layers 10 and 11).

The data stored in the information storage device 1 has a block to be stored on the MO 33, ... Of the third storage layer 30, and the data is the MO number of the MO 33 in which the block to be stored and the block on the MO 33. Managed by address.

Next, the internal structure of the first storage layer 10 (11) will be described. The total number of blocks in the RAM cache 12 is set by the RAM cache management unit 13 by measuring the size of the RAM cache 12 when the first storage layer 10 (11) is powered on.

RAM cache usage number 13b is RAM
Shows the number of dirty blocks on the cache 12.
The number of dirty blocks is the number of blocks in which the data in the RAM cache 12 is different from the data in the HD cache 21 or MO 33 because the data in the RAM cache 12 has been updated, and is set by the RAM cache management unit 13. .

The RAM cache management table 13c is
This is a table for managing the data cached in the RAM cache 12, and all are set by the RAM cache management unit 13.

FIG. 2 shows a RAM cache management table 1
3c shows a configuration example of 3c. Explaining the first stage of the RAM cache management table 13c shown in FIG. 2 as an example, the data cached in the block of the block address 10 of the RAM cache 12 is the data of the block address 50 of the MO 33 of MO number 0. The data block status is "C", which means that the data is clean.

If the block state is clean C, it means H
It indicates that data is read from the D cache 21 or MO 33 to the RAM cache 12 and copied, and since there is data in the HD cache 21 or MO 33, the data of this block of the RAM cache 12 is not saved. Other data can be cached in the block on the RAM cache 12.

Further, in the contents of the second stage, the data cached in the block of the block address 5 of the RAM cache 12 is the data of the block address 100 of the MO 33 of MO number 1, and the block state is "D". Indicates that it is dirty. Then, in the last stage, the block number of MO number and MO33 is "-1".
Then, the block status is "F" and it is free, and RA
This indicates that the block of the block address 300 of the M cache 12 does not cache data.

Further, the RAM cache management table 13
The order of the row of c is sorted and arranged by LRU (Least Recently Used), and the most recently accessed data is cached in the block at the head of the table.

Next, the internal structure of the second memory layer 20 will be described. The total number of blocks in the HD cache 21 is HD
The HD cache management unit 22 inspects and measures the HD cache 21 when the cache 21 is installed, and uses this value after the installation.

The HD cache usage number 22b is the number of dirty D blocks on the HD cache 21, and
The cache management unit 22 sets it. The dirty D block is a state in which the data in the HD cache 21 is different from the data in the MO 33 by writing the dirty D data from the first storage layers 10 and 11 to the block in the HD cache 21. It is a block.

The watermark 22d indicates the upper limit of the HD cache usage number 22b, and is set by the HD cache management unit 22. There are two methods for setting. First
In this method, the total number of blocks in the HD cache 21 is subtracted from the total number of blocks in the plurality of RAM caches 12, and the sum is set at startup. The second method sets a value obtained by subtracting the sum of the plurality of RAM cache usage numbers 13b from the number of blocks of the HD cache 21 each time the RAM cache usage number 13b changes.

The HD cache management table 22c is H
This is a table for managing the data cached in the D-cache 21, and is the HD cache management unit 2
2 sets.

FIG. 3 shows the HD cache management table 22.
It shows a configuration example of c. The first stage of the HD cache management table 22c shown in FIG. 3 will be described as an example.
The data cached in the block of the block address 2 of the HD cache 21 is MO33 of MO number 0.
The data has the block address 20 and the data block state is dirty D.

In the second stage, the data cached in the block of block address 3 of the HD cache 21 is the block address 1 of MO 33 of MO number 1.
It is the data of No. 5 and indicates that the block state is clean C. The block state being clean C means that the data of the block of the HD cache 21 is stored in the MO 33, and therefore when the other data is cached, the data of this block can be overwritten and used.

At the last stage, MO number, MO3
The block address of 3 is “−1”, the block state is free F, and the block address of the HD cache 21 is 10
Block 24 indicates that the data is not cached.

As in the case of the RAM cache 12, the order of this table is sorted and arranged by LRU (Least Recent Used) so that the most recently accessed data is cached in the block at the head of the table.

The first embodiment has a first method and a second method. The first method uses a watermark to create a writable capacity on the HD (second storage layer 20) immediately.
The capacity of the RAM (first storage layers 10 and 11) is secured, the data on the RAM is stored in the secured area at the time of shutdown, and the HD (second storage layer 2) is again stored at the time of start-up.
0) It is a method to secure again.

FIG. 4 shows the structure of the first system, in which the vertical axis shows the amount of dirty data on the HD and the horizontal axis shows the time, and the HD from the shutdown start to the shutdown end, the startup, and the read / write start. Shows the amount of dirty on HD. The vertical axis also shows the maximum capacity of the HD (second storage layer 20) and the watermark at the same time.

The second method uses a watermark as in the first method, but RAM (first storage layers 10 and 11) is used.
When the data inside is updated and must be stored in a storage layer below the HD (second storage layer 20), the watermark value is lowered and the data can be immediately caught on the HD (second storage layer 20). The size of the data,
Newly secure and add. RAM at shutdown
The data on the (first storage layer 10, 11) is saved in this reserved area. In the case of this method, unlike the first method, there is no need to re-secure at startup.

Next, the operation at the time of starting the first storage layer 10 (11) in such a configuration will be described with reference to the flowchart of FIG. First, the RAM cache management unit 13
Performs a memory test of the RAM cache 21 and checks the usable area to obtain the number of blocks. This value is R
The AM cache block count 13a is set, the RAM cache usage count 13b is set to 0, the usable block address is set in the RAM cache block address of the RAM cache management table 13c, and the MO number and MO block address are set to -1. And sets the block state to "F" (ST1).

In the first method, the clean block reservation request command and the number of RAM cache blocks are subsequently set to HD.
Send to the cache management unit 22 and wait for a response (ST2),
Start accepting requests from external I / O 14, 23 (S
T3).

In the second method, the HD cache management unit 2
Request reception from the external I / O 14, 23 is started without transmitting / receiving to / from No. 2 (ST3). In addition, the HD cache management unit 22 and the AC management unit 36 are connected to the external I / O 1 immediately after starting.
Start accepting requests from 4, 23.

Next, the processing operation of the clean block reservation request will be described with reference to the flowchart of FIG. This processing starts when the HD cache management unit 22 receives the clean block reservation request command and the number of blocks (N) from the external I / Os 14 and 23 (ST11).

The HD cache management unit 22 reduces the watermark by N (ST12), and performs a stage-out process from the HD cache 21 to the MO 33 (described later) until the number of used HD caches becomes equal to the watermark.
(ST13, 14) is repeated, and when the watermark is exceeded, the external I / O 14, 23 is notified of the end of processing (ST15).

By this stage-out process, the dirty blocks on the HD cache 21 are cleaned, and the writable area of the RAM cache 12 at the time of shutdown is secured on the HD cache 21.

Next, the processing operation of the stage out from the HD cache 21 to the MO 33 will be described with reference to the flowchart of FIG. First, the HD cache management unit 22
Searches the HD cache management table 22c from the end of the table for a block whose block status is dirty (ST21). The block address of the found HD cache 21 is Ah, the MO number is N,
This process will be described assuming that the MO block address is Ao.

After the search, the HD cache management unit 22
The write command NNAo is transmitted to the C management unit 36.
After receiving the write command N Ao, the AC management unit 36 prepares the MO 33 with the MO number N in the MO drive 34. Therefore, first, the loaded MO number is compared with N, and if they do not match, Using the accessor 32, the MO 33 is removed from the MO drive 34, returned to the slot 35 in the autochanger 31, the MO 33 with MO number N is taken out from the slot 35 and loaded into the MO drive 34, and the MO number being loaded is set to N. (ST22). And H
A response is returned to the D cache management unit 22.

In response to the response from the AC management unit 36, the HD cache management unit 22 reads out one block of data from the block address Ah of the HD cache 21,
It is transmitted to the management unit 36. The AC management unit 36 writes the received data in the block of the block address Ao of the MO 33 loaded in the MO drive 34 (ST2
3). Finally, the corresponding HD cache management table 22c
The block state of is set to clean C, and the HD cache usage number 22b is decremented by "1" (ST24).

Next, the setting of the watermark by the user such as the system administrator will be described. The watermarks so far are the RAM cache management unit 13 and the HD.
Although the cache management unit 22 is an example of automatically determining, if the first storage layers 10 and 11 are file systems such as personal computers on the network, RAM cache management of the first storage layers 10 and 11 is performed. The unit 12 may not have the function of transmitting a clean block reservation request command.

In this case, the HD cache management unit 22
The system administrator needs to set the watermark. The HD cache management unit 22 displays the CRT display unit 42 or the like connected to the external I / O 23 as shown in FIG. 8, and displays the value input from the keyboard or the like connected to the external I / O 23. Received from and set its value.

In the CRT display section 42 shown in FIG. 8, the watermark (900 in this example) obtained by the HD cache management section 22 in the method 1 is set, and the "OK" key to be left as it is It is possible to change the setting by pressing or the "Cancel" key. If the normal watermark is "900" or more, the autochanger 31 may operate during shutdown.

Next, the processing operation when the RAM cache management unit 13 receives a read / write request from the external I / O 14 will be described with reference to the flowchart of FIG. When a read / write request is received from the external I / O 14 (ST
31), the RAM cache management unit 13 uses the external I / O1
The MO number (hereinafter referred to as Or) and the block address (hereinafter referred to as Bo) are read from 4 (ST32).

Subsequently, the RAM cache management table 13
A block whose MO number is equal to Or and whose MO block address is equal to Bo is searched from c (ST33). If found, Br cache block address to Br
And

If not found, a block is secured from the RAM cache 12 (ST34). First, R
The AM cache management table 13c is searched to see if there is a block whose block status is free F or clean C. If there is, the block address of that block is B
r, MO number is Or, MO block address is Bo
Set. If not, a stage-out process from RAM to HD (described later) is performed, and a search is performed again to secure a block in the RAM cache 12.

Then, the read command, Or and Bo are transmitted to the HD cache management unit 22 (ST35). HD
If the response from the cache management unit 22 is “with data”, the data received from the HD cache management unit 22 is transferred to the secured block of the RAM cache 12 (S
T36). The processing of the HD cache management unit 22 will be described later.

If the response from the HD cache management unit 22 is "no data", a read command, Or and Bo are transmitted to the AC management unit 36. If there is a response from the AC management unit 36, the data is received from the AC management unit 36 and RA
Write to the block of the block address Br of the M cache 12. The processing of the AC management unit 36 is performed in the same manner as the above-described write command processing except for the data transfer direction (ST37).

After the data is read into the RAM cache 12 from the lower storage layer, the RAM cache management unit 13
Sets the block state of the block address Br to clean C (ST38).

In the case of the second method, if the write command received by the RAM cache management unit 13 from the external I / O 14 is write, the block state of the block address Br of the RAM cache 12 is checked (ST3
9) If it is a clean C, a clean block reservation request and the number of blocks 1 are sent to the HD cache management unit 22,
Secure a clean block on the cache 21 (ST
40).

Since the data has been prepared in the RAM cache 12, the RAM cache management unit 13 transfers the data of the block at the block address Br of the RAM cache to the outside from the external I / O 14 in the case of the read request, and the write request. In the case of, the data is transferred from the external I / O 14 to the RAM cache 12 (ST41).

Finally, if the request is a write, RA
The M cache management unit 13 checks the block state of the corresponding block in the RAM cache management table 13c, sets it to dirty D when the block is clean C, and adds 1 to the RAM cache usage number 13b (ST42).

Next, the stage-out processing operation from the RAM cache 12 to the HD cache 21 will be described with reference to the flowchart of FIG. The RAM cache management unit 13 searches the RAM cache management table 13c for a block whose block status is dirty D from the end, and sets the block address of the RAM cache 12 of the found block to Br, the MO number is Or, and the MO block address is Bo. To do. The RAM cache management unit 13 sends a write command, Or, Bo to the HD cache management unit 22 (ST51).

The HD cache management unit 22 receives the above command and parameter, and searches the HD cache management table 22c for a block having an MO number of Or and an MO block address of Bo (ST52). If there is a block, its HD cache block address is set to Bh.

When the search is not found, the CPU 2 sets H
A block is secured on the D cache 21. First, HD
The cache management table 22 is searched for a block whose block status is free F or clean C. HD
Since the number of dirty on the cache 21 is limited by the watermark, it should be found by this search. The HD block address of this block is set to Bh (ST53).

The HD cache management unit 22 determines that the block state of this block is free F or clean C,
Further, it is checked whether the number of used HD caches 21 is equal to or more than the value obtained by subtracting "1" from the watermark (ST54).

If this determination is "YES", the CPU
2 executes the stage-out processing from the HD cache 21 to the MO 33 (ST55). HD cache 21
Since the block has been secured above, the HD cache management table 22 indicates that the RAM cache management unit 1 which is the request source.
3 sends a response. The RAM cache management unit 13
When receiving the response from the HD cache management unit 22, the data of the block address Br is transmitted to the HD cache management unit 22. The HD cache management unit 22 writes the received data in the block with the block address Bh. Finally, the HD cache management unit 22 reads the block state of the block address Bh, adds "1" to the HD cache usage number 22b, and sets the block state to dirty D unless it is dirty D. The RAM cache management unit 13 subtracts “1” from the RAM cache usage number 13b, and sets the block state of the block address Br in the RAM cache management table 13c to clean C (ST5).
6).

Next, the read processing operation of the HD cache management unit 22 will be described with reference to the flowchart of FIG. Upon receiving the read command, the MO number (Or), and the MO block address (Bo), the HD cache management unit 22 performs a read process (ST61). Whether the HD cache 21 is cached or not is searched from the HD cache management table 22c for a block whose MO number is equal to Or and whose MO block address is equal to Bo (S).
T62). If found, the HD cache block address is set to Bh.

If not found, the HD cache management unit 22 returns "no data" to the RAM cache management unit 13 (ST63), and terminates the process. If found, the HD cache manager will say "There is data"
To the RAM cache management unit 13 (ST6
4).

Further, the HD cache management unit 22
The block data of the block address Bh of the cache 21 is transmitted to the RAM cache management unit 13. The RAM cache management unit 13 writes the received data to the block of the block address Br of the RAM cache 12, and sets the block state of the block to clean C (ST
65).

Next, the shutdown processing operation when the RAM cache management unit 13 receives a shutdown request from the external I / O 14 will be described with reference to the flowchart of FIG.

If the RAM cache usage number 13b is "0", the process is terminated (ST71). The RAM cache management unit 13 searches the RAM cache management table 13c for a dirty block, sets the block address to Br, the MO number Or, and the MO block address to Bo. The HD cache management unit 22 receives the shutdown write command, Or, Bo. To send.

When the HD cache management unit 22 receives the shutdown write, the HD cache management table 2
The block having the MO number Or and the MO block address Bo is searched from 2c (ST73). If found, set the block address on this HD cache 21 to B
h. After the search, the HD cache management unit 22 is a RAM
It responds to the cache management unit 13.

If not found, a block whose block status is Free F or Clean C is searched from the end of the HD cache management table 22c, and this block address is set to Bh (ST74). HD cache management unit 2
2 responds to the RAM cache management unit 13.

Upon receiving the response, the RAM cache management unit 13 sets the block data of the block address Br to H level.
It is sent to the D cache management unit 22, the block state of this block is set to free F, and the RAM cache usage number 13
Subtract "1" from b. The HD cache management unit 22 receives the data, writes it in the block of the block address Bh, and sets Or in the MO number and Bo in the MO block address. If the block state of this block is not dirty D, the HD cache usage number 22b is incremented by "1",
The watermark is also incremented by "1" (ST75).

Then, the process returns to the comparison of the RAM cache usage number 13b in step ST71. If the RAM cache usage number 13b becomes "0", all the dirty data on the RAM cache 12 can be saved in the non-volatile memory (HD cache 21), so it is safe to turn off the power. Become.

Finally, the external I / O 14 notifies the shutdown completion. Upon receiving this notification, the user can safely turn off the power without losing data. next,
A second embodiment will be described.

FIG. 13 shows a schematic structure of the information storage device according to the second embodiment. The same parts as those of the information storage device 1 of the first embodiment are designated by the same reference numerals and the description thereof will be omitted. That is, the information storage device 3 is entirely controlled by the CPU 2. Further, the information storage device 3 is
RAM that uses RAM as volatile memory as cache
The caches 12, 12 and the RAM cache management unit 13,
First storage layers 10 and 11 composed of 13, a second storage layer 20 composed of a cache 21 that uses HD as a non-volatile memory as a cache and an HD cache management unit 22, an autochanger 31, a MO 33, ..., A MO drive 34 and an AC.
Three layers including the third storage layer 30 including the management unit 36,
It comprises a backup management unit 61 and a backup layer 60 comprising a backup HD 62.

In the backup layer 60, the backup management section 61 has a backup management table 61a, which manages the data stored in the backup HD 62.

These first, second and third storage layers 10, 11,
20, 30, and the backup layer 60, the bus 40
Connected to each other (or network).
An uninterruptible power supply (UPS) 41 is connected to each of the first storage layers 10 and 11 and the second storage layer 20,
Even after a power failure, power is supplied for a few minutes. Even if a power outage occurs using this power,
By shutting down (described later), the system can be safely stopped without losing data in the volatile memory (first storage layers 10 and 11).

All the data handled by this system has blocks to be stored on the MO 33, ... Of the third storage layer 30, and the MO number and the M number of the block to be stored.
It is managed by the block address on O33.

The internal structure of the first memory layer 10 is the same as that of the first embodiment. The internal structure of the second memory layer 20 is the same as that of the first embodiment except that the watermark 22d is deleted.

The internal structure of the backup layer 60 includes a backup management unit 61, which uses a backup management table 61a as shown in FIG.
Manage the data in 2. The backup HD 62 is provided with a block address in block units, and corresponds to the block address of the backup management table 61a.
It indicates that the data indicated by the MO number and MO block address in the same row of the table is stored. In FIG. 14, for example, in the first row, the data of the MO block address “20” with the MO number “3” is stored in the backup HD 62 of the block address “0”.

FIG. 15 shows the method of the second embodiment, in which a backup area (backup HD62) composed of a non-volatile storage medium other than the non-volatile storage medium (HD cache 21) of the second storage layer 20 is provided. Prepared, save the data of the RAM (first storage layer 10, 11) in this backup area (backup HD62) at the time of shutdown, and store the data of the backup area (backup HD62) at the time of start-up in the non-volatile storage of the second storage layer 20. It is moved to the medium (HD cache 21).

Next, the processing operation when the information storage device 3 is activated in such a configuration will be described with reference to the flowchart of FIG. The backup management unit 61 performs the following processing from the beginning on the backup management table 61a with the MO number Or, the MO block address Bo, and the backup block address Bb (ST8).
1).

A write command, Or, Bo is transmitted to the HD cache management unit 22. HD cache management unit 22
If there is a response from, the backup management unit 61 transmits the block data of the backup address Bb and sets the MO number to "-1" (ST82, 83, 84).

Also, the HD cache management unit 22 and the AC management unit 36 start accepting requests from the external I / Os 14 and 23 immediately after starting (ST85). Next, the operation of the write processing from the RAM cache 12 to the HD cache 21 will be described with reference to the flowchart of FIG.

The HD cache management section 22 searches the HD cache management table 22c for the block with the MO number Or and the MO block address Bo (ST91). If found, this block address is set as Bh and a response is returned to the command transmission source (ST92).

If not found, a block whose block status is Free F or Clean C is searched from the HD cache management table 22c (ST93). If not found, the above-mentioned stage-out processing from the HD cache 21 to the MO 33 is executed (ST94), and a block whose block state is clean C is searched again. The block address of the found block on the HD cache 21 is set to Bh, and a response is returned to the command transmission source (ST92).

The HD cache management unit 22 receives the data and writes it in the block of the block address Bh. If this block state is not dirty D, the block state is set to dirty D and the HD cache usage number 22b is incremented by "1" (ST95).

Next, R in the first memory layer 10 (11)
The processing operation when the AM cache management unit 13 receives a read / write request from the external I / O 14 will be described with reference to the flowchart of FIG.

The RAM cache management unit 13 uses the external I / O
The MO number (hereinafter referred to as Or) and the block address (hereinafter referred to as Bo) are read from O14 (ST101). Next, the RAM cache management table 13c is searched for a block whose MO number is equal to Or and whose MO block address is equal to Bo (ST102). R if found
Let the AM cache block address be Br.

If not found, the block is secured from the RAM cache 12. First, the RAM cache management table 13c is searched to see if there is a block whose block status is Free F or Clean C.
If so, the block address of the block is set to Br, and M
Or is set to the O number and Bo is set to the MO block address. If not, a stage-out process (described later) from the RAM cache 12 to the HD cache 21 is performed, and a search is performed again to secure a block on the RAM cache 12 (ST103).

Then, the read command, Or and Bo are transmitted to the HD cache management unit 22. If the response from the HD cache management unit 2 is "data exists", the data received from the HD cache management unit 22 is transferred to the secured block of the RAM cache 12 (ST105). The processing of the HD cache management unit 22 is similar to that of the first embodiment.

If the response from the HD cache management unit 22 is “no data”, a read command, Or and Bo are transmitted to the AC management unit 36. If there is a response from the AC management unit 36, the data is received from the AC management unit 36 and RA
Write to the block of the block address Br of the M cache 12 (ST106). The processing of the AC management unit 36 is the first
This is the same as the embodiment.

After reading data from the lower storage layer to the RAM cache 12, the RAM cache management unit 13
Sets the block state of the block address Br to clean C (ST107).

Since the data has been prepared in the RAM cache 12, the RAM cache management unit 13 makes a block address Br of the RAM cache 12 in the case of a read request.
Block data is transferred from the external I / O 14 to the outside, and in the case of a write request, on the contrary, from the external I / O 14 to R
Transfer to the AM cache 12 (ST108).

Finally, if the request is a write, RA
The M cache management unit 13 checks the block state of the corresponding block in the RAM cache management table 13c, sets it to dirty D when it is clean C, and adds 1 to the RAM cache usage number 13b (ST109).

Next, the stage-out processing from the RAM cache 12 to the HD cache 21 will be described with reference to the flowchart of FIG. RAM cache management unit 1
3 retrieves a block whose block state is dirty D in the RAM cache management table 13c from the end, and finds the block address of the RAM cache 12 of the found block is Br, MO number is Or, MO block address is Bo.
HD cache management unit 22, write command / O
Transmit r · Bo (ST111).

When the RAM cache management unit 13 receives the response from the HD cache management unit 22 (ST112),
The data of the block address Br is transmitted to the HD cache management unit 22, and the RAM cache usage number 13b is set to "1".
Then, the block state of the block address Br in the RAM cache management table 13c is changed to clean C (S
T113).

The read processing of the HD cache management unit 22 is the same as in the first embodiment. Next, the operation of the shutdown process when the RAM cache management unit 13 receives a shutdown request from the external I / O 14 will be described with reference to the flowchart of FIG.

If the RAM cache usage number 13b is "0", the process is terminated (ST121). The RAM cache management unit 13 searches the RAM cache management table 13c for a dirty D block, sets the block address as Br, the MO number Or, and the MO block address as Bo.
-Send Bo (ST122).

When the backup management unit 61 receives the write command, Or, and Bo, it receives the MO number Or and the MO block address B from the backup management table 61a.
A block that is o is searched (ST123). Let this backup block address be Bb. After the search, the backup management unit 61 responds to the RAM cache management unit 13.

If not found, the backup management unit 61 searches the block having the MO number "-1" from the end of the backup management table 61a, sets this block address as Bb, and sets the Or number and MO block to this MO number. Bo is set to the address and the RAM cache management unit is responded to (ST124).

Upon receiving the response, the RAM cache management unit 13 sends the data of the block having the block address Br to the backup management unit 61, sets the block state of this block to free F, and the RAM cache usage number 13b.
Is subtracted by "1". The backup management unit 61 receives the data and writes it in the block of the block address Bb (ST125).

Subsequently, the RAM in step ST121
Return to the comparison of the cache usage number 13b. If the RAM cache usage number 13b becomes "0", the RAM cache 1
Since all the dirty D data on 2 can be stored in the backup HD 62 which is a non-volatile memory, it means that the block state is safe even if the power is turned off.

As described above, according to the embodiment of the present invention, three or more storage layers (for example, RAM, HD,
In a storage device having a MO changer), in order to shorten the time for shutdown and reduce the required power, dirty data on RAM, which is a volatile memory, can be stored in H
Make sure to secure a clean part in HD so that it can be stored in D, save the data in RAM in this part at the time of shutdown, and secure a clean part in HD at startup, thereby increasing the storage battery. Without shutting down, the computer OS can be shut down within the time limit.

[0112]

As described in detail above, according to the present invention,
It is possible to provide a data processing device that can be safely shut down within the time limit of the computer OS without increasing the capacity, price, or housing of the storage battery required to supply power from the uninterruptible power supply. .

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an information storage device according to a first embodiment.

FIG. 2 is a diagram showing a configuration example of a RAM cache management table.

FIG. 3 is a diagram showing a configuration example of an HD cache management table.

FIG. 4 is a diagram showing a configuration of a first method.

FIG. 5 is a flowchart for explaining an operation at the time of starting the first storage layer.

FIG. 6 is a flowchart for explaining a processing operation of a clean block allocation request.

FIG. 7 is a flowchart for explaining processing operation of stage-out from HD cache to MO.

FIG. 8 is a diagram showing an example of a CRT display unit on which a watermark is displayed.

FIG. 9 is a flowchart illustrating a processing operation when the RAM cache management unit receives a read / write request from an external I / O.

FIG. 10 is a flowchart for explaining a stage-out processing operation from RAM cache to HD cache.

FIG. 11 is a flowchart for explaining a read processing operation of the HD cache management unit.

FIG. 12 is a flowchart illustrating a shutdown processing operation when the RAM cache management unit receives a shutdown request from an external I / O.

FIG. 13 is a block diagram showing a schematic configuration of an information storage device according to a second embodiment.

FIG. 14 is a diagram showing a configuration example of a backup management table.

FIG. 15 is a diagram for explaining the system of the second embodiment.

FIG. 16 is a flowchart for explaining a processing operation when the information storage device is activated.

FIG. 17 is a flowchart for explaining the operation of the write process from the RAM cache to the HD cache.

FIG. 18 is a flowchart illustrating a processing operation when the RAM cache management unit in the first storage layer receives a read / write request from an external I / O.

FIG. 19 is a flowchart illustrating a stage-out process from the RAM cache to the HD cache.

FIG. 20 is a flowchart for explaining the operation of shutdown processing when the RAM cache management unit receives a shutdown request from an external I / O.

FIG. 21 is a diagram showing the concept of a hierarchical storage device.

[Description of Reference Signs] 1, 3 ... Information storage device 2 ... CPU 10, 11 ... First storage layer 12 ... RAM cache 13 ... RAM cache management unit 14, 23 ... External I / O 20 ... Second storage layer 21 ... HD Cache 22 ... HD cache management unit 30 ... Third storage layer 31 ... Autochanger 33 ... Magneto-optical disk (MO) 34 ... MO drive 36 ... Autochanger (AC) management unit 40 ... Bus 41 ... Uninterruptible power supply (UPS)

Claims (9)

[Claims] 1. Information comprising a first storage means composed of a volatile storage medium, a second storage means composed of a non-volatile storage medium, and a third storage means composed of a plurality of portable storage media. And an information storage device for storing, which corresponds to a calculation means for calculating a value obtained by subtracting the information storage capacity of the first storage means from the information storage capacity of the second storage means, and a value calculated by the calculation means. The amount of information stored in the second storage means and not stored in the third storage means is made smaller than the information storage capacity of the second storage means. 2. An information storage device, comprising: a control means for performing control to secure at least the overwriteable area at least as much as the information storage capacity of the first storage means. 2. The information storage device according to claim 1, wherein when the first storage means is powered off, the information stored in the first storage means is secured by the control of the first control means. An information storage device, comprising: a second control means for performing control to save in the overwritable area of the second storage means. 3. Information comprising a first storage means composed of a volatile storage medium, a second storage means composed of a non-volatile storage medium, and a third storage means composed of a plurality of portable storage media. An information storage device for storing the setting means for setting a value for securing an overwritable area in the second storage means for storing the information stored in the first storage means; 2. An information storage device, comprising: a changing unit that changes the value set by the setting unit to change the information storage capacity of the overwritable area secured in the second storage unit. 4. A first storage means composed of a plurality of volatile storage media, a second storage means composed of a non-volatile storage medium, and a third storage means composed of a plurality of portable storage media. An information storage device for storing information by means of a first power supply means for supplying power to a plurality of volatile storage media of the first storage means, the second storage means, and the third storage means. Second power supply means for supplying electric power to the second storage means, and control for securing the overwritable area corresponding to the total information storage capacity of the plurality of volatile storage media of the first storage means in the second storage means When the power supply from the first power supply means to the first storage means is cut off, the data is stored in a plurality of volatile storage media of the first storage means. The information obtained by the control of the first control means Information storage device, wherein the second control means for performing control of storing the overwrite area of the storage unit, by comprising a. 5. Information comprising a first storage means made of a volatile storage medium, a second storage means made of a non-volatile storage medium, and a third storage means made of a plurality of portable storage media. A value obtained by subtracting the information amount of the information stored in the first storage means from the information storage capacity of the second storage means, which information is not stored in the second storage means. The information storage capacity of the second storage means corresponding to the value calculated by the calculation means and the value calculated by the calculation means, and stored in the third storage means by the information stored in the second storage means. Control means for reducing the amount of information that has not been written and for ensuring that the second storage means has at least as much overwriteable area as the information storage capacity of the first storage means. Information storage characterized by apparatus. 6. Information comprising: first storage means made of a volatile storage medium, second storage means made of a non-volatile storage medium, and third storage means made of a plurality of portable storage media. In the information storage device for storing the information storage capacity of the second storage means corresponding to the amount of information stored in the first storage means and not stored in the second storage means. First control means for controlling to secure a writable area in the second storage means, and information stored in the first storage means when the first storage means is powered off. Second control means for performing control to save information not stored in the second storage means in the overwritable area of the second storage means secured by the control of the first control means. An information storage device characterized by the above. 7. Information comprising a first storage means made of a volatile storage medium, a second storage means made of a non-volatile storage medium, and a third storage means made of a plurality of portable storage media. In the information storage device for storing the information, an overwriteable area for storing the information stored in the first storage means and not stored in the second storage means is secured in the second storage means. Setting means for setting a value for setting the value, and changing means for changing the value set by the setting means in order to change the information storage capacity of the overwritable area secured in the second storage means. An information storage device characterized by being provided. 8. A first storage means composed of a plurality of volatile storage media, a second storage means composed of a non-volatile storage medium, and a third storage means composed of a plurality of portable storage media. An information storage device for storing information by means of a first power supply means for supplying power to a plurality of volatile storage media of the first storage means, the second storage means, and the third storage means. Second power supply means for supplying electric power to the second storage means, and an overwritable area corresponding to the amount of information stored in the first storage means and not stored in the second storage means. A second control means for controlling the second storage means and a plurality of volatilizations of the first storage means when the power supply from the first power supply means to the first storage means is cut off. Information stored in the second storage means in the second storage means. Second control means for performing control to save unremembered information in the overwritable area of the second storage means secured by the control of the first control means. Information storage device. 9. The information storage device according to claim 8, further comprising: fourth storage means connected to the information storage device and comprising a non-volatile storage medium; A first control means for controlling the storage of the information stored in the first storage means in the fourth storage means; and the information stored in the fourth storage means when the information storage device is activated. An information storage device comprising: a second control unit that performs control for storing in the second storage unit.